Note: These are the original, non-proofed versions of the 
abstracts accepted for presentation at the 1996 meeting of the 
Society for Psychophysiological Research to be held in Vancouver, 
British Columbia, Canada.  Typographical errors that may be present
in these abstracts may have been corrected at the copy-editing stage.
However, no changes -- typographical or substantive -- can now 
be made, as the Abstract Supplement is already in press. 

Sincerely, Margaret Bradley, 1996 SPR Program Chair   

Abstracts of Papers to be Presented at the 36th Annual 
Meeting of the Society for Psychophysiological Research
The Thirty-Sixth Annual Meeting of the Society for 
Psychophysiological Research will be held October 16-20, 1996
at the Hotel Vancouver in Vancouver, British Columbia, Canada.   
Members of the Program Committee are Margaret Bradley (chair), 
Francis Gabbay, Jutta Globisch, Ottmar Lipp, Harald Schupp, 
Robert Simons, Diane Swick, Cyma Van Petten, Scott Vrana and 
Steven Woodward.
	The program includes three invited addresses and the 
Presidential Address.  Specific topics will be addressed in 
twelve Symposia. The majority of research reports will be 
presented and discussed informally at three Poster Sessions.  
	This web page contains abstracts for each presentation 
in the Symposia, and for the Poster Sessions.  The Symposia 
are printed in the order in which they will occur at the 
conference.  The Poster Session abstracts are published 
together in alphabetical order by first author.
	These abstracts will be published and distributed to 
all members of the Society in a supplement to the journal 
Invited Speakers

Patricia Churchland,  University of California-San Diego
Towards a neurobiology of consciousness

John Gottman University of Washington
What predicts divorce? Building a theory

Marta Kutas, University of California-San Diego, 
Presidential Address 
Meeting language head on: Constraints from anatomy, meaning, 
physiology, and structure

Michael Merzenich, Keck Center for Integrative Neuroscience, 
San Francisco
Specific language impairments: Neurological origins, and an 
effective neurologically-based therapy.


Symposium 1

Off to the Races: Ethnicity in Contemporary 

Chair: Robert F. Simons
Participants: Robert F. Simons, Scott Vrana, Cheryl 
Eric Vanman, Jeanne Tsai

Discussant: Gregory A. Miller

A paper presented recently at a meeting of the American Lung 
Association (Glindmeyer, 1996) reported that women who quit 
smoking and remained 'smoke free' regained significant 
breathing capacity, supporting the clinical observations 
that some tobacco-related damage to lung tissue is 
reversible.  Perhaps more striking, however, was the 
additional finding that this recovery applied only to white 
subjects.  Black women, also smoke free for 20 years, failed 
to regain any breathing capacity as a result of their 
abstinence from smoking.
     This report illustrates, as well as any in the 
literature, the importance of race and ethnicity as a topic 
for psychophysiological investigation. What could account 
for such striking between-group differences?  Fundamental 
differences in anatomy or physiology?  Ancillary differences 
in attitudes, life events or emotion-processing that vitiate 
the effects of smoking cessation?
     In juxtaposition to this provocative study and the many 
questions it raises, a survey of the literature undertaken 
to examine the frequency of appearance of ethnicity in 
psychological research (Graham, 1992), reported that such 
studies are on the decline.  This juxtaposition is cause for 
concern.  Representation of diverse subject groups in 
psychophysiological research and the inclusion of ethnicity 
as a substantive independent variable will be necessary in 
order to understand the type of finding reported by 
Glindmeyer and by others.
     The present symposium consists of five presentations 
involving ethnicity, racial attitudes, and human emotion.  
Though the overarching theme of ethnicity is consistent 
across studies, there is considerable diversity in both 
method and approach. These experiments reveal a number of 
emotion-related differences among Black-, White- and Asian-
American subjects while at the same time illustrating some 
interesting and sometimes surprising similarities.


Robert F. Simons
University of Delaware

Over forty years ago it was reported by Rankin and Campbell 
(1955) that experimenters of an opposite race (black) 
provoked increased arousal in subjects (white).  Though this 
experiment had serious limitations (Petty & Cacioppo, 1983), 
it has nonetheless become part of the psychophysiological 
lore.  The present experiment attempted to put the race-of-
experimenter hypothesis to a more rigorous test by employing 
multiple measures, by employing multiple experimenters, by 
counterbalancing both experimenter and subject race, and by 
examining racial attitudes held by subjects.
     Both Caucasian (high- and low-'racist') and African-
American subjects participated in the present two-part 
experiment.  Part 1 consisted of the greeting, explanation, 
informed consent and psychophysiological 'hookup', all 
conducted by a same-race experimenter, and then the 
unexpected introduction of a laboratory assistant of either 
the same or the opposite race who would conduct the 
remainder of the experiment.  Facial EMG, heart rate and  
skin conductance were measured as the assistant conducted a 
brief demographic interview and then conspicuously touched 
the subject while allegedly checking the quality of the 
electrode placements.  Part 2 consisted of a standardized 
interview which required subjects to address topical issues 
that were either stereotypically race relevant or race 
neutral.  Lastly, subjects were asked to provide affective 
reactions to both the experimenter and the assistant. 
     Black assistants prompted greater HR and SC reactivity 
from their subjects, and this was true during both phases of 
the experiment. Surprisingly, this effect was as strong and 
sometimes stronger among African-American than it was among 
Caucasian subjects.  The racism measure was unrelated to 
physiological reactivity, though it did predict their 
affective ratings of the black and white assistants.


Scott R. Vrana
Purdue University

Two assumptions about differences in social and emotional 
responses of Black Americans and White Americans have 
received inadequate empirical test: that African Americans 
are more emotionally expressive because of the values of 
their African cultural heritage; and that Black Americans 
find interethnic interaction particularly stressful, and 
this stress contributes to Blacks' high levels of chronic 
stress and hypertension.
     These assumptions were examined by having both Black 
and White college students participate in emotional imagery 
and a neutral in vivo interaction with people of the same 
and different ethnic background.  Facial EMG (corrugator, 
zygomaticus) and cardiovascular measures (heart rate, blood 
pressure) were recorded.  Facial EMG results showed that 
White subjects displayed a very positive expression in a 
social greeting, whereas Black subjects exhibited a more 
neutral expression.  In contrast, compared to White 
subjects, during imagery Black subjects displayed a higher 
baseline zygomatic and greater smiling during joy imagery; 
and showed a lower corrugator baseline and less frowning 
during negative imagery.  The imagery study replicated 
earlier blood pressure differences between Black and White 
Americans: Blacks had a higher resting blood pressure and 
greater increases while listening to the emotional imagery 
script.  However, in contrast to the hypothesis stated 
above, it was White subjects (especially White males) who 
exhibited the greatest cardiovascular response (increased 
heart rate and blood pressure) to interethnic interaction 
(greeting and being touched by a Black male).
     These results will be discussed in terms of their 
implications for interethnic relations and the social 
context of emotion.


Cheryl A. Armstead1, Kathleen C. Lawler2, Norman B. 
Anderson3, and Chy Relle Thompson1
1University of South Carolina ,2University of 
Tennessee,3Duke University Medical Center

There is a growing body of literature identifying racism as 
a stressor eliciting acute changes in cardiovascular 
functioning in the laboratory. Laboratory speech stressors 
may yield important qualitative information concerning 
perceptions of racism and behavioral styles utilized in 
coping with racism.  The present study  examined the 
influence of emotional and psychosocial factors on 
reactivity and recovery to a racism speech stressor.
     Thirty African American normotensive females were 
administered a 10 minute racism speech stressor, with blood 
pressure measurements taken speech and during recovery.  
Stepwise regressions, with measures of anger as 
predictors,were performed for the speech task  and during 
recovery.  Several anger measures (e.g. expressed hostility, 
habitual failure to suppress anger) predicted diastolic 
reactivity to the racism speech stressor while other 
measures (educational attainment) predicted diastolic 
recovery from the speech task.  Likewise, the same, or 
closely related measures predicted reactivity and recovery 
of the mean arterial blood pressure measure. There were no 
significant anger-associated predictors of  systolic blood 
pressure responses.
     These preliminary analyses suggest that cardiovascular 
reactivity to a racially salient stressor is influenced by 
both experienced and habitual emotional coping styles among 
African-American females.  Educational attainment and anger 
suppression were consistently predictive of degree of 
recovery after this acute racial stressor.  Thus, 
educational level may be associated with  attenuations of  
autonomic arousal after the task.  These results underscore 
the importance of assessing qualitative aspects of racism-
related laboratory stress responses among African-Americans.


Eric Vanman1 and Tiffany Ito2, 
1University of Southern California, 2The Ohio State 

Previous research demonstrates that facial EMG can reliably 
index  racial bias, even when self report measures cannot.  
The growing popularity of the startle eyeblink reflex as a 
measure of affective response suggests that it too might 
serve as a marker of such bias.  In this study we recorded 
facial EMG from the brow, cheek, and eye regions while 
White,Asian, Hispanic, and Black participants viewed 
pictures of White and Black students. For one half of the 
trials participants were instructed to attend to the 
duration of the black (or white) slides, and for the other 
half were told to ignore the duration of the white (or 
black) slides.  During some trials an acoustic startle probe 
was presented at either 300, 800, or 4500 ms following slide 
onset.  Analyses revealed evidence of racial bias as a 
function of the participant's race.  For the Asian and White 
participants, cheek EMG activity was greater for White 
targets and brow activity was greater for Black targets.  
For these same participants, startle eyeblink amplitudes 
were greater when White targets served as prepulses at 800 
and 4500 ms.  The Hispanic and Black participants tended to 
demonstrate the opposite pattern.  However, at 300 ms probes 
all participants exhibited more startle reflex inhibition 
for attended prepulses, regardless of the target's race.  
These findings indicate that in this paradigm the startle 
eyeblink reflex is more sensitive to attentional processes, 
whereas facial EMG is more sensitive to affective processes 
in racial bias.  We discuss implications for research in 
intergroup relations.


Jeanne L. Tsai and Robert W. Levenson
University of California, Berkeley

Notions that culture influences emotional responses are 
wide-spread.  For example, compared to European Americans, 
members of Asian cultures are believed to demonstrate 
greater emotional moderation, somatize their emotional 
states more, and base their emotional states more on the 
interpersonal nature of the social context.  Few studies 
have examined whether the two groups differ in actual 
emotional responding,and if so, whether such differences are 
consistent with these notions. Studies in our laboratory 
have compared emotional responses (autonomic and somatic 
physiology; dimensional and discrete emotional self- 
reports) in a variety of Asian and European American 
cultures using both intrapersonal (films, loud noises, and 
voluntary emotional facial expressions) and interpersonal 
(dyadic interaction) stimuli.
     In this work we have found little evidence that 
purported emotional moderation in Asian cultures extends to 
the physiological level in either intrapersonal or 
interpersonal contexts.  However, in both contexts, Asians 
have demonstrated greater emotional moderation in reported 
subjective emotional experience.  Cultural differences have 
also been found in relationships between  physiology and 
subjective experience.  For both Asians and European 
Americans, physiological and subjective emotional responses 
were correlated when emotions were elicited in interpersonal 
contexts.  In intrapersonal contexts, these correlations 
were found for European Americans, but not for Asians.
     A model of how culture influences the physiological and 
subjective aspects of emotional responding will be 
discussed, highlighting the role of the social context as a 
mediator of cultural influence.  Our findings will be used 
to evaluate empirically several key ethnographic notions 
about emotion in Asian cultures.

Symposium 2

The Lateralized Readiness Potential: Contributions 
to the study of information processing

Chair: K. Richard Ridderinkhof

Participants: Michael G. H. Coles, K. Richard 
Ridderinkhof, Martin Eimer, Fernando Valle-Inclan

Discussant: Michael G. H. Coles 

In recent years, the lateralized readiness potential (LRP) 
has been extremely fruitful in exploring the structural and 
temporal aspects of the human information processing system.  
The LRP has proven to be a reliable and highly sensitive 
real-time index of the relative activation of competing 
responses in a variety of experimental tasks, and can reveal 
whether and when a particular response is preferentially 
activated.  The present symposium is intended to provide an 
overview of the history and ideas behind the LRP, of 
methodological pittfalls and innovations, and of recent 
empirical results and their implications.  The presentations 
of experimental work will focus in particular on how 
irrelevant stimulation (distractor arrows, locational cues, 
and subliminal primes) exerts its effect on the processing 
of target information.  In one study, LRP evidence indicates 
that distractor arrows prime response activation processes 
directly, without undergoing controlled response selection 
first.  The response thus activated is subsequently 
inhibited when the target stimulus signifies that the 
subject should refrain from responding.  In the second 
study, LRP evidence reveals how the response to a target 
stimulus is impeded by the presence of an identical 
subliminally presented prime stimulus.  The subliminal prime 
briefly activates its corresponding response, which is then 
strongly inhibited.  In the third study, LRP evidence 
emphasizes the role of a central decision mechanism rather 
than automatic response activation when the irrelevant 
stimulation is spatial in nature.  Together, these LRP 
studies yield important new insights into the processing 
architecture and temporal dynamics of interference from 
irrelevant information.


Michael G. H. Coles
University of Illinois at Urbana-Champaign

In this presentation, I will consider the use of measures of 
the lateralized readiness potential (LRP) in the study of 
human  information processing.  The LRP provides an 
indication of the relative activation of two responses 
(usually the left- and right-hands), and, when measured 
appropriately, can reveal whether and when a particular 
response is preferentially activated.  The measure has been 
used to explore the architecture and dynamics of the human 
information processing system. Particular insights have been 
gained into the nature of information transmission, and the 
measure has been especially successful in specifying the 
cognitive locus of various kinds of interference effects and 
in characterizing the effects of aging on cognition.  Issues 
still in need of resolution concern the measurement of the 
onset latency of the LRP, and the interpretation of LRP 


K. Richard Ridderinkhof1, Emil R. Lauer1, and Remco H.J. 
1University of Amsterdam , 2Free University of Amsterdam

Responses to target stimuli are typically delayed when 
irrelevant arrow stimuli point in the incorrect direction.  
Recent behavioral evidence indicates that central S-R 
translation processes operate on target information but not 
on distractors:  A right-pointing distractor always primes 
the right-hand response, even when subjects are required to 
respond to a right-pointing target arrow with a left-hand 
(incompatible) response. Thus, distractor arrows are thought 
to by-pass controlled response selection.
     Two LRP studies used an arrow version of the Eriksen 
flanker task to examine this assertion.  Transient effects 
on response activation manifestthemselves in LRP 
lateralization.  Typically, with a left-pointing target, 
right-pointing distractors briefly activate the right-hand 
response, as indicated by incorrect (subthreshold) LRP 
lateralization just prior to the opposite lateralization 
that precedes the left-hand response to the target. This 
transient LRP effect is known as the Gratton-dip.
     In the first experiment, it was shown that the Gratton-
dip was greatly attenuated when a left-pointing target 
required a right-hand (incompatible) response.  In the 
second experiment, it was shown that on incidental NoGo 
trials, right-pointing distractors initially activated the 
right-hand response, even though instructions required 
incompatible responses.  LRPs showed further that this 
initial right-hand response activation turned into 
subsequent inhibition of that response once the target was 
identified as a NoGo stimulus.  Together, these findings 
indicate that distractor arrows primed responses directly 
without undergoing controlled S-R translation processes 


Martin Eimer
University of Munich

A series of experiments is reported where target stimuli 
were presented requiring either a left-hand or right-hand 
response.  Prior to the target, a masked prime stimulus was 
delivered that was either identical to the target 
(compatible trials) or was identical to the alternative 
target stimulus (incompatible trials).  Subject's verbal 
reports and forced choice performance indicated that they 
were not consciously aware of the presence of the prime 
stimulus.  However, behavioral performance was strongly 
influenced by the prime, with faster RTs and less response 
errors in incompatible as compared to compatible trials.  
The Lateralized Readiness Potential (LRP) waveforms revealed 
the mechanisms responsible for these unexpected effects: An 
initial activation of the response indicated by the prime 
was found, which was then followed by a strong inhibition 
process; this latter process resulted in a partial 
activation of the response side opposite to the side 
indicated by the prime, thereby giving rise to the observed 
behavioral effects.  Further experiments showed that these 
effects are critically dependent on the eccentricity of the 
masked prime, since they were only present with foveal prime 
stimuli.  A tentative functional interpretation of the 
'activation-plus-inhibition' pattern and the cental-
peripheral asymmetry will be offered.


Fernando Valle-Inclan
University of La Coruna

In three experiments the lateralized readiness potential 
(LRP) was used to test current theories regarding the 
effects of task-irrelevant information on choice reaction 
time (RT).  The first study tested the hypothesis that the 
influence of irrelevant spatial information on RT (the Simon 
effect) is due to automatic activation of the ipsilateral 
(to stimulus location) response. By presenting the response 
key labels after the stimulus, we showed that such 
activation is not automatic.
     In Experiment 2 we used a a visual search task and 
manipulated the proportion of trials in which stimulus and 
response locations were contralateral (incompatible trials).  
We found that the Simon effect decreased as the frequency of 
such incompatible trials increased.  The LRP onset latencies 
mirrored the RT data and there were no signs of incorrect 
response activation on incompatible trials.  This latter 
finding indicates that spatial and nonspatial information 
was available for response preparation at the same time, or 
alternatively, that the interference was produced before the 
LRP onset.
     Experiment 3 used LRP and EMG recordings to compare two 
current theories about whether the locus of the accessory 
stimulus effect is at a central decision stage (Posner, 
1978) or at a peripheral motor locus (Sanders, 1980). The 
results indicated that irrelevant accessory stimulation 
affected processes preceding the LRP onset, in agreement 
with the central decision mechanism hypothesis.


Symposium 3

Motivated attention 
Chair: Gudrun Sartory
Participants: Arne Ohman, John T. Cacioppo, Don M. Tucker, 
Peter J. Lang

In the laboratory, selective attention is often conceived of 
as a rational, cognitive activity that is manipulated and 
investigated by instructions, by varying the probability of 
stimulus events, or by altering the stimulus array to 
include 'oddballs' that draw attention. In the natural 
world, however, attention to one type of stimulus rather 
than another may often be dictated by the organism's 
existing motivational state - by drives related to hunger, 
sexual arousal,and threat.  The speakers in this symposium 
present data and theory relevant to the thesis that 
attentional processes are enlisted in the service of 
processing stimuli that have motivational significance, 
compared to routine, affectively neutral events.  


Arne Ohman
Karolinska Institute and Hospital

Functional considerations suggest that motivationally 
relevant stimuli should be detected after a preliminary 
analysis by early, parallel processing mechanisms for 
stimulus recognition. This proposition receives support from 
several lines of evidence. First, research from my 
laboratory has demonstrated that autonomic responses can be 
elicited by stimuli that are prevented from reaching 
conscious recognition by the technique of backward masking. 
Thus,we have reported that phobics show elevated skin 
conductance responses to masked phobic stimuli, and that 
similar masking-resistant responses can be classically 
conditioned in normals, provided that the conditioned 
stimulus is fear-relevant. Second, visual search experiments 
in which the subjects look for deviant stimuli in matrices 
of stimuli demonstrate that normals are quicker to discover 
fear-relevant stimuli against a background of fear-
irrelevant stimuli than vice versa. This effect is enhanced 
if the subject actively fear a stimulus. For example, snake 
phobic are faster to locate snakes than spiders against 
fear-irrelevant backgrounds, even though they are faster 
with spiders then with fear-irrelevant control stimuli. 
Finally, contrary to the original report of the "face-in-
the-crowd effect", normals often are quicker to find happy 
than angry faces against a background of neutral faces. 
However, in a series of studies using schematic facial 
expressions,which allows precise experimental control of 
differential features in emotional facial expressions, we 
demonstrated that angry faces are more quickly located than 
happy faces, regardless of background condition. Thus, it 
appears that emotional stimuli are effective in capturing 
attention and in very rapidly activating psychophysiological 


John T. Cacioppo
The Ohio State University

All organisms must be capable of differentiating hostile 
from hospitable stimuli to survive.  Typically, this 
evaluative discrimination, which governs attentional 
processes, is conceptualized as being bipolar (hostile-
hospitable).  This conceptualization is evident in the areas 
of affect and emotion, where bipolar measures and 
experimental conditions focus on net affective 
predispositions.  Contrary to conceptualizations of the 
motivational processes underlying affect as bipolar, 
evidence suggests that distinguishable motivational systems 
underlie assessments of the positive and negative 
significance of a stimulus.  Thus, a stimulus may vary in 
terms of the strength of positive evaluative activation and 
the strength of negative evaluative activation it evokes.  
Low activation of positive and negative evaluative processes 
by a stimulus reflects neutrality or indifference, whereas 
high activation of positive and negative evaluative 
processes reflects conflict and ambivalence.  As such, 
affect can be represented more completely within a bivariate 
space than along a bipolor continuum.  A model (Cacioppo & 
Berntson, 1994; Cacioppo, Gardner, & Berntson, 1997) is 
discussed in which the positive and negative motivational 
processes underlying affect and emotion are distinguishable 
(stochastically & functionally independent), are 
characterized by distinct negatively accelerating activation 
functions (marginal utility functions, positivity offset, & 
negativity bias), are related differentially to ambivalence 
and conflict, have distinguishable antecedents 
(heteroscedacity), are activated to maximize behavioral 
adaptation (modes of evaluative activation), tend to 
gravitate from a bivariate toward a bipolar structure when 
the underlying beliefs are the target of deliberation or a 
guide for behavior (principle of motivational certainty), 
and reflect individual differences in motivational 


Don M. Tucker1 and Phan Luu2
1University of Oregon & Electrical Geodesics, Inc, 
2University of Oregon 

For behavior to be organized effectively requires motivated 
attention. Adaptively significant goals must be represented 
within working memory so that they shape ongoing action 
plans.  The connections between the motivation and memory 
functions of limbic cortex and the motor planning functions 
of frontal neocortex must be fundamental to this process.  
The densely interconnected paralimbic cortices may serve to 
maintain a global motivational context within which specific 
actions are articulated and sequenced within frontal 
neocortical networks.  The paralimbic networks represent the 
visceral and kinesthetic information that is integral to the 
representation of the bodily self.  In a geneal sense, the 
implicit self-representation within paralimbic networks may 
shape the significance of perceptions and the motivational 
context for developing actions.  The network architecture of 
the frontal lobe reflcts the dual limbic origins of frontal 
cortex, in the dorsal archicortical and ventral 
paleocortical structures.  The dorsal limbic mechanisms 
projecting through the cingulate gyrus may be influenced by 
hedonic evaluations, social attachments, and they may 
initiate a mode of motor control that is holistic and 
impulsive. In contrast, the ventral limbic pathway from the 
amygdala to orbital frontal cortex may implement a tight, 
restricted mode of motor control that reflects adaptive 
constraints of self-preservation.  Each of these motor 
pathways may elaborate a unique mode of working memory that 
allows limbic motivational controls to shape the executive 
attentional functions of the frontal lobe.


Peter J. Lang
University of Florida

Human attention is viewed here as information processing  
that involves procedures of selection, sustained focus, and 
the  evaluation of motivationally relevant input, similar to 
that  occurring in an animal as it forages in a field, 
encounters  others, pursues prey or sexual partners, and 
tries to avoid  predators and comparable dangers.  This 
evolutionary conception  of attention also defines the bio-
behavioral context in which  Pavlov (1927) originally 
proposed the concepts of orienting and  defense.  Both fMRI 
and ERP research strongly implicates sites  in the cerebral 
cortex in the executive attentional functions of  orienting.  
On the other hand, animal research has emphasized  deeper 
structures (amygdala, limbic system, hypothalamus, and  
central gray) as the foundation of primary motivational 
systems  that determine the direction and vigor of action-- 
appetitive  approach and defensive withdrawal.  Orienting 
occurs under  moderate motivation, similar for pleasant or 
unpleasant input,  in a cortically active but overtly 
passive organism.  Increased  imminence of an aversive event 
prompts a staged shift from  passive attention towards 
action mobilization, orchestrated by  the sub-cortical 
motivational circuits.  The initial defensive  reaction does 
not, however, involve stimulus rejection or "a  shutting 
down of the analyzers." It involves more focused  attention 
evaluation-- part of a staged defensive process from  broad 
orienting to physiological mobilization and active fight  or 
flight.  Thus, within the usual range of human  
experimentation, increases in the "intensity" of a threat  
stimulus (e.g., judged negative valence and arousal) are  
accompanied by ERP and autonomic responses consistent with 
an enhanced attentional set.

Symposium 4

Psychophysiological correlates of visual-spatial 
processing and its contribution to performance

Chair: Edmund Wascher
Participants: Steven J. Luck, George R. Mangun, Edmund 
Wascher, Shuhei Yamaguchi

Electrophysiological correlates of visual spatial attention 
can be measured over the temporo-parietal cortex both as 
alterations of early ERP-components (e.g. P1, N2) and as 
event-related asymmetries in early time windows 
independently from discrete ERP-components. The use of brain 
imaging techniques permits the localization of attention-
related activity in the human brain. However, visual spatial 
attention is important both for the perceptual and the motor 
system. The importance of intracortical communication based 
on visual spatial attention becomes evident in the deficits 
of neurological patients. Not only by destruction of 
posterior parietal areas of the human brain but also by 
degeneration of projection deficits of neurological 
patients. Not only by destruction of posterior parietal 
areas of the human brain but also by degeneration of 
projection pathways (e.g. in Parkinson's disease) responding 
to spatial information can be impaired.


Steven J. Luck
University of Iowa

A detailed characterization of the neural mechanisms of 
visual selective attention will require an integration of 
human psychophysical experiments and monkey single-unit 
recordings, but such an integration is made difficult by 
many factors.  ERP recordings can help to bridge the gap 
between the human and monkey domains by providing a means of 
assessing neural activity in humans.  We have recently 
conducted two sets of experiments that were designed to 
provide direct links between human ERP and monkey single-
unit recordings.  In one set of experiments, we adapted an 
ERP attention paradigm for use in monkeys to determine the 
relationship between attentional modulations of single-unit 
activity in area V4 of extrastriate visual cortex and 
previously observed modulations of sensory ERP responses.  
Significant differences were observed between the V4 
attention effects and the previous ERP attention effects, 
suggesting that these effects are not homologous.  In a 
second set of experiments, we adapted a single-unit 
attention paradigm for use in human ERP recordings to 
determine whether a specific human ERP component --the 
"N2pc" wave -- reflects an attentional suppression effect 
that has previously been observed in monkey single-unit 
recordings.  Several experimental manipulations that had 
previously been performed in single-unit experiments were 
found to have analogous effects in the ERP recordings, 
suggesting that the N2pc wave may indeed reflect the same 
attentional suppression effects observed in the single-unit 
recordings. Implications for attentional mechanisms and 
cross-modality/cross-species integration will be discussed.


George R. Mangun
University of California, Davis

In earlier research, we combined event-related potentials 
(ERPs) and positron emission tomography (PET) in studies of 
visual selective attention to spatial location (Heinze et 
al., 1994). We found activations in human extrastriate 
cortex (fusiform gyrus) contralateral to the attended 
hemifield, and related this to ERP attention effects in the 
occipital P1 component (latency of 80-110 msec). Here, using 
the 15O-water PET method and 3-D acquisition, we extend 
these findings by showing that attentional activations at 
the border of the fusiform and lingual gyri were defined by 
spatial selective attention and not higher-order task 
requirements. A slight tendency for the fusiform activity 
and the P1 component of the ERP to be greater in magnitude 
during form discrimination (symbol matching) as compared to 
luminance detection (detect illuminated pixels) was 
obtained, and is interpreted as resulting from differential 
spatial attention allocation policies used in the two tasks. 
In addition to the fusiform activations, the present 
findings also included activations in the middle occipital 
gyrus contralateral to the attended location. The present 
findings significantly strengthen our hypothesis of close 
associations between early ERP attention effects and neural 
activity in the human extrastriate cortex. Specifically, 
they identify an early, spatially defined selection process 
that acts to modulate signals from attended regions of the 
visual field. The relationship of the P1 attention effect to 
the posterior fusiform activations versus the lateral 
occipital activations is investigated using integrated 
modeling of ERP and PET activity. [This work was supported 
by NSF, NIMH, NINDS, the James S. McDonnell Foundation, and 
a UC Faculty Research Grant].


Edmund Wascher
Medical University of Luebeck

In a series of studies, the relationship between movement- 
and stimulus-related processes was investigated. Lateral 
presentation of relevant information facilitates manual 
responses if the side of relevant information corresponds to 
the side of the response. Analogously, manual responses are 
disfacilitated if the spatial position of stimulus and 
response interfere. As a psychophysiological correlate for 
this interaction an increase of negativity contralateral to 
the relevant information was found in the EEG over temporo-
parietal sites. This event-related asymmetry was found at 
central sites (over the motor cortex) as well. However, it 
was markedly smaller. This early asymmetry (200 - 300 ms) 
might reflect a shift of visual spatial attention in common 
with a shift of motor spatial attention towards the source 
of stimulation. This shift of attention influences the 
execution of a manual response. Previous studies showed that 
this interaction is restricted in time, i.e. the 
facilitation/disfacilitation of manual responses is not 
visible if the task is more difficult (if basic response 
times are longer). However, in a new task where the subjects 
had to search in a stimulus array for the relevant 
information, the effect on manual response times was visible 
although response times were markedly prolonged in one 
condition. The early asymmetries indicated that the shift of 
spatial attention was delayed to the same amount as basic 
response times. Therefore, there is accumulating evidence 
that the interaction of stimulus- and response-related 
processes due to spatial information is reflected in phasic 
asymmetries of the EEG. These asymmetries most probably 
reflect a controlled shift of spatial attention which is not 
restricted to the perceptual system.


Shuhei Yamaguchi
Shimane Medical University

Brain cortical activities involved in visuospatial attention 
shift were studied by recording event-related evoked 
potentials (ERPs) during a cued priming task. Dopaminergic 
contribution on the activities were also investigated. 
Normal subjects and patients with Parkinson disease (PD) 
were studied in the task with a central and peripheral cue. 
Attention shift-related negativities (ARNs) were generated 
at the posterior scalp sites contralateral to the shifting 
direction in the early stage, and at the anterior sites in 
the late stage after the central cue onset. The PD group 
showed a marked reduction of the ARN over the frontal site 
while the posterior ARN was delayed in onset compared to the 
normal group. The peripheral cue enhanced the N1 component 
over the contralateral posterior cortices. The late ARN was 
observed over the posterior cortices 500 ms after the 
peripheral cue onset. The ARN for the peripheral cue was 
normal in the PD group. However, behavioral and ERP data 
showed impairment of "inhibition of return" in the PD group. 
The present study suggests that distinct neural systems are 
involved in voluntary and automatic attention shift. 
Anterior and posterior attention systems may contribute to 
the different stage in voluntary attention shift mechanism. 
Mesocortical dopaminergic projection to cerebral cortex may 
play a role for modulating the anterior attention network.


Symposium 5

Time-frequency analysis of event-related brain 

Chair: Scott Makeig
Participants: Scott Makeig, Niels Birbaumer, Bin Shen, 
Astrid Von Stein, Wolfgang Klimesch

Currently, the predominant analysis tool for studying event-
related electrical and magnetic brain dynamics is time-
domain response averaging. Its products (ERPs, ERFs) reveal 
brain activity reliably locked in both time and phase 
(positive or negative) to reference events. However, time- 
and phase-locked activity is only one class of event-related 
brain dynamics. Another class, event-related changes in 
power and/or coherence of the ongoing EEG or MEG spectrum, 
occur over a wide range of frequencies (~1-100+ Hz) and time 
scales (msecs-minutes), and can be measured by averaging 
time-frequency transforms of event-related EEG or MEG 
epochs. Narrow-band (ERD/ERS) measures of changes in EEG 
spectral amplitude have been in use for many years; full-
spectrum time-frequency amplitude and coherence transforms 
are newer research tools. Time-frequency averages measure 
event-related perturbations or modulations of the spectral 
character of the EEG/MEG itself, whereas evoked responses 
typically amount to only a small fraction of the recorded 
brain signals. Like ERPs, time-frequency transforms vary 
both with cognitive state and with stimulus significance. 
They reveal cognitive event-related brain dynamics lasting 
longer than the maximum (<1s) latency of most (>DC) ERP 
phenomena, and thus are prime candidates for relating to 
concurrent changes in brain blood flow. The symposium will 
give a technical overview and survey current work in this 
emerging research area.


Scott Makeig1,2 and Tzyy-Ping Jung1,3
1Naval Health Research Center, 2University of California at 
San Diego,
3The Salk Institute for Biological Studies

From Berger's very first EEG reports, nonstationarity in the 
spectral character of the EEG, as well as apparent 
consistency in the EEG spectral changes accompanying changes 
in wakefulness and attention were noted to be outstanding 
traits of EEG signals.  In 1977, Pfurtscheller described a 
method of averaging the time course of event-related 
reductions in power in a narrow frequency-band and called 
the phenomena event-related desynchronization (ERD). During 
the last decade, increasing observations of subcortical 
brain centers which modulate spontaneous and event-related 
cortical activity, and suggestions that subcortical and 
cortical oscillations may serve to transiently synchronize 
or 'bind' activity in separate brain areas have fueled 
interest in observing the dynamics of EEG/MEG modulation and 
synchronization processes noninvasively in humans. Makeig 
(1993) demonstrated that averaging broad band time-frequency 
transforms of event-related EEG epochs in an auditory 
attention experiment revealed precisely-timed, both narrow- 
and wide-band perturbations in the power spectrum of the 
ongoing EEG, phenomena called event-related spectral 
perturbations (ERSPs). ERSPs need not fit into 
traditionally-defined frequency bands, and often exceed the 
durations and latencies of supra-DC components of
event-related potentials. Time-frequency averaging may be 
accomplished by several methods (FFT, wavelet and Wigner 
transforms, matching pursuit, etc.) Both electric and 
magnetic spectral perturbations have definitely been shown 
to depend on cognitive processing as well as on stimulus 
character, and may represent both transient amplitude 
modulation of ongoing brain activity, or transient changes 
in synchronization of activity in previously desynchronous 
brain areas.


N. Birbaumer, F. Pulvermueller, W. Lutzenberger and H. 
University of Tuebingen

In a series of experiments with meaningful and non-
meaningful visual and verbal stimulus material the 
hypothesis of a functional role of high frequency brain 
oscillations for the construction of meaning was tested.  
EEG-frequency bands were analyzed in 10 Hz wide windows 
between 10 and 80 Hz in the time-frequency domain as 
described by S. Makeig. In the visual modality, an irregular 
pattern of horizontal lines in the four visual fields 
changed into an orderly Gestalt-like pattern (waterfall 
appearance). Over occipital cortices only, spectral power in 
the 35-45 Hz frequency band increased significantly during 
the gestalt-like patterns. A similar effect was seen in the 
25-35 Hz range over left perisylvian cortices for meaningful 
words in comparison to meaningless pseudowords. These data 
support a Hebbian model of Gestalt processing in the brain 
as deduced from animal experiments with simultaneously 
moving visual stimuli (Singer 1995). 
Supported by the German Research Society (DFG).


B. Shen1,2 and John J. B. Allen3
1New Jersey Neuroscience Institute, 2JFK medical center,
3University of Arizona

To enhance the signal-to-noise ratio, event-related 
potentials (ERPs) are often averaged over many trials of 
repeated stimuli. However, this approach presumes that 
cognitive processes remain stationary over those trials, 
which may not be valid in many experimental paradigms. 
Traditional single-trial methods are heuristic and have 
failed to extract information from noisy signals reliably. A 
recently developed technique, matching pursuit (Mallat and 
Zhang, 1993), offers adaptive time-frequency decomposition 
and has been shown to be effective in hippocampal EEG 
analysis (Shen et al., 1995) and ERP studies (Allen et al., 
1995). In this study, we use matching pursuit analysis to 
study the statistics and the time frequency character of the 
classic P3 components of the ERPs in a memory-assessment 
task. The P3 was mathematically modeled by an adaptive 
Gaussian waveform while oscillatory wavelets were screened 
out. This method offered a better measure of P3s because its 
signal-to-noise ratio of a single-trial ERP was larger.  The 
statistics of the amplitudes, the latencies and the 
durations of single-trial P3s were more informative than the 
averaged ERP amplitudes.  The correlation between the P3s 
and other time-frequency components suggested that there 
existed a distinct time-frequency character of the P3 
components.  This approach provides a new method for 
examining ERP data, and a means of accessing the non-
stationarity of cognitive processes.


Astrid von Stein
 University of Vienna, Austria

A paradigm shift has occurred in Neurobiology. The old 
concept of single cell coding is being replaced by coding in 
distributed neuronal cell ensembles. results from cat 
intracortical recordings have shown that synchronization 
among the members of a cell ensemble play a major role in 
this process. This makes it desirable to measure 
synchronization also in humans. In my talk I will show that 
synchronization of neuronal cell ensembles can be detected 
in the frequency components of human scalp EEG. According to 
our results, spectral power in the high beta and in lower 
frequency ranges reflects the degree of synchronization 
between column within a cortical area (von Stein et al. 
1995). Furthermore coherence, the normalized crosspower 
between two cortical signals, gives us the possibility to 
determine synchronization between different cortical areas. 
These interareal synchronizations are of specific importance 
for higher cortical processes where neuronal ensembles are 
thought to extend over large distances. I will present data 
on local and interareal synchronization during auditory and 
visual perception and during several higher cognitive tasks. 
We present stimuli lasting for 1-2 s and measure the 
averaged power/ crosspower spectra over 30-50 trials. We 
find significant task-specific increases of coherence 
between visual cortex and associative cortex during visual 
perception, between several sites in parietal cortex of both 
hemispheres during spatial imagery tasks, between temporal 
and parietal cortex during feature integration, and more. 
Changes occur in single frequency ranges or in combinations 
of frequency ranges. These induced interareal 
synchronizations are usually reproducible within single 
subjects and reveal significant consistencies for groups of 
subjects. Thus, EEG coherence analysis seems to be an 
adequate tool for measuring synchronization within extended 
neuronal ensembles directly in humans.


Wolfgang Klimesch
University of Salzburg, Austria

The results of two memory experiments are reported which 
indicate that theta synchronization (increase in event-
related band power) is associated with episodic memory 
processes. In Experiment 1, a recognition task, the EEG was 
recorded during the study phase in which 96 target words 
were presented and during actual recognition. The results 
indicate that a) in the study phase those words that can 
later be remembered show significantly more theta 
synchronization as compared to words that cannot be 
recognized later and that b) during actual recognition, 
correctly recognized targets exhibit significantly more 
theta synchronization as compared to distractors and not 
recognized targets. For the alpha band it was found that 
successful encoding is associated with desynchronization in 
the lower in the upper alpha band. In order to rule out the 
possibility that theta synchronization simply reflects 
increased effort and/or attention that accompanies 
successful encoding or retrieval, an incidental free recall 
task was carried in Experiment 2. In the study phase, 
subjects categorized the target words of Experiment 1. Then, 
without prior warning subjects had to free recall the words. 
As in Experiment 1, those words that were remembered later 
showed significantly stronger synchronization in the theta 
band. In the alpha band, no significant differences between 
remembered and not remembered words were observed. The 
findings are discussed with respect to a possible 
involvement of hippocampal theta, induced in the cortex via 
hippocampo-cortical feedback loops.

Symposium 6

Dysregulated social behavior during development: 
meanings, biological correlates, and 
psychopathological outcomes

Chairs: Nathan A. Fox and Louis A. Schmidt

Participants: Susan D. Calkins, Louis A. Schmidt, Theodore 
P. Zahn, Adrian Raine

Dysregulated social behavior has multiple meanings and is 
associated with different psychopathological outcomes during 
development.  The four papers presented in this symposium 
will examine the multiple biological contributions and 
developmental course of different types of dysregulated 
social behavior in normal as well as atypical development.  
This symposium will integrate knowledge derived from several 
different psychophysiological approaches and developmental 
time periods.  The symposium will address the following 
issues:  (1) multiple meanings, assessment, and origins of 
dysregulated social behavior during development, (2) 
multiple psychophysiological correlates, (3) behavioral as 
well as psychopathological outcomes, and (4) protective 
factors.  Susan Calkins presents data on the relation 
between cardiac vagal tone and temperament in toddlers.  She 
examines cardiac vagal tone indices of behavioral reactivity 
and regulation and their relation to normal social 
development in children.  Louis Schmidt and Nathan Fox 
discuss the role of frontal cortex in the dysregulation of 
emotion in children.  They report on the contribution of 
frontal EEG activation to maladaptive social behavior in 
toddlers, preschoolers, and seven year-olds.  Theodore Zahn 
examines the autonomic correlates of diagnosis and symptoms 
in childhood psychopathology.  He reports on autonomic 
activity in children diagnosed with behavioral problems and 
other dysregulated behaviors.  Adrian Raine and Peter 
Venables present longitudinal data on autonomic markers and 
protective factors associated with antisocial behavior in 
young adults.  They examine the role of high autonomic 
arousal and information-processing as possible buffers to 
subsequent criminality.  Nathan Fox will serve as the 
discussant and provides perspective on the biological 
contributions and the course of maladaptive social behavior 
during development.


Susan D. Calkins
University of North Carolina at Greensboro

Vagal tone and dynamic changes in Vna (vagal suppression) 
may be physiological indicators or markers of both emotional 
reactivity and regulation, and thus may play a role in the 
development of particular social behaviors and behavior 
problems.  However, the relation between vagal suppression 
and social outcomes may be mediated by behavioral strategies 
such as self-comforting, help-seeking, and distraction that 
assist the child in managing emotional responses.  This 
investigation examined the relations among (1) physiological 
reactivity and regulation (2) emotional reactivity and 
regulation and (3) social behavior in a sample of 44 two- 
and three-year-old children.  Children were observed 
individually across tasks designed to elicit emotionality 
and emotion regulation.  Baseline and task measures of vagal 
tone were computed during these assessments.  Children were 
also observed in naturalistic play with peers several times 
over the course of one month.  Of interest were the 
relations between individual task behaviors and social, 
aggressive, and solitary play behaviors with peers.  The 
data indicate that baseline measures of vagal tone and the 
ability to suppress vagal tone during tasks were 
differentially related to emotional reactivity and emotional 
regulation.  And, children who displayed adaptive emotion 
regulating strategies and displayed an underlying pattern of 
appropriate physiological reactivity and regulation were 
less likely to display problematic social behavior problems 
such as aggression and withdrawal.  These findings will be 
discussed in terms of the adaptive value of physiological 
regulation in the development of self-regulatory behaviors 
that may be critical to social development.


Louis A. Schmidt and Nathan A. Fox
University of Maryland

Recent studies have shown that right frontal EEG activation 
is associated with social withdrawal as well as aggression 
in children.  In the present paper, we compare data from 
three separate studies of toddlers, preschoolers, and seven 
year-olds on (1) measures of frontal EEG asymmetry, (2) 
social behaviors involving peers, and (3) behavioral 
outcomes.  In Study 1, baseline EEG was measured in a group 
of toddlers, some of whom displayed a high proportion of 
externalizing problems.  This group was characterized by a 
pattern of greater relative right frontal EEG asymmetry and 
lower global EEG activation compared with controls. In Study 
2, baseline EEG was measured in a group of preschoolers, 
some of whom displayed a high proportion of shyness and 
sociability during peer play.  Children who were right 
frontal and shy were at risk for internalizing problems, 
while children who were right frontal and highly exuberant 
were at risk for externalizing problems.  In Study 3, EEG 
was measured during a task designed to induce social anxiety 
in a group of seven year-olds.  Children who were classified 
as socially anxious displayed a shift towards right frontal 
EEG asymmetry and an increase in right frontal EEG 
activation during the task in response to challenge.  
Collectively, these data suggest that right frontal EEG 
activation asymmetry may be a biological marker of emotion 
dysregulation and maladaptive social behavioral profiles 
during the first seven years of life.


Theodore P. Zahn
National Institute of Mental Health

Data from studies of children and adolescents with 
Disruptive Behavior Disorders, Obsessive Compulsive 
Disorder, and Childhood-Onset Schizophrenia will be compared 
in an attempt to discern autonomic correlates of different 
types of abnormal psychological functioning.  In all 
studies, skin conductance (SC) and heart rate (HR) were 
recorded during a rest period, an orienting response 
paradigm, and a simple reaction time task.  In addition to 
comparisons with appropriate normal control groups, 
correlations of selected autonomic variables with target 
symptoms were computed.  Disruptive subjects had normal SC 
base levels but attenuated tonic responses to the task.  
Those without a subdiagnosis of conduct disorder had higher 
HR than controls. Autonomic base levels were positively 
correlated with impulsivity and significantly correlated 
with aggression, but the direction depended on the 
aggression measure.  Obsessive patients did not differ 
autonomically from controls, but SC activity correlated 
positively with severity of obsessive-compulsive symptoms. 
Schizophrenia patients showed high resting autonomic 
activity and were greatly impaired on phasic and tonic 
responses throughout the protocol.  A failure of orienting 
was associated with negative symptoms and total symptoms 
while high resting spontaneous SC activity was associated 
with positive symptoms.  In general, the data indicate that 
high levels of autonomic activity are associated with 
behavior control problems -- impulsivity, some types of 
aggression, obsessions, compulsions, and positive symptoms 
of schizophrenia -- whereas low reactivity is associated 
with impaired effortful information processing and social 


Adrian Raine1 and Peter H. Venables2
1University of Southern California; 2York University, 

Nothing is known about biological factors that protect a 
predisposed individual from becoming a criminal.  This 14-
year prospective study tested the hypothesis that antisocial 
adolescents who desist from crime by age 29 years have 
greater autonomic arousal and information-process than 
antisocial adolescents who do become adult criminals.  Skin 
conductance and heart rate measures were taken at age 15 
years in 101 unselected schoolboys.  Of these, 17 antisocial 
adolescents who desisted from adult crime (Desistors) were 
matched on adolescent antisocial behavior and demographic 
variables with 17 antisocial adolescents who became criminal 
by age 29 (Criminals), and 17 nonantisocial, noncriminal 
normal subjects (Non-Antisocials).  Desistors had 
significantly higher resting heart rate, resting skin 
conductance activity, and skin conductance conditioning, and 
faster skin conductance half-recovery times relative to 
criminals.  They also showed higher arousal and conditioning 
relative to Non-Antisocials.  These initial findings 
indicate that better autonomic arousal and information-
processing may serve as factors that protect an otherwise 
predisposed individual from becoming criminal.


Symposium 7

Brain potentials and memory: Recent developments

Chair: Michael D. Rugg

Participants: Ray Johnson Jr., Frank Roesler, Ken Paller, 
Michael D. Rugg

Discussant: Cyma Van Petten

Event-related brain potentials (ERPs) are increasingly 
finding favor as a means of studying human memory, and ERP 
studies have begun to address important issues in several 
domains of memory research. The aim of the symposium is to 
illustrate the diversity of topics in human memory to which 
ERPs are presently being applied, and to show how recent 
findings from ERP studies add to knowledge about both the 
functional architecture and the neural basis of memory 
processes. Each presentations focuses on a differenthem they 
cover working memory, mechanisms of retrieval from long-term 
memory, dissociations between explicit and implicit memory, 
and the component processes underlying recognition memory. 
Common themes arising from these presentations, and 
important questions for future research, will be highlighted 
by the discussant.

Working Memory: An event-related brain potential analysis

Ray Johnson, Jr.1 and Daniel S. Ruchkin2
1Queens College, 2University of Maryland

Ideas about the nature of short-term memory processes have 
undergone substantial change in the past two decades. The 
old view of short-term memory as a unitary store has been 
reconceptualized as a system consisting of three major 
components: 1) an articulatory loop for verbal material, 2) 
a visuo-spatial sketchpad for rehearsing visual material, 
and 3) a central executive which exerts control over the 
other two systems (Baddeley, 1986). Since parts of this 
model remain under-specified and little is known about the 
timing of short-term storage operations, we conducted a 
series of ERP studies of working memory (Ruchkin et al., 
1990, 1992, 1994,1995, submitted). These ERP studies have 
demonstrated the anatomical separation of different working 
memory stores in normal humans. Retention of visuo-spatial 
information elicited negative slow wave activity that was 
largest over right central/parietal scalp whereas retention 
of phonological information elicited long-lasting negative 
slow wave activity that was maximal over left frontal scalp. 
For both types of stimuli, the amplitude of the retention-
related slow waves increased with the amount of information 
held in working memory. The topography of slow wave activity 
that occurred during retention of visuo-spatial information 
further demonstrated that the visuo-spatial sketchpad 
consists of separate object and spatial stores. The 
implications of these and other findings for models of 
working memory will be discussed.

Evidence for memory traces in slow cortical brain activity

Frank Rosler, Martin Heil and Erwin Hennighausen
Philipps-University Marburg

Theories on the architecture of the human memory system 
agree on the assumption that the engram is stored as a set 
of synaptic changes in neocortical tissue. These changes are 
assumed to take place in those brain areas in which the 
information is initially processed.  Retrieval of 
representations is understood as a recreation of the very 
same activation patterns which prevail during initial 
processing. This implies (1) that the same cortical modules 
should be activated during both storage and retrieval, and 
(2) that representations of different quality should be 
reactivated in distinct cortical areas.
     We studied these hypotheses by means of slow event-
related brainpotentials. The main findings are: (1) 
Retrieving associations from long-term memory is accompanied 
by a slow negative shift of 5 - 10microV which prevails 
about as long as the retrieval process lasts, i.e. in our 
experiments, for a period of several seconds. (2) When 
different types of representations have to be reactivated 
the slow negative wave shows a clearly distinct topography. 
Retrieval of different types of representations have to be 
reactivated the slow negative wave shows a clearly distinct 
topography. Retrieval of verbal, spatial, and color 
representations goes together with negative slow waves over 
left-frontal, parietal and occipital areas, respectively. 
(3) The amplitude of the topographic maximum increases with 
the number of representations which have to be reactivated. 
(4) The very same topographic pattern of slow waves is 
recorded during anticipation learning (primary storage of an 
engram) and cued recall (retrieval of well established 
representations). These findings are compatible with the 
idea that memory retrieval implies a reactivation of those 
cortical cell assemblies in the cortex in which the 
constituting features of a mnestic entity had originally 
been processed.

Memory dissociations and associated brain potentials

Ken A. Paller
Northwestern University

Neuropsychological studies of amnesic patients have shown 
that certain types of memory are dependent on processing 
carried out in a set of brain areas that includes medial 
temporal and midline diencephalic regions, whereas other 
types of memory do not require this processing.  One way to 
describe these results is in terms of a distinction between 
"declarative" and "nondeclarative" memory. Performance 
dissociations in normal subjects also lend support to this 
distinction.  For example, the extent of semantic processing 
at encoding tends to have large effects on recognition but 
not on various sorts of priming.  On the other hand, certain 
perceptual manipulations have larger influences on priming 
than on recognition. Evidence from event-related potentials 
(ERPs) has confirmed that, at the time of retrieval, 
different neural processing occurs during recognition and 
priming.  Encoding manipulations that primarily affect 
recognition are associated with ERP effects that can be 
interpreted as correlates of recollection.  Encoding 
manipulations that primarily affect priming are associated 
with ERP effects that can be interpreted as correlates of 
priming.  Differences between ERP correlates of recollection 
and ERP correlates of priming attest to the distinction and 
can be mapped on to the distinctive neural events underlying 
recollection and priming as also studied with other 


Michael D. Rugg
University of St Andrews

The distinction between recognition memory judgments that 
are accompanied or unaccompanied by retrieval of the study 
context is central to so-called 'dual-process' models of 
recognition memory. According to these models, recognition 
memory can be based on two independent kinds of information: 
recollection - the successful retrieval of the original 
study episode (recollection); and familiarity - the feeling 
that a test item has been experienced recently in the 
absence of explicit memory of the study episode. A critical 
distinction between these two forms of recognition memory is 
that only recollection is associated with the retrieval of 
information about study context; recognition based on 
familiarity is acontextual. Thus the two kinds of 
recognition memory can be dissociated by segregating 
accurate recognition memory judgments according to whether 
the recognised items can be correctly assigned to their 
study context. This procedure was employed to study the ERP 
correlates of recognition memory based on recollection and 
familiarity. Two components, one with a left parietal 
distribution, and the other maximal over the right frontal 
scalp, differentiated ERPs elicited by recollected' and 
'unrecollected' items, demonstrating that ERPs are sensitive 
to the processes supporting memory for context (source). 
Subsequent studies demonstrated that the two components 
index dissociable processes, hypothesised to be associated 
with the retrieval of information from episodic memory, and 
the 'post-retrieval' processing of such information, 
respectively. No consistent evidence has yet been found of 
an ERP signature of recognition based upon familiarity; the 
implications of this negative finding for dual-process 
models will be discussed.

Symposium 8

The construction of social reality in the 
psychophysiological laboratory

Chair: William Gerin
Participants: Wolfgang Linden, Kevin Larkin, Nicko 
Christenfeld, Timothy W. Smith,

Discussant: Douglas Carroll

A recent trend in cardiovascular reactivity testing concerns 
the use of stressors which involve social interactions.  It 
seems likely that much of the stress we face in daily life 
comes from the social environment; thus, valid 
representations of social interaction may provide new 
insight into the manner in which stress produces 
physiological effects which may have long-term health 
consequences. However, the development of social interaction 
paradigms is difficult, since we are attempting to create a 
sitaution that engages the subject, and causes her/him to 
behave as though the situation were occurring outside the 
laboratory.  The panel will deal with some of the more 
important issues that must be considered in these sorts of 
studies. Wolfgang Linden discusses gender differences in 
response to an anger-provocation task, and shows that in 
order to elicit reliable responses from female subjects, the 
provocation must be severe, in contrast to males.  Kevin 
Larkin describes social confrontation procedures that may 
provide an ecologically valid means of eliciting 
cardiovascular responses in the laboratory; this work is 
concerned equally with the affective/cognitive, as well as 
physiological, responses.  Nicko Christenfeld discusses the 
role of task engagement as it may affect active coping, and 
its interaction with the social context provided.  Finally, 
Tim Smith describes his work on the interpersonal circumplex 
as a means of understanding social motivation in 
interactions between married couples. Doug Carroll will 
provide an integrative discussion of the factors one must 
consider in attempting to create an ecologically valid 
social interaction in the laboratory.


Wolfgang Linden, Thomas Rutledge, and Tracey Earle
University of British Columbia

Review of the literature suggests that anger provocations in 
the lab trigger large and reliable cardiovascular responses 
that exceed those observed under less challenging 
conditions.  We have now had an opportunity to apply anger 
provocations in three different studies and observed less-
than-perfect replicability of effects.  In Study 1, blood 
pressure and heart rate (BP, HR) changes were twice as large 
as noted in comparable active coping challenges without 
anger provocation.  Men showed greater responses than women.  
In Study 2, the gender difference was replicated.  Anger 
provocation in addition to arithmetic triggered larger 
responses than were apparent during arithmetic alone, but 
that was not true for diastolic BP or for cortisol responses 
in women.  In Study 3, with a women-only sample, an attempt 
at anger provocation superimposed on a 3-task protocol was 
completely ineffective for augmenting BP and HR responses.  
In comparing results from these three studies, it becomes 
clear that anger provocation effects are difficult to find 
in samples of women, require fairly nasty interruption 
statements, and are more likely replicable when the tasks 
involve vocal speech.


Kevin T. Larkin, Nicole Frazer, Elizabeth Semenchuk, and 
Sonia Suchday
West Virginia University

One of the major criticisms confronting investigators of 
cardiovascular responding to stress is the relative lack of 
ecologically-valid strategies for eliciting heart rate and 
blood pressure reactions.  The purpose of the three studies 
presented in this paper is to describe the development and 
validation of a laboratory social confrontation procedure 
designed to measure not only cardiovascular responses, but 
concomitant behavioral and self-reported affective/cognitive 
responses.  From a list of 30 potential laboratory conflict 
situations, two were selected based upon ratings of ecologic 
validity obtained from 40 undergraduate students: a scene 
where a participant is requested to confront a noisy 
neighbor, and another where a participant is requested to 
confront a messy roommate.  In the first study, students 
with a parental history of hypertension evidenced 
significantly greater SBP and more negative confrontive 
behaviors during both scenes.  In the second study, 
instructions for the laboratory task were varied so that 
participants were instructed to express their anger freely 
during one scene and suppress their anger during the other. 
Behavioral, affective, cognitive, and DBP responses were 
elevated during anger expression in contrast to anger 
suppression.  In the third study, students from families 
differing in levels of cohesion and adaptability showed 
differences on cardiovascular and behavioral measures during 
the social confrontation procedure. Based upon these initial 
investigations, we have found that the social confrontation 
challenge provides a reliable method for assessing response 
parameters concurrent to cardiovascular responding, as well 
as measuring more representative real life tasks without 
losing the standardization the laboratory provides.


Nicholas Christenfeld1, Laura Glynn1, James Kulik1, and 
William Gerin2
1University of California-San Diego, 2Cornell University

Studies of cardiovascular reactivity generally compare the 
level of a physiological variable, such as blood pressure, 
during a task to the resting level.  The difference is 
interpreted as reflecting the effect of that task.  However, 
such changes may reflect the extent to which the subject is 
engaged in the task as much as the nature of the task 
itself. In the present study we examine two variables that 
might influence task engagement.  Subjects performed a word 
search task while blood pressure and heart rate were 
continuously monitored.  Half received performance norms and 
half received no information on how many words could be 
found. Subjects provided with norms showed significantly 
greater systolic blood pressure (SBP) increases during the 
task (10.9 vs. 5.4 mmHg).  The second manipulation explored 
whether the same difference could be created by simply 
altering the experimenter's behavior.  For half the 
subjects, the experimenter provided encouragement (though no 
performance feedback), and for half the experimenter did not 
provide encouragement.  The encouraged subjects showed a 
significantly greater increase in SBP (10.5 vs. 5.6 mmHg).  
Encouraged subjects also indicated that they felt greater 
control over the outcome.  The results indicate that the way 
the task is approached, not just the nature of the task, 
determines physiological responses.  They also indicate that 
positive regard from an experimenter does not always produce 
decreased reactivity (the social support effect), but may 
increase arousal by increasing the effort put into the task. 
Increased reactivity is not necessarily a marker of stress, 
but can also indicate enjoyable active coping.


Timothy W. Smith and Linda C. Gallo
University of Utah

The interpersonal circumplex is a useful conceptual and 
methodological tool in studies of social determinants of 
cardiovascular reactivity.  Two broad classes of such 
influences are motives involving agency (i.e., competition, 
dominance, achievement) and communion (i.e., relatedness, 
conflict).  To evaluate the effects of these social 
motivations on cardiovascular reactivity during marital 
interaction, 60 married couples participated in a current 
events discussion task.  In a factorial design, half of the 
couples were told that their verbal intelligence would be 
evaluated (vs simple clarity of speech) and half were 
assigned to opposite sides of the topic to be discussed (vs. 
the same side).  Responses to Wiggins' Interpersonal 
Adjective Checklist indicated that disagreement led wives to 
view their husbands as more hostile and less friendly, but 
did not alter husbands' descriptions of their wives' 
behavior during the task.  Evaluative threat led husbands to 
rate their wives as more dominant and less submissive, but 
did not alter wives' ratings of their husbands.  Evaluative 
threat(but not disagreement) increased systolic, diastolic 
and heart rate responses of husbands, but not wives.  
Disagreement (but not evaluative threat) increased systolic, 
diastolic, and heart rate responses of wives, but not 
husbands. Thus, the interpersonal circumplex provided a 
dimensional conceptualization of sex differences in 
reactivity in this context, as well as a sensitive 
assessment of the corresponding interpersonal appraisals.


Symposium 9

GO-NOGO: Cerebral inhibitive control mechanisms and 

Chair: C.H.M.Brunia

Participants: J. Richard Jennings, G.J.M. van Boxtel, 
Martin Eimer, C.H.M. Brunia

In recent theories about attention its intimate relation to 
action is stressed. Both attention and action are based upon 
inhibitory mechanisms, by which selection in the input and 
the output channels of the brain is possible. The first 
contribution is theoretical: it applies the selection for 
action view upon psycho-physiological results obtained 
during the foreperiod of reaction time tasks. In the next 
two contributions electrophysiological responses obtained in 
a number of experiments will be interpreted in terms of 
patterns in inhibition and excitation. In the fourth 
contribution the possible neuro-anatomical pathways 
underlying inhibition in attention and action will be 
described. The aim of this symposium is to attempt to find a 
common framework in which behavioral, psychophysiological 
and neuroanatomical data concerning response activation and 
inhibition can be put together.


J. Richard Jennings1 and Maurits W. van der Molen2
1University of Pittsburgh, 2University of Amsterdam

As argued by Neumann, Allport, and others, both perceptual 
and motoric selection may arise from the requirements of 
action. Physical actions must be coordinated not only 
between motor effectors, but also with perceptions of the 
environment and physiological state.  This view provides an 
alternative to resource theories of attention.  Further, 
coordination for action implies the psychophysiological 
changes occurring during preparation. Preparation is 
examined in warned reaction time paradigms by varying the 
timing and predictability of and information about the 
anticipated stimulus and the response required. Preparation 
necessarily implies a cost. Other actions--including the 
maintenance and formation of perceptions linked to action--
must be inhibited to permit the programming and execution of 
the anticipated action.  Preparation time can be minimized 
and efficient with information about the timing and nature 
of the anticipated response. An assessment of the precision 
of the anticipated perception-action linkage and how it 
interacts with competing perception-action linkages will 
predict physiological changes occurring in the preparatory 
period. Results with heart rate and anticipatory cortical 
potentials can be considered with the assumption that 
transient heart rate change is one aspect of the 
coordination of response initiation and regional cortical 
changes interpreted as facilitation and/or inhibition of 
different perceptual-motor linkages. This analysis is most 
compelling when performance at a fixed instant requires that 
motor programming deal solely with the anticipated action, 
inducing a blocking of interfering actions.  Coordination 
for action will be illustrated during both mnemonic 
retrieval and during selection of a simple motor response.  


G.J.M. van Boxtel1,2, W.P.M. van den Wildenberg1, M.W. van 
der Molen2, J.R. Jennings3, and C.H.M. Brunia1
1Tilburg University, 2University of Amsterdam, 3University 
of Pittsburgh

Event-related brain potentials (ERPs) from 28 mostly 
anterior electrode positions, heart rate, respiration, 
agonist and antagonist EMG, and continuous force output were 
recorded in a stop-signal task. The respond stimulus was an 
arrow pointing left or right. On 70% of the trials, it was 
colored green and the subjects had to make a response with 
the left or the right index finger (GO trials). On 10% of 
the trials, it was colored red, and no response was to 
follow (NOGO trials). On 20% of the trials, the arrow was 
green, but after a variable interval it briefly turned red, 
upon which the subjects had to inhibit their response (STOP 
trials). The latency of the STOP signal was chosen so that 
the subjects could inhibit their response on half of the 
trials. The behavioral results fitted into a horse race 
model, in which independent activation and inhibition 
processes race for completion. Response inhibition was 
accompanied by heartbeat slowing, which was largest when the 
stimulus was presented late in the cardiac cycle. The ERPs 
were characterized by a small transient (pre)frontal 
negativity followed by a large broad central positivity. 
Both components were larger on inhibition trials than in 
response trials. The first negative component was much 
smaller or absent in STOP trials than in NOGO trials. 
Results are discussed in terms of current behavioral and 
neurophysiological theories of inhibitory control.


Martin Eimer
University of Munich

Processes of selective attention have recently been 
characterized in terms of selection-for-action, thus 
suggesting a close relationship between attentional 
mechanisms and motor preparation and activation processes. 
Electrophysiological correlates of perceptual and motor 
selectivity will be discussed, focussing mainly on ERP 
indicators of attentional and motor inhibition processes. 
Three areas of research will be reviewed: (1) Studies 
investigating the time course and distribution of ERP 
effects related to attention and motor inhibition elicited 
by single visual and auditory stimuli in a precueing 
paradigm; (2) Studies investigating mechanisms of selecting 
response-relevant information (or supressing irrelevant 
information) in visual search tasks; (3) Studies 
investigating the inhibition of response tendencies that are 
elicited by masked non-targets with the help of the 
Lateralized Readiness Potential (LRP).        


Tilburg University

Order in our behavior is based upon spatiotemporal selective 
processes that take place both in the input channels and in 
the output channels. In order to understand behavior it is 
necessary to understand how selection takes place and where. 
It is a common notion that the sensory cortex is a target 
for afferent volleys, informing the brain about both the 
outside world and our body, while the motor cortex is 
considered a starting point for efferent volleys, that via 
the motoneurons give rise to muscle contractions to react or 
act in the outside world. However, just as the sensory 
cortex, the motor cortex needs to be informed via 
subcortical pathways before the efferent stream of 
information actually can give rise to any change in our 
behavior. The thalamus is a crucial relay center for sensory 
information from different modalities to separated cortical 
areas, known as the primary projection areas. The motor 
nuclei in the thalamus are also part of separated thalamo-
cortical systems. Input from all cortical areas is sent to 
the neostriatum, via which these separate systems 
areactivated. It is hypothesized that comparable control 
mechanisms from the prefrontal cortex are active in the 
patterning of inhibition in the anticipation  of attention 
and action. We will try to interpret the results of the 
first three papers in terms of the (expanded) model, 
presented at the San Diego meeting. 

Symposium 10
The Psychophysiology of Fatigue

Chairs: Evan A. Byrne and Wolfram Boucsein

Participants: Akihiro Yagi, June J. Skelly, James C. 
Miller, Mark R. Rosekind

Discussant: John A. Stern

The problem of fatigue is becoming increasingly recognized 
as an important safety concern in many human endeavors 
ranging from workplace safety to commercial transport.  In 
laboratory research fatigue is an established construct in 
theories of attention and vigilance.  Fatigue is a dominant 
research area for many applied psychophysiologists and 
crosses the domains of human factors and ergonomics, 
transportation and health, and basic attention processes.  
Each paper in this symposium addresses the role of  
psychophysiological methods as an approach to understand 
fatigue.  Using distinct methodology the speakers in this 
symposium will demonstrate the role of psychophysiology in 
the study and detection of fatigue in laboratory and applied 
settings.  Yagi presents data on task enjoyment and fatigue 
using a new measure based on cortical potentials associated 
with eye movements.  Skelly will present data on the use of 
EOG-based measures to index the magnitude of attention in 
sustained task performance.  Miller will present EEG data 
collected on long-haul truck drivers using a unique 
methodological approach to detect periods of fatigue.  
Finally, Rosekind will describe the use of physiological 
measures, including EEG, EOG, EMG, & HR, in the study of 
fatigue in aviation environments as part of the NASA Ames 
Fatigue Countermeasures Program.  These presentations 
capture the breadth of research questions, methodology, and 
environments encountered by psychophysiologists interested 
in fatigue.


Akihiro Yagi
Kawnsei Gakuin University

Application of ERP is difficult in ergonomics because of 
eye-movement artifacts.  When a subject observes something, 
the eye movement record shows a step-like pattern consisting 
of saccades and eye fixations.  We can obtain a specific ERP 
called the eye fixation related potential (EFRP) with 
averaging EEG time-locked to onset of eye fixations (i.e.  
offset of saccades).  Two experiments were conducted 
evaluating the EFRP as a potential index of fatigue.  In the 
first experiment, 10 subjects performed figure drawing tasks 
on positive (white) and negative (black) screens for 40 min 
while EFRP was measured.  EFRP differences were evident 
between the positive and negative screens.  A small 
decrement in the EFRP with time on task was observed in some 
subjects.  All subjects enjoyed the task.  In the second 
experiment, 13 subjects performed VDT tasks during a 1.5 
hour task period.  Tasks included a free search task and two 
kinds of computer games.  EFRP was obtained during the task 
period and during rest conditions immediately pre- and post-
task.  EFRP latency decreased slightly during the free 
search task for 15 min.  Subjects enjoyed the VDT tasks and 
games and few reports of fatigue symptoms occurred during 
these conditions and EFRP shows few changes.  However, 
fatigue symptoms increased post-task and post-task EFRP 
latency was significantly delayed compared to latencies 
obtained pre-task.  The EFRP may be applicable as an index 
of fatigue in long lasting tasks.


June J. Skelly
Wright-Patterson AFB, OH

A series of studies investigated the effects of portraying 
temporally patterned information on subjects' ability to:  
(1) maintain a state of sustained attention over time; (2) 
switch attention among different streams of sequential 
information.  Subjects engaged in a simple decision making 
task where task relevant information was variably mapped 
across different temporally patterned sequential information 
streams.  Subjects participated in three studies over a 3 
month period and performed the task over 4 consecutive days 
in each study.  Subjects monitored two co-occurring, 
temporally patterned information sources for target 
information for 2.5 hours in each daily session.  Results 
indicated that: (1) response times and EOG measures were 
differentially affected by temporal patterns associated with 
both relevant and irrelevant information sequences; (2) 
residual effects on attentional flexibility, due to exposure 
to certain timing patterns, were evident across days within 
a study, and across studies separated by at least one month; 
(3) temporal pattern complexity appeared to affect a 
subject's attentional state and ability to maintain a high 
level of attention.  A later study attempted to design a 
Temporal Interface for an F-16 cockpit display to evaluate 
temporally patterned sequential information in a real world 
setting.  Using temporal structures from earlier studies, 
but with rates increased tenfold, a warning cue was 
periodically presented on a modified Radar Warning Receiver.  
Pilots had to initiate countermeasures when they detected 
the signal.  Results indicated attention was captured by 
certain timing patterns of information.  These studies 
suggest ways to exploit and energize an operator's 
attentional mechanism in either a slow vigilance task 
setting, or a rapidly changing, dynamically complex work 


James C. Miller
Miller Ergonomics and The Scripps Research Institute 

The data came from 320 round-trips contributed by 80 
commercial drivers (4 or 5 trips each) driving both day and 
night revenue cargo runs of 10 or 13 hours each.  The sleep 
and driving EEG data were collected with Medilog recorders 
at 128 samples/sec. Non-artifactual data collected when the 
truck was moving faster than 45 mph were reduced through 
Fourier transform to compressed band arrays for 20-sec 
epochs, with associated manual scoring to identify 
artifactual and drowsy epochs.  Each 20-second-epoch was 
represented by the mean of five 4-second epochs.  The 
arcsines of relative amplitude estimate within each band of 
interest (theta, TH; alpha, AL) were used, and were 
standardized with reference to the relative spectral 
amplitude of arcsine relative beta (BE).  A within-subject 
z-score was based upon the mean and standard deviation of 
the first 100 useable 20-second epochs that occurred when 
the truck was traveling greater than 45 mph during the 
drivers first trip of the week. The z-scores were averaged 
into sequential, 1-minute z-scores.  The means and the 
rectilinear regression slopes for the standard scores of  
TH/BE and AL/BE were used to estimate EEG values at the 
beginning and end of each outbound and inbound leg of each 
trip for each driver.  One-week-long time series plots of 
TH/BE and AL/BE were prepared, showing the slopes and 
intercepts of each leg of each trip.  The time series showed 
evidence of acute fatigue and cumulative fatigue, and 
circadian and semi-circadian patterns of drowsiness.

Evaluating fatigue in operational settings:  The NASA Ames 
Fatigue Countermeasures Program

Mark R. Rosekind1, Kevin Gregory2, Donna Miller2, Lissa 
Webbon2, and Ray Oyung3
1NASA Ames Research Center, 2Sterling Software,3San Jose 
State University

n response to a 1980 Congressional request, NASAAmes 
initiated a program to examine fatigue in flight 
operations.The Program objectives are to examine fatigue, 
sleep loss, andcircadian disruption in flight operations, 
determine the effectsof these factors on flight crew 
performance, and the developmentof fatigue countermeasures.  
The NASA Ames FatigueCountermeasures Program conducts 
controlled laboratoryexperiments, full-mission flight 
simulations, and field studies.A range of subjective, 
behavioral, performance, physiological,and environmental 
measures are used depending on studyobjectives.  The Program 
has developed substantial expertise ingathering data during 
actual flight operations and in other worksettings.  This 
has required the development of ambulatory andother measures 
that can be carried throughout the world and usedat 41,000' 
in aircraft cockpits.  The NASA Ames FatigueCountermeasures 
Program has examined fatigue in shorthaul,longhaul, 
overnight cargo, and helicopter operations.  A recentstudy 
of planned cockpit rest periods demonstrated 
theeffectiveness of a brief inflight nap to improve 
pilotperformance and alertness.  This study involved 
inflight reactiontime/vigilance performance testing and 
EEG/EOG measures ofphysiological alertness.  The NASA Ames 
Fatigue CountermeasuresProgram has applied scientific 
findings to the development ofeducation and training 
materials on fatigue countermeasures,input to federal 
regulatory activities on pilot flight, duty, andrest 
requirements, and support of National Transportation 
SafetyBoard accident investigations.  Current activities are 
examiningfatigue in nonaugmented longhaul flights, 
regional/commuterflight operations, corporate/business 
aviation, andpsychophysiological variables related to 


Symposium 11

Electrophysiologic studies of depression and 
anxiety: New findings and theoretical synthesis

Chairs: Gerard E. Bruder and Richard J. Davidson

Participants: Richard J. Davidson, Wendy Heller, Gerard 
Bruder, Bruce Cuthbert

Discussant: David Watson 

There is considerable comorbidity of depressive and anxiety 
disorders, and self-ratings of depression and anxiety are 
highly correlated. The extent to which depressive and 
anxiety disorders have a common or different pathophysiology 
is an important issue addressed in this symposium. Recent 
efforts have been directed at developing psychometric scales 
that can distinguish symptom subtypes of depression and 
anxiety. Relatively little attention has, however, been paid 
to assessing the differential effects of depressive and 
anxiety symptoms in electrophysiologic and neuroimaging 
studies. There is evidence that depression and anxiety have 
different effects on regional hemispheric activity, which 
could explain some conflicting findings in anxiety have 
different effects on regional hemispheric activity, which 
could explain some conflicting findings in this area. The 
presentations in this symposium review new findings from 
electrophysiologic and imaging studies that evaluate the 
effects of comorbidity, diagnostic subtype, and specific 
symptoms of depression and anxiety on a variety of measures, 
including quantitative EEG, PET scans, event-related brain 
potentials, startle responses, autonomic responses, and 
neuropsychologic tests. Theoretical formulations are also 
presented for integrating these findings. The presentations 
will be discussed in relation to the "tripartite model" of 
anxiety and depression (Clark &Watson, 1991, J. Abnorm. 
Psychol., 316-336).


Richard J. Davidson
University of Wisconsin-Madison

Depression and anxiety share certain symptoms in common and 
also feature particular distinguishing symptoms. A common 
observation in clinical research is the extensive co-
morbidity between depression and anxiety. These facts raise 
questions about the extent to which they share certain 
biological substrates in common and also whether the current 
diagnostic categories adequately capture the variance 
commonly observed in nature. This talk will first present a 
model based upon prior data and theory of the similarities 
and differences in the cortical and subcortical substrates 
of anxiety and depression, focusing on prefrontal and 
parietal cortex and the amygdala. Data from several 
electrophysiological and neuroimaging studies on cortical 
asymmetries and amygdala activation in both depressed and 
anxious subjects will be presented. Different patterns of 
prefrontal electrophysiology characterize depressives with 
or without certain features of anxiety. All depressives 
appear to show increased metabolic activity in the amygdala 
and greater metabolic activity in the amygdala is associated 
with increased negative affect. Subjects with social phobia 
show increased right-sided prefrontal and parietal 
activation. These findings lead to the suggestion that 
activation of the amygdala may be common to both depression 
and anxiety, while cortical activity is likely to differ 
between them. The implications of these data for parsing 
mood and anxiety disorders is considered, along with 
specifying the particular subcomponents of affective 
processing that may be impaired in these different types of 


Wendy Heller
University of Illinois

Numerous studies have documented specific patterns of 
anterior and posterior brain activity associated with 
depression and anxiety. However, discrepancies, failures-to-
replicate, and opposing findings abound. In an attempt to 
arrive at a better understanding of the brain mechanisms 
involved, I will review this literature in light of a model 
of emotion and regional brain activity which integrates 
neuropsychological and psychophysiological data on 
cognitive, emotional, and autonomic functioning with the 
valence and arousal dimensions of the circumplex structure 
of emotion. In previous work, we have related core symptoms 
of depression and anxiety to the valence and arousal 
dimensions of the circumplex model of emotion, and have, in 
turn, linked these dimensions to specific patterns of brain 
activity. Following from the theory and our previous 
findings, I propose that attention be given to two crucial 
issues that are too often overlooked in the relevant 
literature, specifically, subtypes and comorbidity. In 
particular, a review of the literature and our own empirical 
findings indicate that melancholic versus non-melancholic 
depression and anxious apprehension versus anxious arousal 
are important subtypes of depression and anxiety that appear 
to be associated with different, at times even opposing, 
patterns of regional brain activity. Therefore, the 
comorbidity of depression and anxiety (and their subtypes) 
may substantially confound findings for a given sample. In 
sum, the inconsistencies in the literature may be mitigated 
by taking into account subtypes and comorbidity in future 
research paradigms examining depression and anxiety.


Gerard Bruder, Craig Tenke, Regan Fong, Paul Leite,
James Towey, Jonathan Stewart, and Frederic Quitkin
New York State Psychiatric Institute

Recent theories and behavioral laterality findings emphasize 
the importance of distinguishing effects of depression from 
those of anxiety. Based on a model of asymmetric hemispheric 
activity in depression and anxiety (Heller et al., 1995, J. 
Abnorm. Psychol.,327-333), it was predicted that anxious and 
nonanxious depressed patients would differ in EEG alpha 
asymmetry at parietotemporal sites. Resting EEG (eyes closed 
and open) was recorded from 44 unmedicated patients having a 
unipolar major depressive disorder (19 with and 25 without 
an anxiety disorder), and 26 normal controls using 30 scalp 
electrodes. As predicted, depressed patients with an anxiety 
disorder differed from those without an anxiety disorder in 
alpha asymmetry at posterior but not at anterior sites. 
Nonanxious depressed patients showed an alpha asymmetry 
indicative of relatively less activation over right 
parietotemporal sites, whereas anxious depressed patients 
showed the opposite alpha asymmetry. In the same study, ERPs 
were recorded during an oddball task with complex tones. 
Normal controls and patients with low scores on a physical 
anhedonia scale had greater P3 amplitude over right than 
left central sites, whereas patients with high anhedonia 
scores did not show this right hemisphere advantage. These 
findings are consistent with a model in which anhedonic 
depression is associated with right parietotemporal 
hypoactivation, whereas anxious arousal is associated with 
hyperactivation in this region.

Fear and anxiety: Theoretical distinction and clinical test

Bruce N. Cuthbert, Margaret M. Bradley, and Peter J. Lang
University of Florida

Clinical studies  of anxiety disorder patients are described 
that utilize multi-measure psychophysiological recording to 
assess emotional reactivity in perception, imagination, and 
action. This research presents the data from a large (n=150) 
group of patients, including those with simple phobia, 
social phobia, panic disorder, or post-traumatic stress 
disorder (PTSD). The assessment procedure involves interview 
and questionnaire measures, followed by psychophysiological 
monitoring during emotional imagery and picture processing 
as well as during standard physical (e.g., step test) and 
psychological challenge tasks (e.g., speech performance).  
In addition to dependent measures of autonomic and 
respiratory activity, these studies assess affective 
modulation of the probe startle reflex.  The data 
demonstrate significant variation among patients in the 
relationships between affective judgments and autonomic and 
startle responses.   Panic and PTSD patients showed 
significantly larger base startle responses than simple or 
social phobics, whereas specific clinically-relevant fear 
stimuli prompted greater startle potentiation for those 
diagnosed with simple or social phobia.   Marked differences 
in  autonomic and probe startle reflex responses were 
obtained between patients showing high and low negative 
affect and generalized psychopathology.   This data base is  
discussed in terms of how it  aids in sharpening 
psychophysiological distinctions between fear (i.e., phobia) 
and anxiety (as in generalized anxiety disorder),  as well 
as how it illuminates issues regarding differential 
diagnosis and the significance of depressive co-morbidity in 
anxiety disorders.


David Watson
University of Iowa

Clark and Watson (1991, J. Abnorm. Psychol., 316-336) 
proposed a 'tripartite model' of depression and anxiety that 
divides relevant symptoms into 3 groups: symptoms of general 
distress that are largely nonspecific, manifestations of 
anhedonia and low positive affect that are specific to 
depression, and symptoms of somatic arousal that are 
relatively unique to anxiety. The model originally was 
developed from measures of self-rated symptoms. However, the 
papers in this symposium will examine the differentiation 
between anxiety and depression using various 
psychophysiological indices. I will integrate the findings 
from the individual papers and examine the extent to which 
they are compatible with the general tripartite model.

Symposium 12

On becoming selective: Integrating cognitive and 
electrophysiological approaches to the development 
of attention

Chair: Lourdes Anllo-Vento
Participants: James T. Enns, James M. Swanson, David 
Friedman, Margot J. Taylor 

Researchers interested in understanding the developmental 
time-course of attention have taken widely varied approaches 
to the issue. Some have used development as a tool to help 
test the validity of a particular model of attention. Others 
have resorted to attentional theories to shed light upon 
developmental disorders of attention. And still others have 
employed attentional theories and psychopathology to 
determine how attention develops.  This symposium will 
examine what can be accomplished by integrating well-known 
cognitive and psychophysiological paradigms and applying 
them to the study of normal and abnormal development of 
attention.  The first speaker will illustrate the benefits 
of using well-defined behavioral tests to assess changes in 
attentional capabilites over the lifespan.  We shall see how 
different attention tasks seem to follow distinct 
developmental courses, information that might prove useful 
in conceptualizing the sub-components of the attentional 
system.  The next speaker will illustrate how theory-driven 
behavioral analyses can help establish a taxonomy of 
attention deficits. The last two speakers will focus on 
electrophysiological measures of auditory and visual 
selective attention, respectively, and will consider the 
benefits afforded by jointly probing neural and behavioral 
indices of cognitive development. The aim of the symposium 
is to help us establish effective experimental strategies 
that incorporate the knowledge accrued from cognitive and 
psychophysiological approaches to the question at hand.  


James T. Enns
University of British Columbia

I will highlight some of the behavioral patterns we have 
seen emerge in our laboratory over the past decade or so on 
research on the development of visual attention.  We began 
our studies of visual attention in school-age children 
(e.g., Enns & Girgus, 1985; Enns & Brodeur, 1989) but in the 
past few years have begun to compare these trends with those 
that can be seen at the other end of life (Brodeur & Enns, 
submitted; Plude, Enns & Brodeur, 1994; Trick & Enns, 
submitted; Trick, Enns & Brodeur, in press).  The data I 
will summarize will focus on three separable methods of 
attention measurement that are known to vary in childhood as 
well as in old age: covert visual orienting,  visual search, 
and visual enumeration.  In each case it will be shown that 
there are differential rates of development for sub-
components of these tasks. In some cases there is almost no 
measureable change across the full span of life, in other 
cases there is improvement early in life but no decline in 
later life, and in still other cases the oft-cited inverted-
U function is observed.  These results have alternately 
provoked optimistism and pessimism in us, as we have 
explored the possibility of linking these behavioral changes 
over the lifespan to theories of development, to theories of 
attention, and to changes in the observable nervous system.


James M. Swanson
University of California at Irvine

Over the past few years, a team of investigators from UC 
Irvine, UCLA, and the University of Oregon has applied a 
combined approach to the description and measurement of 
cognitive deficits in children with attention deficit 
hyperactivity disorder (ADHD). One group within this project 
has focused on behavioral subtyping, a second on the 
measurement of structural and functional abnormalities in 
brain architecture, and a third has taken as a point of 
departure the idea advanced by Posner and his colleagues 
that a neuroanatomical network mediates the various aspects 
of attentive behavior: alerting (right frontal), orienting 
(right parietal), and executive control (anterior 
cingulate).  In this vein, we have recently extended our 
understanding of visual-spatial orienting in ADHD children 
by investigating the impact of comorbid conditions on a 
child's performance on the Posner cuing task.  In this task, 
a "pure" ADHD subgroup showed the abnormal pattern of 
reaction-time differences (RVF>LVF) when responding to 
targets after 800 ms SOAs (Swanson et al.,1991), while a 
"mixed" subgroup with ADHD and comorbid disorders did not 
(Ottolini, 1995).  This finding suggests that specific ADHD 
subgroups may not have the same type of attention deficits.  
We are now carrying out high density electrophysiological 
recordings on neuropsychologicallly homogeneous groups of 
ADHD children in order to characterize the temporal aspects 
of their information processing deficits.  We expect this 
approach will help us understand why different attention 
deficits are manifested by distinct subtypes of children 
with ADHD.


David Friedman1 and Steve Berman2
1New York State Psychiatric Institute, 2UCLA 
Neuropsychiatric Institute

The development of auditory selective attention was assessed 
using event-related potentials (ERPs) and behavioral 
measures, with children, adolescents, and young adults as 
subjects. In separate blocks, subjects heard two sequences 
of pure tones (low- and high-pitched) or consonant-vowel 
syllables (CVs; e.g., ba vs da). Subjects were required to 
attend to one of the two stimuli in order to detect a 
deviant (longer-duration) target embedded within the 
attended sequence, while ignoring the sequence comprised of 
the other stimulus (which also contained standard and 
deviant stimuli). The effect of selective attention was 
operationalized using the Nd difference waveform (ERP 
elicited by the unattended standard subtracted from that 
elicited by the attended standard). There was an increase in 
early Nd amplitude and a decrease in its latency for both 
pure tones and CVs from childhood through young adulthood. 
For the amplitude measure, this effect was much more marked 
for CVs. The major effect of age involved reduction of 
negative-going ERP amplitude elicited by stimuli in the 
unattended channel, which was accompanied by an age-related 
decrease in the number of behavioral responses to deviants 
in the unattended channel. The age-related reduction in 
negativity to stimuli in the unattended channel was 
interpreted as indicating that with increasing age there is 
an improvement in the narrowing of the attentional focus, 
with the major change taking the form of greater facility in 
rejecting stimuli in the unattended channel. The data appear 
to fit Naatanen's model (1982; 1990) of selective attention 
based on the processing negativity.


Margot J. Taylor
University of Toronto

We have investigated the development of attentional 
processes in normal and clinical groups of school-aged 
children.  I will review three even-related potential (ERP) 
studies with emphasis on the earlier components.  In the 
first, visual search paradigms were used to study parallel 
processing of pop-out stimuli and serial processing of 
feature conjunctions.  Posterior P1 and N2, and anterior P2 
and N2 were measured. In 7-8 yr-olds there was little effect 
of task; with increasing age, the processing associated with 
the colour pop-out became faster than that of the size on 
serial tasks.  These data suggest that developmental changes 
in early stages of visual selection are continuing at this 
age. Data from normal children will be compared to those 
from children with attention deficit hyperactivity disorder 
(ADHD).  The second study assessed the effects of medication 
on a continuous performance task.  On medication ADHD 
children showed a normalization of P3 and behavioural 
measures, but increased N2 and P2 latencies, suggesting more 
controlled decision processes leading to improved 
behavioural scores.  The third study examined patterns of 
hemispheric specialization during reading tasks in 8-10 yr-
olds with very low birth weight (VLBW).  While control 
children showed LR at N2, VLBW 
children showed only Rdergraduate
student subjects (3 male, 21 female). The 10 imagery scripts were based
on those used by Fiorito & Simons (1994, Psychophysiology,513-521) and
were designed to elicit emotions within all four quadrants of the
emotion circumplex defined by the affective dimensions of valence and
arousal. Each imagery trial consisted of a 20 second initial relaxation
period, a 30 second period during which the script was presented to the
subject, and a 30 second period during which the subject was required
to imagine the scene described. Following this, subjects made ratings
of the valence, arousal, and vividness of the emotional imagery.
Startle probes were presented either 12, 18 or 24 seconds into the 30
second imagery period. Factor analyses of subject's ratings of the
scripts allowed them to be placed into four groupings; (1) high
arousal, positive valence, (2) high arousal, negative valence, (3) low
arousal, positive valence, and (4) low arousal, negative valence. A
series of 2 x 2 ANOVAs indicated that although vividness ratings were
similar across the classes of scripts, startle magnitude was
significantly greater during the two high arousal scripts than the two
low arousal scripts. Emotional valence did not have a significant
influence on startle magnitude. These results suggest that the
emotional determinants of startle reflex magnitude differ between
perception of internally generated and externally generated emotional

Effects of misinformation on the Concealed Knowledge Test
Susan Amato-Henderson1, Charles R. Honts2, and Joseph J. Plaud1
1University of North Dakota, 2Boise State University

The current study used a repeated measures design to examine the
effects of misinformation on the validity of the Concealed Knowledge
Test (CKT).  Ninety-six subjects were made knowledgable by watching a
12-minute videotaped crime of a burglary.  One week later subjects were
given misinformation concerning three details of the crime.  The
misinformation was contained in a narrative presented to the subjects
as a memory refresher.  Subjects then took the CKT, which inquired
about 3 non-misinformed and 3 misinformed crime details.  The
non-misinformed CKT series consisted of the key and 5 foils, whereas
the misinformed series consisted of the key, the misinformation and 4
foils.  To determine the effect of misinformation on CKT performance,
skin resistance amplitude data was scored using Lykken's procedure
(1959, Journal of Applied Psychology, 43, 385-388).  Scores on the 3
non-misled series (M = 3.57, SD = 1.65) were significantly larger than
scores on the three misled series [M = 2.52, SD = 1.71; t(91) = -4.80,
p < .000].  Furthermore, no significant differences existed between
scores associated with the key (M = 2.52, SD = 1.71) on the
non-misinformed series and scores associated with the misinformation (M
= 2.46, SD = 1.54) on the three misled series [t(93) = .24, p =.813].
In real-world applications it would not be possible to differentiate
misinformed items from other foils.  Results of this study suggest that
the mere introduction of misinformation (rather than a demonstrated
misinformation effect) would lead to a higher rate of false negative
errors.  These findings should lead one to question the utility of the
CKT in real-world situations due to the many possible sources of
misinformation to which a guilty individual may be exposed (e.g.,
inaccurate media reports, interrogation).

Effects of caffeine and expectancy of caffeine on the human startle
Siobhan E. Andrews1, Terry D. Blumenthal1, and Magne A. Flaten2
1Wake Forest University, 2University of Tromso

Repeated use of a drug such as caffeine may constitute a classical
conditioning procedure, where the effect of caffeine is the
unconditioned stimulus, and stimuli reliably accompanying the intake of
caffeine, e.g. the sight and smell taste of coffee, may constitute
conditioned stimuli that can elicit conditioned responses. An
experiment was performed that investigated the effect of caffeine and
expectancy of caffeine on the startle reflex (periorbital eyeblink
EMG).  Nineteen habitual caffeine users received caffeinated coffee,
caffeinated juice, decaffeinated coffee, and decaffeinated juice in
four counterbalanced sessions spaced one week apart.  Juice without
caffeine constituted the baseline, whereas juice with caffeine should
give information about the effect of caffeine without any conditioning
or expectancy effects. Responses to decaffeinated coffee should give
information about conditioned responses, and responses to the
caffeinated coffee should give information about the interaction of
caffeine with conditioned responses. The results showed that caffeine
increased startle eyeblink amplitude, but had no effect on startle
probability or latency. A caffeine by solution interaction was
significant for response latency.  In the decaffeinated coffee
condition, subjects had significantly longer reflex latencies than in
the other three conditions, indicating that the ingestion of coffee may
have elicited a compensatory response that was overridden by the
caffeine response. Thus, two processes may have been acting at the same
time, a conditioned slowing of responding based on the expectation of
caffeine, and a speeding up of responding back to some optimal level by
caffeine itself.

Following the time course of feature extraction with event-related
brain potentials
Lourdes Anllo-Vento and Steven A. Hillyard
University of California, San Diego

During selective attention to multifeature stimuli, it has been
proposed that an initial stage of parallel feature extraction gives way
to attentive processing of feature conjunctions.  This experiment was
designed to determine the time course and topographical distribution of
event-related potentials (ERPs) associated with selective processing of
color and orientation features and their conjunctions.  ERPs were
recorded while subjects viewed foveally presented vertical or
horizontal red or orange bars and tried to detect infrequent
shortenings of one of the bars of the designated color and
orientation.  EEG was recorded from 42 scalp electrodes and averaged
separately for each of the four color-orientation combinations under
each of four attention conditions, where each combination was attended
in turn.  The ERP waveforms revealed several distinct phases in the
processing of color and orientation conjunctions.  Between 100-200 ms
there was evidence of independent selection of color and orientation
features, but not of the relevant conjunction.  At about 200 ms
post-stimulus, there was a significant attentional modulation due to
the selective processing of the relevant feature conjunction, along
with evidence of continued parallel processing of each of the two
features.  A still later phase of the ERP waveform reflected selective
processing of the target feature of bar length.  Thus, ERPs provide an
effective means of delineating the temporal course of independent and
conjunctive stages of feature selection.

Effects of personality on skin conductance habituation.
Peter Annas, Lisa Ekselius, Lars von Knorring, and Mats Fredrikson
Uppsala University,

Habituation is a form of non-associative learning by which the
organism learns not to respond to repeated irrelevant stimuli. Studies
have shown that anxious and neurotic subjects tend to habituate slower
than normal controls. However, results have been conflicting for
studies on sensation seeking and habituation. Both differences and
similarities between subjects being high in sensation seeking and
normal controls have been reported. The aim was to study habituation of
skin conductances responses in 190 subjects high or low (median split)
in somatic anxiety, psychastenia and monotony avoidance measured
through the Karolinska Scale of Personality. Both somatic anxiety and
psychastenia are components in neuroticism and monotony avoidance is
akin to sensation seeking.  Skin conductance responses were measured to
4 seconds, 80 dB tones presented through headphones with an interval
ranging from 10 to 15 seconds. Subjects with high somatic anxiety and
psychastenia habituated significantly slower as compared to those being
low. A reversed pattern appeared for monotony avoidance. Subjects high
in monotony avoidance showed markedly faster habituation than those
being low. These results indicate that the anxiety component in
neuroticism is involved in determining habituation and that sensation
seekers, estimated through monotony avoidandce, habituate faster than
normal controls.

Genetic influences on the skin conductance orienting reaction to
fear-relevant and -irrelevant stimuli.
Peter Annas and Mats Fredrikson
Uppsala University

The orienting reaction reflects attention to novel stimuli.  Previous
studies indicate enhanced orienting reactions to fear-relevant stimuli
(e.g. pictures of snakes and spiders) as compared to fear-irrelevant
stimuli (e.g. pictures of circles and triangles).  The aim was to
investigate whether orienting reactions to these stimuli are
genetically determined. 77 monozygotic (MZ) and 72 dizygotic (DZ) twin
pairs were shown  4 slides with fear-relevant or fear-irrelevant
content presented for 8 seconds during which skin conductance were
measured.  Skin conductance responses (SCR's) were range corrected and
averaged over trials.  Correlations between SCR magnitude in twin pairs
were higher for fear-relevant than for -irrelevant stimuli for both
zygosity groups.  The correlations in MZ twins was double that of DZ
twins for both types of stimuli indicating additative genetic effects.
With linear structual modeling a stronger genetic effect was evident
for fear-relevant stimuli than for fear-irrelevant stimuli.  We
conclude that the orienting reaction in part is genetically modulated
and that the genetic influence is greater for fear-relevant than for
fear- irrelevant stimuli.

EEG correlates of psychometric intelligence in adolescents: Coherence
and dimensional complexity
Andrey P. Anokhin
Washington University School of Medicine

This study investigated relationships between global properties of
brain electric activity under different conditions and intelligence.
EEG was recordered monopolarly from 10 symmetric leads (10-20 system)
in 37 (17 males) healthy subjects (mean age 13.7 yrs) under resting
condition with eyes closed and during performance of two visually
presented cognitive tasks, verbal (semantic grouping) and spatial
(mental rotation). On another occasion, the subjects were administered
the Amthauer's Intelligence Structure Test (IST). Both total IST score
and some individual subtests of specific abilities showed significant
positive correlations with EEG coherence in the theta band and
significant negative relationships with the correlation dimension of
the EEG (a measure reflecting the complexity and unpredictability of of
neural dynamics underlying the EEG time series).  Furthermore these two
EEG parameters were inversely related to each other. Taken together,
these EEG metrics accounted for about 30% of the total IQ variance in
this sample. No significant effects of the task type (spatial vs.
verbal) or specific abilities were observed. Long-distance coherence
indices (fronto-parietal and fronto-occipital) showed the most
consistent relationship with cognitive abilities. The results suggest
that the order to chaos ratio in task-related brain dynamics may be one
of the biological factors underlying cognitive performance in

The role of stimulus preceding negativity and heart rate deceleration
as an index of attention
Ross Apparies, Brad Hatfield, Laine SantaMaria, and Thomas Spalding
University of Maryland

The current study endeavors to directly assess the role of attention
in eliciting the SPN, and to identify the SPN's relationship with heart
rate deceleration (HRD).  Accordingly, 20 right handed college age
males performed a cued discrimination and reaction time task.  This
task consisted of a tone (S1) which provided information about the
upcoming difficulty of S2, (consisting of two bars of varying height,
easy difference 3", difficult difference 1/4") presented on a computer
screen for 100 ms.  Since the subject was only required to make a
decision of which bar (right or left) was higher, the S1-S2 interval
was not contaminated by motor preparation.  The subject was required to
respond according to their decision at S3.  EEG, referenced to average
ears, was obtained at C3, C4, P3, and P4, using a lowpass filter of 30
HZ, and amplified 50,000 times.  Heart rate was obtained in interbeat
intervals accurate to the millisecond.
The EEG data were subjected to a 2 x 2 x 2 (Difficulty x
Laterality x Region) ANOVA.  This ANOVA yielded no significant
effects, however the trends in the data show that the difficult
discrimination is preceded by a greater negativity then the easy
discrimination.  The HRD data, which was subjected to a t-test, found
significant differences between the easy and difficult trials.
Although the trends in the SPN and HRD data are in the same direction,
no correlation was found between these measures.  The trends suggest,
that the SPN and HRD, show increased effects prior to a task that
requires increased cognitive demands.

Event-related potentials during recognition memory for pictures
Maria Luisa Armilio, Terence W. Picton and Fergus I.M. Craik
University of Toronto

Event-related potentials (ERPs) were recorded from 46 electrodes
during encoding and retrieval of complex, coloured pictures.  Subjects
determined whether the pictures were 'old' (seen previously) or 'new'
(never before presented) during recognition.  All responses during
recognition showed an N150 and a prominent P240 over occipitoparietal
sites; both waves were larger over the right hemisphere.  These may
represent the automatic processing of nonverbal stimuli.  A frontal
negativity at 450 ms (N450) that was more prominent to correctly
recognized new pictures may signify some early memory process
associated with novelty detection.  The successful retrieval of the old
pictures from memory may be denoted by the centroparietal positive wave
(P550) that was more prominent and earlier in latency to correctly
identified old pictures.  Beginning at approximately 600 ms, a positive
slow wave arose over prefrontal sites in some subjects and lasted until
the end of the recording epoch.  This was more positive in the ERPs
elicited by the new pictures and suggests some further processing of
those stimuli recognized as novel.  These late, cognitive components
provide a likely time course for these operations in memory.

Slow cortical positivity in 6-year-old children during an S1-S2
Gregory J. Austin, W. Keith Berg, and Helen Fields
University of Florida

During a fixed-foreperiod interval, adult's slow brain potential moves
negatively in a pattern known as the CNV.  To our knowledge there is no
published study of this phenomenon in young children.  However, Warren
and Karrer's (1992, Annals of the New York Academy of Sciences, 425,
489-495) report of brief positive premotor potentials suggests
children's slow potentials may differ from adults.
Both adults (N=20) and 6-year-old children (N=16) were presented
with 30 and 28 trials respectively in which a warning light was
followed after 6 seconds by an interesting, age-appropriate video was
presented as a "go" stimulus for a speeded button press.  Slow brain
potential from Cz and vertical EOG were recorded with a 18 s time
constant as was second-by-second HR.
Adults showed the typical slow negativity gradually increasing
over the 6 s interstimulus interval as well as the typical three
component anticipatory HR pattern.  In contrast, the children's slow
cortical potential was significantly positive during the interval as
demonstrated by a linear seconds effect, F(1,15) = 9.25, p < .01.  The
heart rate pattern was very similar to adults.  EOG was not
consistently correlated with brain potential, being positively
correlated on some slow positive trial blocks, and negatively
correlated on others.  Of the 16 subjects, 11 showed clearly positive
slow potentials, 3 little or no slow change, and 2 negative change.
Careful examination of the ERP to the warning onset verified that lead
polarity was not inadvertently reversed.
The results suggest that cortical dipoles associated with slow
potentials must undergo dramatic changes between the ages of 6 and
adulthood, possibly indicative of major brain reorganization.

Vocal expression of emotion is associated with vocal fold vibration and
vocal tract resonance
Jo-Anne Bachorowski1 and Michael J. Owren2
1Vanderbilt University, 2Reed College

Acoustic properties of speech likely provide external cues about
internal emotions, a phenomenon called "vocal expression of emotion."
Most empirical work in this area has emphasized global measures, such
as pitch and speech rate. In this experiment, associations between
induced positive and negative emotions and more fine-grained acoustic
characteristics were examined.  Twelve subjects were shown 20 slides
selected from the International Affective Picture System to elicit
emotional responses ranging from strongly negative to strongly
positive. Following each slide presentation, subjects provided a
free-form description of the feelings and thoughts evoked by the
picture, preceded by the stock phrase "This test picture..."  At the
end of this narrative, subjects prompted for the next slide by saying
"Next test picture." Acoustic analyses of the frequency and spectral
properties of the /e/ phoneme, drawn from the word "test" in each stock
phrase, showed statistically significant interaction effects involving
emotional valence and trait differences in emotional intensity.
Discriminant function analyses indicated that acoustic cues associated
both with vocal tract resonance and vocal fold vibration rate and
variability are related to discrete emotional states. The effects
observed are of sufficient magnitude to be perceptible, and indicate
that an emotion experienced during vowel production can affect the same
acoustic cues widely held to be used in a listener's speech processing.
These results demonstrate both that acoustic properties of speech carry
information about emotional processes and that individual differences
in intensity of typical emotional responding may be an important
mediator of form of expression.

Psychophysiological responses to emotion-antecedent appraisal of
critical events in a computer game
Rainer Banse1, Alexandre Etter2, Carien van Reekum2, and Klaus R. 
1Humboldt University, Germany, 2University of Geneva, Switzerland

A computer game was used to investigate psychophysiological reactions
to emotion-antecedent appraisal. Three appraisal dimensions were
studied in a factorial design: pertinence and conduciveness of an event
to the major goal in the game, and intrinsic pleasantness of game
events.  Appraisals of goal conduciveness vs. goal obstruction were
operationalized by selecting critical game events in which the player
either succeeds to pass to a higher level or looses a 'life'. The
current level of the player's performance relative to his or her
personal high score was used to operationalize the pertinence of an
event to the current goal. Intrinsic pleasantness was manipulated by
presenting either pleasant or unpleasant sounds marking the critical
game event. 29 subjects practiced the game and then played for 45
minutes. Emotion self report was assessed for a sample of critical
events. Cardiac activity, respiration, skin conductance, skin
temperature, and EMG at the frontalis lateralis, left and right forearm
extensor muscle sites were continuously recorded. The self report data
indicate that appraisals induced the intended emotions: subjects
reported more joy/amusement and pride after goal conducive events, and
more anger after goal obstructive events. The factors conduciveness and
pertinence accounted for more and stronger psychophysiological effects
than intrinsic pleasantness. This result suggests that the pertinence
of an event to goals with high priority may moderate the link between
appraisal and physiological reactions. The specific results are
discussed in the framework of the component-process theory of emotion
(Scherer, 1986).  This research was supported by the Swiss National
Fund for Research

Speed of processing and tested intelligence: A chronopsychophysiological 
Theodore R. Bashore
University of Northern Colorado

Reaction time (RT) studies suggest that faster processing time is
associated with higher levels of tested intelligence.  However, these
studies have not determined if differences in speed are evident at all
or only some stages of processing.  To examine the extent to which
higher levels of intelligence are associated with general or
stage-specific benefits to speeded decision-making, 68 young adults
(mean age=28), grouped according to full scale WAIS IQ scores (105,113,
124, 134), completed a choice RT task in which stimulus
discriminability and S-R compatiblity were manipulated orthogonally.
Stimulus discriminability was varied by requiring subjects to respond
to the word LEFT or RIGHT, embedded in a matrix of number signs (#) or
of letters chosen randomly from the sets A-G or A-Z.  S-R compatibility
was varied by requiring subjects to press a button with either the hand
indicated by the target word or the opposite hand.  ERP activity was
recorded from Fz, Cz, Pz, and Oz.  The choice reaction elicited a
series of components in the ERP, identified as the N60, N160, P200,
N260, and P300.  Differences related to IQ were found only for N260
latency and RT.  These differences indicated that the slowing incurred
when an incompatible response was executed reduced in magnitude as IQ
increased.  This pattern of results suggests that the benefit in
processing speed that accrues to those with higher tested IQs is not
general, but derives from the speed with which they are able to select
a response output.

Frontal P300 decrements, childhood conduct disorder, and the prediction
of relapse among abstinent cocaine abusers
Lance O. Bauer
University of Connecticut School of Medicine

P300 event related brain potentials were studied in 49 cocaine
dependent patients, abstinent for 1-5 months, and 20 healthy,
non-drug-dependent controls.  Patients were assigned to one of two
subgroups based on the presence/absence of a DSM-IIIR diagnosis of
antisocial personality disorder (ASPD).  Analyses of P300s recorded
during a visual discrimination task (rare target, rare nontarget,
frequent nontarget) revealed reduced amplitudes at frontal electrode
sites among patients with ASPD, relative to the ASPD negative patient
and control groups. The frontal P300 decrement was only present in the
rare nontarget ERP.  It correlated significantly with the number of
childhood conduct disorder symptoms, but not with the number of adult
ASPD symptoms.  A secondary analysis examined the relationship between
P300 amplitude among cocaine dependent patients and their future
behavior, i.e., relapse versus continued abstinence.
Discriminant function analysis revealed that P300 amplitude alone
accurately identified 70.6 % of the patients who would later relapse,
and 46.7 % of the patients who did not.

Paternal alcoholism and smooth pursuit eye movement abnormalities in
abstinent cocaine abusers
Lance O. Bauer.
University of Connecticut School of Medicine

The present study evaluated smooth pursuit eye movement (SPEM)
function in 36 cocaine-dependent patients, abstinent for 3 months, and
12 non-drug-dependent normal volunteers.  None of the subjects in
either group
met DSM-IIIR diagnostic criteria for schizophrenia, or delusional,
major affective, or schizotypal personality disorders. None possessed
a history of seizures, significant head injury, or HIV-1 infection.
Forty-eight percent of the cocaine-dependent patients also met criteria
for alcohol dependence.
Subjects were medication-free at the time of testing.  SPEMs were
elicited by a pendulum, oscillated at 0.5 Hz, and recorded using
electro-oculographic techniques.  Tracking accuracy was estimated  by
the power of the horizontal EOG at the stimulus oscillation frequency.
Analyses revealed above normal SPEM tracking performance among patients
with no paternal history of alcoholism, which correlated positively
with years of cocaine use; and subnormal SPEM tracking among patients
with a paternal history of alcoholism.  These differences could not be
explained by other family history, demographic, or drug use variables.

The effect of hypoxia on the preprocessing stage in an auditory oddball
task employing reaction time and P300.
Catherine Beach and Barry Fowler
York University

An Additive Factors Method experiment was conducted to investigate
the slowing effects of hypoxia at the preprocessing stage, using
auditory reaction time (RT) and P300 latency.  Hypoxia was induced with
low oxygen mixtures and controlled by maintaining saturated
oxyhemoglobin at 65%.  Two equiprobable intensities were presented in
an oddball task to 12 subjects, who pressed one of two finger switches
when they heard the tones.  RT and P300 latency increased in response
to both low-intensity tones and hypoxia, resulting in an interactive
effect on RT but an additive effect on P300 latency.  P300 amplitude
was not affected by intensity or hypoxia.  The results from this and
previous visual hypoxia experiments in our laboratory show a consistent
pattern of slowing at the preprocessing stage of information
processing, while central processing stages are unaffected.  Two
explanations are presented to account for this pattern which
contradicts the accepted global impairment view that hypoxia causes a
generalized impairment of cognitive functions.

Frontal EEG activation in 8-month-old infants during a looking version
of the classic A-not-B object permanence task
Martha Ann Bell
Virginia Polytechnic Institute and State University

A looking version of the classic Piagetian object permanence task
(A-not-B) was designed to allow recording of EEG activity during this
marker of sensorimotor development. The reaching movements required by
the infant in the classic version negate use of EEG during the task.  A
large body of EEG work with adults forms the core of our scientific
knowledge of mature brain function during cognition. Knowledge of adult
cognitive processing is not applicable to infancy, however.
Sixty-three 8 month-old infants were assessed on both the reaching
and the looking versions of the A-not-B task, with EEG recorded during
looking. Order of testing was counterbalanced, with a one-minute
baseline EEG recording made just prior to the looking task.  EEG was
recorded from 6 anterior and 4 posterior scalp locations.
Infants in this within-subjects design performed at a higher level
on the looking version of the A-not-B task than on the reaching
version.  This is consistent with recent between-subjects research
using a Delayed Response reaching versus looking paradigm (Hofstadter &
Reznick, 1996, Child Development, 646-658) and may reflect the simpler
response required by looking compared with reaching. There were
activation differences between baseline and task-related EEG recordings
for the frontal (F3/F4) scalp leads, but not for the posterior leads.
Performance on the classic reaching version of A-not-B has been
associated with maturation of the frontal cortical area, measured via
baseline EEG (Bell & Fox, 1992, Child Development, 1142-1163).  These
data suggest that it is possible to study infant brain function during
classic cognitive tasks associated with the sensorimotor period of

Effects of substituting components of nonsignificant stimuli on
reinstatement of the electrodermal orienting responses and
Gershon Ben-Shakhar, Itamar Gati, and Naomi Benbassat
The Hebrew University, Jerusalem

This study examined several aspects of a feature-matching theory (Gati
& Ben-Shakhar, 1990, Journal of Experimental Psychology:  General,
251-263) formulated to account for the roles of stimulus novelty and
significance in psychophysiological orientation processes. The
prediction that stimulus novelty is negatively related to the measure
of common features, shared by the stimulus input and representation of
preceding events, and positively related to the measure of their
distinctive features was tested. A simple habituation paradigm was used
and sequences of nonsignificant verbal (descriptions of people) and
pictorial (schematic faces) compound stimuli were presented to 80
subjects. A test stimulus (TS) was presented after 9 repetitions of a
standard stimulus (SS), followed by 3 additional repetitions of SS. TS
was created by substituting 0, 1, or 2 stimulus components of SS, and
stimulus change was manipulated both within categories (e.g.,
substituting one type of glasses by another type) and between
categories (e.g., substituting glasses by a hat). The dependent measure
was the electrodermal component of the OR to both the TS (OR
reinstatement) and the SS immediately following TS (dishabituation). As
predicted, OR reinstatement was affected by substituting components,
and between-categories substitution was more effective than
within-categories substitution. However, OR magnitude was unaffected by
the number of substituted components, and no dishabituation effects
were obtained. The same pattern of results was obtained regardless of
whether verbal or pictorial stimuli were used.

Affect and unconscious processing: An event-related potential study
Edward Bernat1, Scott Bunce1, Howard Shevrin1, Stephen Hibbard2, and 
Mike Snodgrass1
1University of Michigan, 2Pacific Graduate School of Psychology

This study investigates the relationship between the experience of
positive and negative affect and parameters of the ERP associated with
conscious and unconscious processes.  Specifically, we hypothesized
that: 1) Negative ratings (unpleasant stimuli) would be correlated with
ERP amplitude positivity.  2) Subject ratings of the stimuli would show
some systematic relationship with the subliminal ERPs (Shevrin, et al,
Consciousness and Cognition, 1992, 1, 340-366).   Two sets of data were
analyzed: one set used words as stimuli and the other used schematic
faces (happy/sad).  All stimuli were presented at subliminal and
supraliminal durations.  The data were analyzed using standard
components within a one second ERP window (N1, P2, P3, N4).  The data
using words were analyzed for affect valence using individual subject
ratings on the Osgood evaluative dimension which is made up of five
items.  Both the means and standard deviations among the items were
analyzed.  Osgood means predicted component amplitudes across durations
and electrodes; a lower mean (more negative rating) was associated with
a more positive component amplitude (F=6.3, p<.001).  Amplitude
components P3 and N4 were primarily responsible for the effect (F=11.6,
p<.004; F=10.8, p<.005).  Importantly, the Osgood means effect was also
significant for the subliminal duration alone (F=4.03, p<.006).  We
also found an interaction between the standard deviation and duration
(F=18.2, p<.001).  For the happy/sad faces, ERPs significantly
differentiated the faces in both supraliminal (F=6.4, p<.033) and
subliminal (F=5.3, p<.01) conditions.  Results support the hypothesis
that affect experience can be related to ERP parameters for both
conscious and unconscious processing.

Biofeedback of slow cortical potentials changes brain blood flow in
attentional systems
N. Birbaumer, F. Pulverm|ller, H. Preissl, C. Tempelmann, and H.J. 
University of Tuebingen

Biofeedback of slow cortical potentials from various brain regions was
shown to modify behavior on several tasks (Birbaumer et al 1990).
Clinically it was used successfully to treat seizure disorders
(Rockstroh et al 1989). The neurophysiological mechanisms of cortical
self- regulation are unknown, except that self-produced cortical
negativity improves behavioral and cognitive demands, while positivity
had a disfacilitatory effect. Therefore well-trained healthy subjects
were investigated during self-generated negativity or positivity, while
cortical and sub- cortical blood flow was measured with a 3Tesla fMRI
device. During negativity and positivity highly distinct brain areas
were activated. During negativity bilateral prefrontal lobes, right
dorsomedial frontal regions, basal ganglia and thalamus were activated.
During positivity all the mentioned regions remained silent and
activation of right inferior frontal and occipital regions was
observed. These results are interpreted on the idea that self-produced
negativity involves those brain structures responsible for attentional
resource allocation while self- generated positivity acts through
disfacilitation of attentional networks.
Supported by the German Research Society (DFG).

Interhemispheric transfer: Latency adjusted ERP averaging in normals
and acallosals
Mark D. Bjerke, Gary C. Galbraith, and Warren S. Brown
Travis Institute, Fuller Graduate School of Psychology; Mental 
Retardation Research Group at 
Lanterman Developmental Center, UCLA

	Visual ERPs have been used to study interhemispheric transfer
time in normals and those with callosal dysfunction. Latency
adjusted averaging (LAA)was applied to the data of  normals and
acallosal patients to assess its utility in detecting cross-callosal
signal degradation evident in signal-to-noise ratio (S/N) and latency
	Subjects were 20 normal, two commissurotomy patients, and two
	adolescents with agenesis of the corpus callosum. An ERP
template from each electrode (left and right parietal) and visual field
was obtained by time-locked averaging.  Correlations were computed
between this template and time lagged single-trial ERPs, testing for
point of greatest correlation.  Trials correlating below .50 were
discarded, while the remainder were reaveraged, adjusting for latency.
S/N was the proportion of single trials above .50.  Latency variability
was the standard deviation of single trial latency adjustments.
	For normals, hemisphere-by-visual field ANOVAs for latency of
N1 and P1 from LAA-ERPs showed clear evidence of callosal
transfer delay [N1: p<.001; P1: p<.001].  Direct ERPs had higher S/N
than cross-callosal ERPs [p<.02].  There was a trend for greater
latency variance in cross-callosal ERPs [p=.11].  The four acallosals
showed direct vs. cross-callosal differences for both S/N and variance
that were much greater than differences for the normal group.
	The higher S/N and lower variance for direct responses than for
cross-callosal ERPs in normals suggests that callosal transfer results
in signal degradation. The markedly lower S/N and greater ERP latency
variability in the cross-callosal LAA-ERPs of acallosals supports use
of LAA in the study of callosal pathology.

Differences in startle reactivity as a function of extraversion
Terry D. Blumenthal, Lynda Gioia, and Jennifer T. Scruggs
Wake Forest University

In an attempt to explore information processing differences underlying
the personality construct of extraversion, startle eyeblink reactivity
(periorbital EMG) was measured in introverts and extraverts.  In
Experiment 1 (N = 30), a 60 dB noise prepulse preceded the startle
stimulus on some trials, at an SOA of 120 ms. On control trials,
extraverts showed larger startle amplitude than did introverts, and the
prepulse resulted in more inhibition of startle amplitude for
extraverts than for introverts. In Experiment 2, (N = 42), subjects
were instructed to look at either a wall (no-task), a drawing (visual
task), or a mirror (visual task with self-focused attention).  Startle
amplitude was significantly greater for extraverts than introverts in
the no-task condition, but not in the two task conditions. When
required to perform the tasks, startle reactivity decreased for
extraverts and increased for introverts. These data may be explained by
a valence-matching theory, wherein outwardly directed tasks are
pleasant for extraverts and unpleasant for introverts. The tasks
decreased startle reactivity in extraverts because of a valence
mismatch between the aversive startle stimulus and the appetitive task.
In introverts, the startle stimulus and the external task are both
aversive, resulting in greater startle reactivity during the task than
during the control situation. Alternatively, a personality
trait/attribution explanation would suggest that extraverts are more
reactive to suprathreshold stimuli in an unstructured situation, where
dispositional factors prevail, but that this personality difference is
no longer found during the performance of a task, which is
situationally constrained.

Comparing measures of eyeblink EMG magnitude during startle
Terry D. Blumenthal
Wake Forest University

Several ways of quantifying startle eyeblink magnitude are highly
correlated with each other when the elicitation of the blink reflex is
studied. In an effort to broaden the range of situations in which these
measures can be compared, three measures of blink magnitude where
compared within-subjects under conditions of reflex inhibition or
facilitation. The peak and area of the EMG activity in a window 25-90
ms after startle stimulus onset was measured for both raw rectified EMG
(90-250 Hz passband) and integrated EMG (10 ms time constant).
Correlations between these four measures were calculated
within-subject, then the squared correlation coefficients were averaged
across subjects, and the square root of this average was taken as the
between-measures correlation.  In Experiment 1 (N = 33), blink reflexes
were elicited by noise bursts at 85, 95, or 105 dB(A), preceded on some
trials by noise prestimuli at 60 or 70 dB(A) at a lead interval of 120
ms.  Prestimuli reliably inhibited blink reflexes.  Correlations
between EMG measures were above +.84 in all cases.  In Experiment 2 (N
= 28), blinks were elicited by presenting electrical pulses to the
forehead, paired on some trials with a 70 dB(A) noise burst. Blinks
were inhibited when the noise pulse preceded the electrical stimulus by
120 ms, and facilitated when the noise pulse followed the electrical
stimulus by 25 or 50 ms.  Correlations between EMG measures were above
+.80 in all cases. Therefore, the four measures of eyeblink EMG
magnitude compared here provide similar information during eyeblink
elicitation, inhibition, and facilitation.

A social encounter attenuates the electrically-elicited startle blink
Terry D. Blumenthal1, A. J. W. Boelhouwer2, M. Scott Bovelsky1, Lynda 
Gioia1, and Becky J. 
1Wake Forest University, 2Tilburg University

The purpose of the present study was to evaluate the effect of a social
encounter on the eyeblink reflex elicited by electrical stimulation of
the trigeminal nerve. Previous research has shown that the human
acoustic startle eyeblink reflex is decreased during a social
encounter. The present study measured eyeblinks (periorbital EMG) to
electrical pulses presented to the forehead (supraorbital branch of the
trigeminal nerve). Subjects (N = 32) participated in three testing
session, one week apart. The first two sessions were identical for all
subjects, and the third session was identical to the first two for 16
subjects. For the other 16 subjects, the third session included a
social encounter, during which an observer sat behind the subject.
During all sessions, eyeblinks were elicited by a brief electrical
pulse to the forehead.  A significant interaction was found between
Subject Group and Session for eyeblink magnitude, due to the fact that
the two groups did not differ significantly in the first two sessions,
but blink magnitude was significantly lower in the third session for
the Encounter Group than for the Control Group.  Questionnaires
assessing subject arousal, anxiety, and self-consciousness did not
differ between the two groups, but the Encounter Group reported more
nervousness during the third session than did the Control Group. These
data show that a social encounter causes an attenuation of the eyeblink
reflex, an effect opposite to that found in the presence of negatively
valenced background stimuli.

Prestimulus modality and effects of attention on startle eyeblink
Andreas H. Boehmelt 1, Michael E. Dawson 1, Eric J. Vanman 1, and Anne 
M. Schell 2
1 University of Southern California, 2 Occidental College

Several studies in our laboratory have shown that modification of the
startle eyeblink by lead stimulus with short and long lead intervals is
modulated by stimulus significance. Studies using a discrimination task
that requires subjects to determine the duration of tones of one of two
pitches show that the to-be-attended prepulse generates enhanced
startle prepulse inhibition with a lead interval of 120 ms as well as
pronounced startle facilitation with lead intervals over 2000 ms.
However, two experiments using visual instead of auditory lead stimuli
failed to replicate this effect. Therefore, the present study was
designed to compare tones with simple visual stimuli (slides of bars in
two different spatial orientations) as lead stimuli in a
counterbalanced complete within subjects design, to compare amounts of
short lead interval inhibition and long lead interval facilitation of
auditory startle, generated by lead stimuli presented in the differing
modalities. Attentional modulation of startle was tested by using the
temporal discrimination task. The results show that auditory compared
to visual prestimuli generate more short lead interval inhibition and
more long lead interval startle facilitation. Nevertheless, short lead
interval inhibition to visual prestimuli was found to be significant.
Whereas the attentional manipulation did not yield short lead interval
effects, long lead interval facilitation was significantly enhanced on
to-be-attended trials, independent of sensory modality. The results
suggest early automatic modality effects. Furthermore, the results show
that late effects are both, modality specific reflected by different
levels of facilitation andmodality independent reflected by
non-differential attentional modification of startle.

Increased dimensional EEG complexity indicates a widening of attention
following ACTH.
J. Born, M. Moelle, L. Marshall, and H.L. Fehm
University of Luebeck

An essential aspect of the organism's response to stress is the
release of adrenocorticotropin (ACTH) from the pituitary. Within the
brain, ACTH as a neuropeptide may concurrently influence central
nervous stimulus processing. A dose dependent impairing influence of
ACTH on evoked potential response (ERP) indicators of selective
attention (the Nd) has been observed in several foregoing experiments.
This study examined the effects of the ACTH fragment ACTH 4-10 on the
dimensional complexity in the ongoing EEG activity.  The EEG's
dimensional complexity indexing the brain's mode of stimulus
processing, was evaluated while subjects performed tasks of different
attentional demands. Sixteen healthy men (23-33 years) were tested once
after placebo and another time after administration of ACTH 4-10 (1.2
mg iv., 30 min prior to testing). The EEG was recorded while subjects
were presented with a dichotic listening task. Subjects either (i)
listened to tone pips in both ears (divided attention), or (ii)
selectively to pips in one ear (selective attention) or (iii) ignored
all pips.  Following placebo administration, dimensional complexity of
the EEG activity over the frontal cortex was higher during divided than
selective attention (p<0.01). ACTH significantly increased the EEG
complexity during selective attention, in particular over the
midfrontal cortex (Fz, Cz). This effect was associated with a marked
decrease in the Nd component of the ERP, confirming an impairing effect
of ACTH on selectivity of attention. The increasing effect of ACTH on
the EEG's dimensional complexity which was similar to that of dividing
attention following placebo, points at a widening influence of the
peptide on the attentional focus.

Retinal contributions to photic blink EMG and how to remove them
A.R.Bos, A.J.W.Boelhouwer, M.M.C. van den Berg-Lenssen, & C.H.M. Brunia
Tilburg University

Photic blink EMG can be contaminated with retinalpotentials.
Precautions to avoid this include stimulating only one eye and
recording from the other, highpass filtering, and decreasing electrode
distance.  We evaluated the effectiveness of these precautions.  Muscle
activity was distinguished from retinal activity in three ways:  1. the
signal/noise ratio of retinal potentials benefits from averaging
directly because signal shape is consistent across trials;  muscle
potentials, consisting of random signals, must be rectified before
averaging.  2. If only one eye is stimulated and the other is patched,
retinal potentials will be larger under the stimulated eye but muscle
activity will not differ;  3. If gaze direction is varied while
electrode position and stimulation remain the same, retinal potentials
will vary, but muscle activity will not.  To check the assumptions with
regard to muscle activity, blinks were elicited acoustically as well as
photically.  Subjects turned their head to the left, center and right
of a computer screen, while continually fixating the center.  One of
their eyes was patched.  Photic and acoustic blink reflexes were
elicited by a flasher above the screen and by 95 dB white noise
respectively.  Reflex activity was recorded with interelectrode
distances of 12 and 24 mm.  Data were digitally high-pass filtered with
a -3 dB cut-off frequency of 30, 65 and 100 Hz.  Retinal potentials
could be removed by 30 Hz highpass filtering when recording from a
patched eye, and by 65 Hz high pass filtering when recording from a
directly stimulated eye.  These conclusions were valid for both
electrode distances.

Blood pressure recovery in normotensives and hypertensives under
different stress settings
Wolfram Boucsein1, Reingard Seibt2, Klaus Scheuch2, and Andreas Grass1
1University of Wuppertal, Germany; 2Technical University of Dresden, 

Essential hypertension may develop from repetitive incomplete recovery
from cardiovascular strain. Two blood pressure assessment techniques
were used to compare the recovery of normo- and hypertensives from
traditional mental stressors relative to a computerized task
independently varying task demand and decision latitude (according to
Karasek's model).
The blood pressures of ninety subjects (18-45 years) were recorded
for five days. Hypertensives (non-medicated), borderline hypertensives
and controls (20 each) were tested under both stress settings in
permuted order with an interval of one day. Each stress session
consisted of a baseline, four periods of six minutes stress each with
interspersed three minutes rest breaks, and a 10 minute recovery phase.
Blood pressure was recorded both intermittently from the brachial
artery (Riva-Rocci) and continuously from the finger (Finapres). Stress
values were compared to everyday blood pressure averages obtained from
a separate 24 hour recording.
Systolic blood pressure as recorded by either measure was
significantly higher during the traditional stress setting compared to
the task-demand/decision-latitude setting. No such differences were
found in diastolic blood pressure. Riva-Rocci values were generally
lower than Finapres values. Stress dependent blood pressure elevations
were not significantly different for the groups.  Hypertensives were,
however, characterized by a slower recovery of their pressures compared
to normotensives and borderline hypertensives.
An attempt to predict everyday blood pressures from stress-
induced values was unsuccessful. Baseline values predicted better than
stress values. The percentage of explained variance in everyday
pressure related to stress was slightly higher, however, in
hypertensives compared to normotensives.

A probe for all reasons: Reflex and RT measures in perception
Margaret M. Bradley, Diana Drobes, and Peter J. Lang
University of Florida

	Attentional and emotional processes in picture processing were  
investigated by presenting startling or non-startling acoustic probes at 
different times in a picture viewing interval.  Voluntary RT and 
involuntary blink reflexes were measured. Previous studies found 
that startle probes produce: 1) attentional modulation, evidenced by 
inhibited blink reflexes early in the viewing interval, particularly for 
arousing stimuli, and, 2) affective modulation, evidenced by augmented 
reflexes for unpleasant, compared to pleasant, stimuli, later in the 
interval.  In the current studies, brief, non-startling tone probes or 
startling noise bursts were presented at varying intervals after picture 
onset, as subjects viewed pleasant, neutral, and unpleasant pictures; 
the task was to respond rapidly to the probe with a button press.  
Patterns of responses for voluntary reaction times (RT) to the tone and 
startle probes were compared with each other, and both were compared 
with the temporal pattern found for involuntary blink reflexes.  
	For blink magnitude, the function obtained over time exactly 
replicated earlier studies, even though in the present experiments the 
subjects explicitly attended to the startle probe.  Voluntary RT to 
probes -- either tone or startle stimuli -- mirrored the pattern found 
for involuntary reflexes early in the interval, with significant 
inhibition of responses (i.e. slower RTs), particularly for arousing 
materials.  Later in the interval, unlike the affective modulation found 
for the blink reflex, highly arousing pictures -- both pleasant or 
unpleasant -- speeded probe reaction time.  These data suggest a 
dissociation between involuntary (reflex) and voluntary (RT) probe 
responses late in the perceptual interval, and suggest a method for 
discriminating between attention, activation, and emotion in picture 

fMRI and affective picture processing
Margaret M. Bradley, Peter J. Lang, Bruce N. Cuthbert, Jeffrey R. 
Fitzsimmons, James Scott, Vijay Nangia, and Salvatore Gintoli
University of Florida

Functional maps of processing in the occipital cortex have
demonstrated localized activation for a variety of visual stimuli.
Here, we assessed fMRI in the context of processing pictures that
varied in terms of pleasantness in order to determine 1) whether
photographic pictures produce measurable functional activity and 2)
whether the affective valence of the picture results in different
patterns of activity.
	Male and female subjects first participated in control runs
with lateralized checkerboard stimuli to establish appropriate
contralateral activation.  Then, pleasant, neutral, and unpleasant
pictures were presented in blocks, in which each picture was presented
for 12 s (flashing at a 3 Hz rate), followed by a 12 s interpicture
interval. Eight pictures were presented in each block. Functional
images were acquired using a multislice spiral scan technique on a
conventional 1.5T GE Signa scanner.
	For each condition, a cross-correlation was computed for
activity in each voxel across time using a reference function that
mirrored the 12 s picture ('ON') and 12 s ITI ('OFF') temporal stream.
Functional activity in the occipital cortex was present for all
conditions.  When viewing emotional pictures -- pleasant or unpleasant 
--subjects showed more widespread activation in the occipital area than
for neutral pictures, and, uniquely for these more arousing stimuli, the
area of functional activation extended into the parietal
region.  These data are consistent with the hypothesis that more
interesting stimuli (in this case, emotional) utilize more attentional
resources, resulting in re-entrant activation of primary and secondary
cortical sensory areas.

Blood pressure variability during the workday is buffered by cardiac
autonomic control
E. Brondolo1, J. Stores1, E. Bagiella2, P.A. Shapiro2, R.P. Sloan2
1St. Johns University, 2Columbia University

We have developed a psychophysiological model for the development of
coronary artery disease (CAD) and acute coronary syndromes which
integrates clinical and experimental observations of reactivity to
psychological challenge, behavioral/psychological characteristics such
as work stress and hostility, and autonomic and neuroendocrine
function.  Central to this model is evidence suggesting that autonomic
control of the heart exerts a buffering influence on blood pressure
variability (BPV) responses to challenge.  While laboratory tests of
this buffering effect generally are supportive of the hypothesis,
extension from the lab to the field is essential.  In this pilot study,
we examined the relationship between resting heart rate variability and
blood pressure variability during a working day.
Subjects were 9 New York City Traffic Enforcement Agents.  During
a quiet baseline in a laboratory psychophysiology session, HR and
BP were measured by a Cor Medical model 7000 monitor set to output
6-beat averages every cardiac cycle.  On another day, BP was recorded
every 20 min throughout the working day with an Accutracker ambulatory
BP monitor. As predicted, BP variability (SD of all measurements during
the workday) was inversely related to resting heart rate variability
(SD) (r =3D -.54, p =.14, and r =-.77, p = .01, for SBP and DBP
respectively).  Since recent evidence indicates that BPV is a risk
factor for heart disease independent of mean arterial pressure, these
findings suggest that diminished cardiac buffering of blood pressure
oscillations during the workday may be a mechanism by which
psychological factors contribute to the development of CAD.

Orienting to threat in trait anxiety
Niall Broomfield and Dr. Graham Turpin
University of Sheffield

Previous research demonstrates that individuals with elevated trait
anxiety are characterised by an attentional bias for threat. The
current study seeks to extend this research by identifying which
components of visual attention are sensitive to threat cues using
Posners paradigm.
Subjects were screened and allocated to one of three groups: high
trait anxious (N=20), low trait anxious (N=20) and repressors
(N=20). Each participant completed 96 reaction time (RT) trials on a
modified Posner cue-target task. 16 threatening and 16 neutral words
served as cues.  Subjects were instructed to maintain fixation at a
central point and respond to a target appearing in either the cued or
uncued hemifield. 2/3 of the trials were validly cued, 1/3 invalidly
cued. Cue-target intervals (CTI) were 100 ms (short) or 500 ms (long).
Inter-trial intervals were randomised between 25 and 35 seconds.
Electroculographic, electrodermal and heart rate activity was recorded
using a Biopac system.
Results indicated that all subjects displayed slowed reactions to
targets invalidly cued by threat words on the long CTI (p < 0.05).
Subjects demonstrated enhanced detection of targets on invalidly cued
trials (p < 0.0005). High trait anxious subjects appeared to
demonstrate more uninstructed eye movements towards threat cues (p <
0.005). Skin conductance analysis indicated a word by trial effect (p <
Results are consistent with the notion that emotional stimuli
attract greater attentional resources but suggest that trait
anxiety may only modulate bias for threat effects at the overt level.

Effects of gender and comorbidity on regional brain asymmetries in
major depression
Gerard Bruder, Regan Fong, Craig Tenke, Paul Leite, James Towey, 
Jonathan Stewart, and Frederic 
New York State Psychiatric Institute

Gender differences have been found in the incidence of depression and
in lateralized cognitive processing (Heller, 1993, J.  Affect. Disord.,
129-143). This study examined gender differences in regional
asymmetries of the electroencephalograph (EEG) in depressive patients
with and without an anxiety disorder. Resting EEG (eyes open and
closed) was recorded from 44 unmedicated outpatients having a major
depressive disorder (MDD) and 26 normal controls (13 females, 13
males).  Patients were subgrouped by comorbidity, with 19 having both a
MDD and an anxiety disorder (9 females, 10 males), and 25 having a MDD
only (13 females, 12 males). EEG was recorded from 30 scalp electrodes
(13 homologous pairs over the two hemispheres and 4 midline sites)
using a nose reference. Differences in regional alpha asymmetries among
the anxious depressed, nonanxious depressed and control groups were
greater for females than males (Gender x Group x Region x Hemisphere
interaction; p<.01). Depressed women (with or without an anxiety
disorder) showed evidence of greater activation (less alpha) over the
right than left anterior region, whereas normal controls did not
(p<.01). In the posterior region, nonanxious depressed women showed
less activation (greater alpha) over the right than left hemisphere,
but anxious depressed women had the opposite asymmetry, indicative of
greater right posterior activation (p<.0005). There was no significant
difference among groups in the alpha asymmetry for men. These data
indicate that both gender and comorbidity are important in
understanding regional brain asymmetries in depression.

Stimulus preceding negativity prior to auditory and visual verbal and
non-verbal knowledge of results stimuli.
Tilburg University

Movement Preceding Negativity (MPN) can be separated from Stimulus
Preceding Negativity (SPN) using a time estimation task.  Subjects had
to press a button 3s after a warning stimulus and were informed about
their performance 2s after the movement. A MPN was found prior to the
button press, and an SPN was found preceding the Knowledge of Results
(KR) stimulus. In a number of experiments with different co-workers it
was demonstrated that the SPN only shows up if the stimulus is a KR
stimulus and only if the KR information is real.  In general a right
hemisphere preponderance was found. In the present experiment we
challenged the right hemisphere preponderance by presenting also verbal
KR stimuli to a group of 20 subjects, both in the visual and in the
auditory modality. Our common KR stimuli were used as well. We tested
the following hypotheses: 1) The potential distribution of the SPN in
this experiment will be affected by a contribution from the primary
projection areas. 2) The right hemisphere preponderance of the SPN is
nevertheless independent on modality. 3) Assuming that the effects of
KR are brought about by a specific network, it should only partly be
vulnerable to the use of verbal KR stimuli. Therefore the right
hemisphere preponderance should at least partly remain intact. SPN was
recorded in a 200ms epoch prior to the presentation of the KR stimulus.
ANOVAs were carried out (Geisser/Greenhouse) with Factors: Electrode
Position (P:F3,F4,C3,C4, P3,P4,O1,O2,T3'and T4'), Hemisphere (H:left/
right), Extremity (E:  left/right), Language (L: verbal/non- verbal)
and Modality (M:  visual/auditory). A significant main effect was found
for Electrode Position and Hemisphere (P: df 4,76, F 4.93; H: df 1,19
F: 10.83). The right hemisphere preponderance of the SPN remained

When positive becomes negative: ERP evidence for differential
processing of affective stimuli in subjects with parental loss.
Scott C. Bunce1, Edward Bernat1, Stephen Hibbard2, and Howard Shevrin1
1University of Michigan, 2Pacific Graduate School of Psychology

Traumatized individuals report a paucity of positive excitement and
intimate moods (PEI) relative to non-traumatized individuals (Bunce et
al., 1995; J. Personality, 63(2) 165-188).  This study examines ERPs
from PEI word stimuli in individuals with early parental loss.
	Psychodynamic and attachment theories suggest early parental
	loss can produce conflict over intimate/excited feelings.
Thus, positive moods were hypothesized to produce more negative
response in people with loss than those without.  Negative affect has
been associated with relatively greater right versus left frontal
activation (Davidson, 1993; Cognition & Emotion, 7(1) 115-138).  We
predicted 1) differential P2 and P3 amplitudes to PEI words, 2) people
with loss would exhibit greater right versus left frontal amplitudes
relative to controls, and 3) an interaction with subliminal versus
supraliminal presentation.
	Seven loss subjects reported less PEI over 30 days than 9
controls.  32 words, representing four mood factors (PEI, Depression,
Anxiety, Calm), were presented at both subliminal (1 ms) and
supraliminal (40 ms) durations.  Results indicated that for PEI, loss
subjects showed greater right (F4) versus left (F3) amplitudes at both
P2 and P3, regardless of duration, relative to controls.  Supraliminal
findings at both P2 and P3 indicate the effect is dependent on both the
relative right frontal amplitude of the loss group and the relative
left amplitude of controls.  Subliminal findings at P2 show a
significant effect only for loss, again showing greater right versus
left amplitude.  These results suggest positive/intimate words may
evoke more negative responses in individuals with parental loss both
consciously and unconsciously.

Event-related desynchronisation (ERD) of the EEG during recognition
memory for words and designs
Adrian P. Burgess and John H. Gruzelier
Charing Cross and Westminster Medical School

Although changes in the EEG which are phase-locked to an event (i.e.
ERPs) have been widely used to study human episodic memory, relatively
little work has investigated non-phase locked changes (i.e.  ERD). In
this study ERD was recorded during two continuous recognition memory
tasks with the aim of determining whether there was evidence for a
repetition effect comparable to that seen using ERPs. EEG was recorded
at 28 scalp sites from 15 healthy right handed subjects while they
performed a continuous recognition memory task for words and faces. ERD
was calculated on individually defined lower and upper alpha ranges.
ERPs were calculated using the same data and the results were compared.
Analysis was by repeated measures ANOVA with the Greenhouse-Geisser
correction for non- spericity. There was a significant ERD repetition
effect in upper alpha for both memory tasks. Desynchronisation was
greater for repeated stimuli than for new ones from 500ms after
stimulus presentation, with nearly a 40% reduction in power compared
with the baseline period (Words: p<0.001); Designs: p<0.03). These
effects remained significant when the data were reanalysed co-varying
for the ERP. The results suggest that there are non- phase locked
changes in the EEG associated with memory processes that are
independent of the ERP. This means that ERD may prove to be a useful
method of investigating the time course of memory-related cortical
processes that will supplement conventional ERP analyses.

Long-term estrogen replacement therapy: Effects on autonomic and immune
stress reactivity
Mary H. Burleson1, John T. Cacioppo2, Kirsten M. Poehlmann2, Gary 
Berntson2, Janice K. Kiecolt-
Glaser1, Ronald Glaser1, and William B. Malarkey1
1Ohio State University College of Medicine, 2Ohio State University, 

Previous research has demonstrated that postmenopausal women show
higher autonomic and neuroendocrine reactivity to brief laboratory
stressors than premenopausal women, and that short-term treatment with
estradiol also lowers reactivity.  Estrogens and progestins also can
impact baseline levels of neuroendocrine and immune function.  Taken
together, these results support a role for endogenous gonadal steroids
in regulating stress reactivity, and suggest pathways through which
estrogen replacement therapy (ERT) may impact health.  The current
study investigated the effects of chronic ERT on baseline function and
on brief stress reactivity.  Postmenopausal women (aged 50-80 yrs) who
were long-term (more than 2 yrs) users of conjugated estrogens alone or
in combination with medroxyprogesterone (n=30), along with 25 controls,
performed 6-min speech and 6-min math tasks.  Indicators of autonomic,
neuroendocrine, and immune function were assessed at baseline and
immediately post-stress.  Heart rate, epinephrine, adrenocorticotropic
hormone, and cortisol increased, and respiratory sinus arrhythmia (RSA)
decreased after the tasks, indicating that they elicited a mild stress
response.  Women using ERT had greater tachycardia and a larger RSA
decrease in response to the stressors than did the controls, suggesting
that ERT may delay age-related declines in parasympathetic
responsiveness.  Across both measurement periods, mitogen-stimulated
lymphocyte blastogenesis was higher in the ERT group than the
controls.  In addition, the chronic ERT group did not show a
stress-related decrease in mitogen-stimulated blastogenesis, whereas
the controls showed a decrease similar to that found in previous
studies.  These results suggest that ERT may upregulate cellular immune
function both tonically and after stress.

Effects of triazolam on sleep and inertia
J. Lynn Caldwell and John A. Caldwell
U.S. Army Aeromedical Research Laboratory

Ten male U.S. Army pilots were tested to determine triazolam's effect
on sleep architecture and wake-up times.  Triazolam (0.25 mg) or
placebo was administered on 4 separate nights in a counter-balanced
design.  A control day separated each drug day.  During one triazolam
and one placebo night, subjects were awakened at midnight to perform a
simulator flight task and then returned to bed to complete the sleep
period.  A 2 (drug) X 2 (wake-up condition) repeated measures ANOVA
indicated that during the triazolam nights, subjects showed less stage
1 sleep (F(1,9)=23.82, p=.0009), more stage 2 (F(1,9)=17.13), less
stage 3 (F(1,9)=5.74, p=.0060), and less awake time after sleep onset
(F(1,9)=12.75, p=.0004) compared to the placebo nights.  Analysis of
the morning wakeup responses indicated a slower reaction time after
triazolam than after placebo (7.7 seconds vs 5.93 seconds, p=.08).
Midnight wakeup responses also were slower after triazolam than after
placebo (53.80 seconds vs 8.4 seconds), but  the high variability under
triazolam (110.52) led to a nonsignificant F value.  Two of the
subjects were unable to remember details of the midnight flight on the
morning after triazolam.  These results indicate that people who are
considering triazolam as a sleep aid on nights during which early
awakening is possible must be aware that the time to awaken is slow and
the details of the night activity may not be remembered.

A factor-analytic study of performance, mood, and EEG in sleep-deprived
John A. Caldwell, Jr. and J. Lynn Caldwell
US Army Aeromedical Research Laboratory

Factor analysis was conducted on flight performance scores, subjective
mood ratings, and electroencephalographic (EEG) data  from a study
using dextroamphetamine to alleviate fatigue.  The flight scores were
collected during "flights" in a helicopter simulator, the mood ratings
were from the Profile of Mood States, and the EEG data consisted of
delta, theta, alpha, and beta power recorded from Cz with eyes closed.
A principal component analysis with varimax rotation extracted a
4-factor solution which explained 73% of the variance in the 16
original variables.  Using a factor loading cutoff of 0.6, the 6
performance variables loaded significantly on factor 1, the 6 mood
variables loaded on factor 2, and the 4 EEG variables were divided
equally between factors 3 (slow EEG) and 4 (fast EEG).  These results
indicate each type of measure (performance, mood, and EEG) contributed
uniquely to an understanding of the effects of fatigue and a
stimulant.  Analysis of variance (ANOVA) on the resultant factor scores
revealed significant effects on performance and mood scores due to drug
(placebo versus dextroamphetamine), session (0100, 0500, 0900, 1300,
and 1700), and the interaction between drug and session.  There were
drug effects on EEG factor scores as well, but no session effects, and
only 1 drug-by-session interaction (for fast EEG).  With minor
exceptions, conclusions drawn from the ANOVA on the 4 factor scores
were consistent with those based on the 16 original variables.  These
findings suggest it may be possible to construct global indices to more
efficiently analyze and interpret complex data sets of this type.

Alcohol dependence and related disorders in subjects with  small,
average, and large P300 event-related potential amplitude.
Scott R. Carlson, Amy K. Mertz, Joanna Katsanis, and William G. Iacono
University of Minnesota

	The P300 component of the event-related brain potential (ERP)
may serve as an  index of risk for developing alcoholism.  Children of
alcoholics have been found to have diminished P300 amplitude relative
to children of normal controls.  In the present study, we selected  95
16 - 18 year old subjects (from a total sample of 443) based on their
having small, average, or large P300 waves and examined them for the
presence of psychopathology.
	Visually evoked ERPs were recorded from the PZ scalp site using
the "rotated heads" task of Begleiter et al. (1994, Science,
225, 1493-1496).  Subjects whose P300 amplitudes lied at the top 7%
(n=29) of the distibution comprised the large amplitude group and those
with amplitudes at the bottom 7% (n=31) of the distribution the small
group.  An average group was also formed by selecting subjects whose
P300 amplitude was at the mean of the distribution (n=35).  Subjects
were assessed for the presence of externalizing (i.e. antisocial
behavior, antisocial personality disorder, conduct disorder,
oppositional defiant disorder), internalizing (i.e., depression), and
substance use (alcohol abuse/dependence, nicotine dependence, and
illicit drug abuse/ dependence) disorders using structured interviews.
	The small and average amplitude groups were found to have more
alcohol dependence diagnoses than the large amplitude subjects.  No
significant differences in the presence of psychopathology were found
between the small and average groups, and there were no significant
group differences for the other disorders.  These findings suggest that
increased P300 amplitude might serve as a protective factor against the
development of alcoholism.

The beneficial and negative influences of marital quality on immune
Sybil Carrere, John M. Gottman, and Hans D. Ochs
University of Washington

This study examined the impact of marital quality on
immunocompetence.  Forty-eight newlywed couples participated in the
study.  Each couple spent 24 hours in an apartment laboratory
interacting as they would in their own home.  Couples were interviewed
about the history of their relationship (Oral History Interview).
Three dimensions of the interview (cohesion, gender role stereotypy,
and volatility) were related to immunocompetence.  Greater marital
cohesion was associated with enhanced immune function for the wives
(higher blastogenic response to 2 mitogens).  Women with more
traditional gender roles in their marriage showed suppresed immune
function (reduced blastogenic response to 3 mitogens and higher
antibody titers to latent Epstein-Barr virus).  Women in emotionally
volatile relationships had greater antibody titers to latent
Epstein-Barr virus.  However, men in volatile relationships had
enhanced immune function (higher natural killer cell activity).
Emotional communication patterns during marital conflict were also
linked to immunocompetence.  Husbands' use of positive emotions (e.g.,
affection, validation, humor) was associated with wives' greater
lymphocyte proliferative response to 2 mitogens.  Wives who displayed
positive emotions had a higher response rate to 2 mitogen challenges.
Men exhibiting negative emotions (anger, contempt, and belligerence)
had higher levels of antibody titers to Epstein-Barr virus.  The wives'
negative emotional communication was associated with the husbands'
lower blastogenic response to 2 mitogen challenges and the wives' own
reduced response to 3 mitogens.  These data provide additional support
for the link between marital discord and immunosuppression.  The data
also indicate that positive qualities of marriage, not just the absence
of marital discord, are linked to enhanced immune function.

Blood pressure reactions to the cold pressor test and future blood
pressure status
Douglas Carroll1, George Davey Smith2, David Sheffield3, Gonneke H.M. 
Willemsen1, Peter N. 
Sweetnam4, and Peter C. Elwood4
1University of Birmingham, 2University of Bristol, 3University of North 
Carolina, 4MRC 
Epidemiology Unit

Following determination at an initial screening session, blood
pressure was recorded at rest and in reaction to the cold pressor
test.  Follow-up screening blood pressure was determined five years
later.  The effective sample was 1059 men, with an average age of 56.5
years at initial screening.  Reactions to the cold pressor test
averaged 24.1mmHg systolic and 12.2mmHg diastolic pressure. Step-wise
multiple regression, with age always entered first, indicated that
blood pressure reactions to the cold pressor test provided minimal
independent prediction of follow-up blood pressure over and above that
provided by blood pressure at initial screening.  In the case of
systolic pressure, a model including only age and initial screening
blood pressure accounted for 38 per cent of the variance in follow-up
screening blood pressure; systolic blood pressure reactions to the cold
pressor test did not enter the equation.  In the case of follow-up
diastolic pressure, diastolic blood pressure at initial screening
accounted for 22 per cent of the variance, and, while diastolic blood
pressure reaction to the cold pressor test entered the equation, it
accounted for only an additional one per cent of the variance.  These
results raise doubts, certainly for older populations, about whether
blood pressure reactions to the cold pressor test provide a useful
clinical index of the course of future blood pressure.

Information processing during slow wave sleep
Elisha A. Chambers, Alexandra J. Ostaniewicz, Sidney J. Segalowitz, 
Robert D. Ogilvie, and Sharon 
A. Mercier
Brock University

Previous work has suggested that the late components of the auditory
ERP which are usually interpreted as indications of effortful
information processing are nonetheless present during various sleep
stages, including Slow Wave Sleep (SWS). However, few studies have
examined the differentiation of target from nontarget stimuli during
SWS. In an investigation of the presence and implications of high
versus low GSR states during SWS, we presented an ongoing sequence of
rare high tones and frequent low tones to subjects once they were
asleep, and divided their SWS into episodes of high versus low GSR. The
auditory ERPs to the target tones produced a recognizable negativity at
approximately 200 ms (-3 to -4 uV), a positivity at 400-600 ms (4-6
uV), and a positivity at 1000-1200 ms (4-6 uV). The lateness of these
components are expected during states of very low arousal such as SWS.
Targets (T) and nontargets (NT) were differentiated by the amplitude of
the positive components (p=.004), but not the earlier negative
component (interaction: p=.003) with NTs having lower amplitude (4.91
vs 2.88 uV). There was also a stimulus x component latency interaction
such that the NT slightly preceded the T at the early positivity by 25
ms, but followed the T at the later positivity by 125 ms (p=.006).
There were no reliable effects of high versus low GSR.  We have
demonstrated that there is stimulus differentiation in the ERP during
SWS in components akin to the P2 and P3 but not N1, indicating that
consciousness is not necessary for producing electrophysiological
indicies of information processing.

Self-efficacy as a predictor of salivary immunoglobulin A concentration
changes under examination-related stress.
Charles C. Chan1 and John A. Spinks2
1Belmont Private Hospital, 2The University of Hong Kong

The relationship between examination-related stress and parameters of
the immune system has been the focus of many naturalistic
psychoimmunological studies over the past decade.  Recent studies by
the present authors have shown that this relationship may be predicted
by parameters of sympathetic activity, such as skin conductance
response derived measures.  Several previous human
psychoneuroimmunological studies have reported suppressed salivary
immunoglobulin A (S-IgA) concentration levels in college student
subjects on the actual examination dates.  When conflicting results
were found, researchers have pointed to differences in saliva sample
collection and assay methods, or simply stated that it was difficult to
understand.  A recent prospective field study has been able to build in
strict controls of the subjects' cognitive demands and attentional
conditions, oral cavity environments, as well as uniformity in saliva
sample collection time and assay methods.  The study hypothesized,
based on a social-cognitive view of examination preparation as a
process of self-regulated learning, that as college students' academic
self-efficacy naturally increased approaching the actual examination
dates, there should be a parallel trend in S-IgA concentration levels.
Results supported the above assertion.  The study also incorporated an
experimental intervention to modify subjects' academic self-efficacy.
This is one of the few studies in this area which would allow for
cause-effect inference.  Although the results indicated that the
intervention program had significant effects on S-IgA concentration
levels, the interpretation of these changes require further

Distance effects in semantic memory as assessed by event-related
potentials: Effects of instruction
Dorothee J. Chwilla, Herman H.J. Kolk, and Patrick J.W. Oor
University of Nijmegen

Distance in semantic memory refers to the number of associative steps
between words.  In case of a mediated relationship words are not
directly but indirectly related via one intervening word.  RT studies
demonstrated Mediated Priming Effects (MPEs) in lexical decision.
Recently we observed a small but significant ERP MPE in the absence of
RT effects.
This study investigated the reliability of this ERP effect, and
assessed the effects of instruction. Previous research suggests that
informing subjects about the presence of indirectly related wordpairs
and instructing them to use such relationships helps to find RT MPEs.
EEG was recorded while 25 subjects performed a visual lexical decision
on target letter strings. The first list included mediated wordpairs
only, the second list also included directly related wordpairs. To
study order effects the mediated list was divided into two blocks.
Instruction, presentation order, and list composition affected the
ERPs.  Instruction enhanced the ERP MPE and yielded an early distance
effect that was absent in our previous study.  In the first block of
the mediated list an early negativity (200-350 ms post-target) was
larger for unrelated than mediated wordpairs. In the second block an
ERP MPE was observed (an increase in negativity within 350-500 ms for
unrelated than mediated wordpairs). No ERP MPE was found for the list
including directly related wordpairs. For RT no MPEs were observed.
This study confirms our finding of ERP MPEs in the absence of RT
MPEs. The implications for models of spreading of activation and
integrative mechanisms will be discussed.

Modifications of electrogastric activity during the viewing of film
Maurizio Codispoti, Giovanni Tuozzi, Roberto Bolzani, Bruno Baldaro, 
Daniela Palomba, Marco W. 
Battacchi, and Giancarlo Trombini
University of Bologna

There are few studies on electrogastric activity modifications during
film sequences.  The purpose of this exploratory study was to compare
the electrogastrographic activity (EGG) during  two different film
sequences: a Disney cartoon scene and a surgery scene that depicts the
phases of a thoracic operation. Previous studies have shown that the
3-cpm frequency of the EGG reflects the frequency of pacesetter
potentials occurring at the serosal and mucosal surfaces of the
stomach.  In many situations, increases in the amplitude of the EGG are
indicators of increased gastric contractile activity (Stern, Koch,&
Twenty-four female students viewed the movie scenes while
electrogastrographic activity and heart rate were recorded.  The
physiological measures were recorded for 10 minutes before and 10
minutes during the projection of each film sequence. EGG signals were
filtered using a moving average digital filter; a Fourier analysis was
then performed for each recording period. The results showed that EGG 3
cpm wave amplitude increased significantly during the viewing of the
surgery sequence whereas there was a decrease during the cartoon

Probing the mind's eye: Reflex modulation for briefly presented
Maurizio Codispoti, Margaret M. Bradley, and Peter J. Lang
University Of Florida

      In a 6 s picture viewing period, startle reflexes elicited at
different times after picture onset are differentially modified by
affective valence early and late in the viewing interval. In addition,
blinks elicited by early probes are greatly inhibited, presumably due
to the 'prepulse' effect of picture onset. In this study, we further
investigated the mechanism of prepulse inhibition and affective 
modulation by presenting pictures for 
a brief, 500 ms period.  Two questions were asked: 1) Would the 
inhibition function differ in form, 
given the absence of a perceptual foreground, and 2) Would late effects 
of affective valence appear in 
the absence of a picture foreground?
      Acoustic startle probes were presented 300, 800, 1300, 1800, 2800,
4800 ms following the onset of a 500 ms presentation of pleasant,
neutral,and unpleasant pictures. Heart rate, skin conductance, and
corrugator EMG activity were also recorded.
      The temporal function relating blink magnitude to picture valence
was practically identical to that obtained in previous studies with a 6
s picture presentation: Blinks were inhibited early in the interval,
and more so for arousing pictures.  Later in the interval (> 800 ms),
blink responses were potentiated for unpleasant pictures, compared to
pleasant stimuli. Responses in skin conductance and corrugator EMG
systems were also identical to those obtained with a 6 s viewing
interval.  These data suggest that, in the absence of a perceptual
mask, picture processing follows the same time course regardless of
whether the visual stimulus is present or absent, and that startle
reflex modulation indexes the affective quality of a persisting mental

Effects of posture on autonomic reactivity to psychological stress in
Beth Colaluca1, Kathleen Soderlund1, and Robert M. Kelsey2
1University of North Texas, 2State University of New York at Stony Brook

The effects of posture on cardiovascular and electrodermal reactivity
to psychological stress were evaluated in 28 undergraduate women during
two 5-min mental arithmetic tasks.  Subjects performed one task while
sitting and the other task while standing, with each task preceded by a
5-min baseline period in the same posture.  Half of the subjects
received the sitting baseline and task periods first, and the other
half received the standing baseline and task periods first.
Cardiovascular measures included heart period (HP), preejection period
(PEP), and total peripheral resistance (TPR).  Skin conductance
responses (SCRs) served as a separate index of sympathetic activation.
During the baseline periods, standing was associated with shorter
HP, longer PEP, and more SCRs as compared to sitting (p<.05).  Baseline
TPR did not vary as a function of posture.  There was a significant
multivariate effect of posture on cardiovascular reactivity (changes
from baseline) during the tasks, due to greater decreases in HP while
sitting and greater increases in TPR while standing (p<.01).  Decreases
in PEP and increases in SCRs during stress were significant (p<.005),
but did not vary as a function of posture.  Likewise, posture did not
affect patterns of reactivity over task minutes.  Correlations between
sitting and standing baseline cardiovascular measures exceeded 0.80,
but correlations between sitting and standing reactivity measures
exceeded 0.60 only for HP and PEP.  Results indicate moderate
consistency of chronotropic and inotropic cardiac reactivity to
psychological stress across postures, and little impact of posture on
relatively pure indices of sympathetic reactivity to stress (PEP,

ERP measures of short-term memory scanning
Edward M. Conley and Arnold Starr
University of California, Irvine

Auditory event-related potentials (ERPs) were studied during the
scanning of short-term memory stores.  Subjects were presented lists of
digits and asked to perform two different processes: a digit
memorization (MEM) task or a digit classification (ODD/EVEN) task.  In
both tasks, lists of 1, 3, 5, or 7 digits (MEM: digits 0-12; ODD/EVEN:
digits 1-12) were presented, followed by a probe digit.  In the MEM
task, subjects made an appropriate button press whether the probe was
or was not a member of the preceding list.  In the ODD/EVEN task,
subjects decided whether the probe was odd or even.  Accuracy decreased
and reaction times (RTs) increased in the MEM task as list size
increased.  In the ODD/EVEN task, accuracy and RTs did not change with
list size.  Three ERP components recorded after probe presentation were
affected by the two tasks.  An early negative component, N180,
decreased in amplitude as list size increased in the MEM task, but did
not change with list size in the ODD/EVEN task.  A lateralized
negativity (N485), maximal at C3, and a later positivity (P800),
maximal at Pz, were affected by list size with some measures
(amplitude, latency, and duration) distinguishing between the two
tasks.  The results suggest that early and late brain events after a
probe in a memory scanning task have certain attributes that can be
related to the scanning of short-term memory stores.

Assessment of cognitive functioning using event-related potentials.
John F. Connolly
Dalhousie University

Previous work has demonstrated the feasibility of adapting a
neuropsychological test, Peabody Picture Vocabulary Test-Revised
(PPVT-R), for computerized presentation and ERP recording (Connolly et
al.,1995, J. Clin. Exp. Neuro.,17,548).  More recent work will be
presented in which ERPs were recorded to two subtests of the Wechsler
Adult Intelligence Scale-Revised as a Neurological Instrument
(WAIS-R-NI); Vocabulary and Similarities. In the Vocabulary test each
target word had five possible definitions: correct, incorrect-related,
phonetically similar-incorrect, semantically related-incorrect, and
incorrect.  ERPs to the correct definition differed from those to all
other definitions by having P300 responses which occurred in the 350-
400 ms range. This ERP-definition effect was observed regardless of
whether behavioural responses were required to the definitions. In the
Similarities test each target pair had four possible similar pairs:
most correct, partially correct, and two levels of incorrect. ERPs to
the most correct pair differed from those to all other categories by
having P300 responses which occurred in the 450-500 ms range and had a
strong parietal distribution. These effects were seen for individual
subjects not just groups. Results will be compared to those with the
PPVT-R in which correct responses were characterised more by
negative-going components (Connolly & Phillips, 1994, J. Cog.
Neuro.,6,256). The importance of these results to the clinical
application of cognitive ERPs will be discussed with particular
reference to the assessment of cognitive functioning in individuals who
are impossible or difficult to test (e.g., global aphasic patients)
using traditional methods.

Gender differences in heart rate variability in 60-Hz magnetic fields
Mary R. Cook, Charles Graham, Antonio Sastre, Steven J. Hoffman, and 
Mary M. Gerkovich
Midwest Research Institute

Quantitative spectral analysis of HRV provides a useful in vivo
indicator of beat-to-beat variations in sympathetic and parasympathetic
nerve activity, and has clinical prognostic value for cardiovascular
risk.  In two previous double-blind, laboratory-based studies involving
a total of 52 healthy young men, we demonstrated that compared to sham
exposure, all-night exposure to an intermittent magnetic field (60Hz,
200 mG) decreases spectral power in the "low" band (0.0-0.1Hz; p =
.017) and increases power in the respiratory arrhythmia band (0,15-0.4
Hz; p = .008).  More recently, we performed a third study under similar
test conditions to determine if continuous rather than intermittent
field exposure, had differential effects on HRV.  Spectral measures of
HRV during continuous exposure did not differ from sham control
conditions. We have now completed a double-blind study of intermittent
magnetic field exposure effects on HRV in healthy young women.  Each
woman slept in our laboratory two nights. On one night the woman was
exposed to the same intermittent magnetic field exposure conditions
used previously; the other night was sham exposure.  Cardiac interbeat
interval was recorded continuously through the night for HRV analysis.
HRV patterns over the night were different in women than in men, even
under the sham exposure condition.  The implications of the complex
gender differences observed during both sham and magnetic field
exposure will be illustrated and discussed.

Specificity of startle modulation revisited: Relationships of affective
and prepulse modification to fearfulness and schizotypy
Edwin W. Cook III, Darin W. Goates, Larry W. Hawk, and Andrew D. 
University of Alabama at Birmingham

Previous research has demonstrated relationships between startle
modulation and individual differences related to psychopathology.  For
example, highly fearful subjects typically show startle modulation by
emotional valence whereas low fear subjects, as well as psychopaths, do
not.  Persons scoring high on the Chapmans' schizotypy scales also show
enhanced affective modulation of startle, along with deficits in
prepulse inhibition and habituation.  The present study further
investigated the specificity of relationships of these startle
modulation phenomena to fearfulness and schizotypy.  Undergraduates
(N=64) were selected based on questionnaire scores and viewed pleasant,
neutral, and aversive pictures.  Acoustic startle probes were presented
during most of the pictures, and half the probes were preceded by a
prepulse (120 ms SOA, 8 dB above background).  As in prior research,
high fear compared to low fear subjects showed greater startle
potentiation during aversive pictures.  Low fear subjects showed a
pattern previously reported for psychopaths: smaller startles during
both aversive and pleasant pictures compared to neutral pictures.
Prepulse inhibition was reduced in high schizotypal subjects, but only
during the pleasant pictures, and high fear subjects also showed a
reliable reduction in PPI.  Although the present findings provide
additional evidence for covariation between startle modulation and
dimensions of psychopathology, they suggest caution in associating
specific modulation deficits and excesses with specific behavioral

Changes in event-related potentials during the transition from
wakefulness to sleep
Kimberly A. Cote, Andrea Perrino, Duncan R. de Lugt, and Kenneth B. 
University of Ottawa

The present study investigated changes in late ERP components as a
function of changing EEG arousal level and behavioural responsivity
during the transition from wakefulnessto sleep. Ten subjects, aged
18-32 were presented with an auditory oddball task during repeated
sleep onset periods. Standard 1000 Hz tone pips were presented on 96%
of trials, and 2000 Hztargets on 4% of trials. Subjects were required
to button press upon detection of the target. Trials were sorted by
sleep stage and by reaction time (RT), in which bins 1-3 represented
increasing time to respond and bin 4 represented a failure to respond.
During wakefulness, misses were rare.  A large amplitude P3 was
observed following target presentation. During stage 1, subjects
responded on approximately 50% of trials. P3 amplitude remained large
to these detected targets.  There is thus good evidence of conscious
awareness during stage 1. P3 latency was relatively constant, while RT
varied considerably. Slowed RTs therefore appear to be due to response
production, rather than stimulus classification processes. The P3 was
not apparent when targets were not detected. The ERPs to the 19
standard stimuli prior to the target were averaged in an attempt to
predict behavioural responsiveness. N1 amplitude gradually declined as
RT increased during stage 1. It was at or near basline during misses.
N1 therefore appears to closely monitor the gradual loss of
consciousness, as reflected by increasing RT and the eventual failure
to respond. The sleep onset period is marked by a gradual decrease in
N1 amplitude, increasing RTs, and loss of P3 when target is not

The effects of odor intensity on the olfactory event-related potential
in young and older adults.
James W. Covington1, Mark W. Geisler1,2, Charlie D. Morgan3, and Claire 
1San Diego State University, 2University of California School of 
Medicine, San Diego, 
3SDSU/UCSD Joint Doctoral Program in Clinical Psychology

Event-related brain potentials have been utilized for decades in the
study of auditory, visual, and somatosensory functioning.  Olfaction
has been neglected for the most part due to the difficulties involved
in recording true olfactory event-related potentials (OERP).  However,
improvements in recording techniques have made the OERP a reliable
method of assessing the psychophysiology of olfaction.  Employing a
single stimulus paradigm, the present study investigated the effects of
odor intensity and age on the OERP.  Recordings of OERPs were obtained
monopolarly at the Fz, Cz, and Pz electrode sites referenced to linked
earlobes in 14 young adults (7M, 7F) and 14 older adults (7M, 7F).  The
stimulus odorant consisted of three concentrations of isoamyl acetate
(low-10% medium-50%, and high-100%) presented with a 60 sec
interstimulus interval.  Component anaylsis was conducted on the N1-P2
and N2-P3 interpeak amplitude and N1, P2, P3 latency.  The P3 component
was elecited with a magnitude estimation procedure. Results
demonstrated that medium and high intensities produced significantly
larger amplitudes compared to the low intensity, across age groups.
There was no effect of odor intensity on latency values.  N1, P2, and
P3 latencies were affected by age such that young adults demonstrated
significantly faster processing of olfactory information compared to
older adults, replicating previous OERP findings. These results are
consistent with other ERP modalities and suggest that OERPs have
similar clinical and research utility.  Supported by NIH grant# DC02064
(CM) and training grant# DC00032 (MWG).

Estimation of missing data in psychophysiological research: Habituation
should not be ignored.
John J. Curtin and Christopher J. Patrick
Florida State University

Frequently in psychophysiological experiments one or more data points
are missing for many participants.  In within subject designs, mean
scores for each experimental condition (cell) are typically based on
several trials, with no special method of handling missing data
employed.  Cell means are simply derived from the remaining available
data.  This method  equivalent to estimating missing scores using the
mean of other trials in that cell would be adequate if all trials in
the cell were comparable.  However, this is often not the case.  Many
psychophysiological measures exhibit pronounced habituation of response
magnitude across trials.  Thus, the accuracy of a cell mean estimate
will depend upon the serial position of the missing data, with greater
error at either end of the habituation function.
An alternate regression based estimation, which accounts for
habituation by incorporating serial position of the missing
values, is proposed.  A Monte Carlo study was conducted to compare the
accuracy of the two methods in estimating startle magnitude for nine
subjects in a paradigm employing affective slides and varying probe
times. A substantial reduction (mean=18%) in the standard error of
estimate, an index of the mean difference between observed and
estimated scores, was obtained for the regression over the cell mean
method across varying numbers of missing values.  The effects of
including additional predictors (e.g., probe time) in the regression
and increasing numbers of missing values are evaluated.  It is
concluded that error variance due to missing scores can be reduced by
employing regression instead of default cell mean estimation.

Affective picture viewing: Task and stimulus effects on startle P3 and
Bruce N. Cuthbert, Harald T. Schupp, Margaret M. Bradley, Mark H. 
McManis, and Peter J. Lang
University of Florida

Previous research has shown that startle probe blinks covary with the
pleasantness of picture stimuli, while probe initiated cortical P3s are
modulated by a picture's attentional/arousal value.  These phenomena
were re-examined here, varying the task relevance of the probe
stimuli.  Stimuli were 18 pleasant, 18 neutral, and 18 unpleasant 
pictures were selected from the 
International Affective Picture System.  Each picture was seen for 6 s.  
Probes were either soft tones or 
acoustic startle stimuli (50/50 probability), presented during each 
picture and again during a 6 s post-
processing period.  Half of the subjects (n=40) performed a choice 
reaction time (RT) task at the 
second probe, pressing a button to indicate whether it was the same or 
different from the probe 
presented in the viewing interval; the other subjects were told to 
ignore the acoustic stimuli.  Affective 
modulation of the blink was again observed, and probe P3s during picture 
processing were again 
reduced in the context of affectively arousing pictures (both pleasant 
and unpleasant).  In addition, 
particularly for subjects performing the RT task, the cortical N1 wave 
was significantly larger for 
unpleasant than for either pleasant or neutral pictures.  Finally, in 
the post-picture period, P3s and N1s 
were systematically larger to startle probes that were preceded by tone 
probes, and this effect was most 
pronounced for subjects performing the RT task.  Overall, picture 
modulation of the cortical evoked 
potential appears to be accentuated when probes are task relevant.

Recognition memory following picture fragment completion: Effects of
memory instructions on ERP indices
Yael M. Cycowicz and David Friedman
New York State Psychiatric Institute

A recognition memory task was given to two groups of adults subjects
following picture fragment completion and free recall tasks.  The
groups differed in the instructions they had received prior to the
picture fragment completion task.  Subjects in the explicit memory
group were asked to memorize the pictures for a subsequent direct
memory test, while subjects in the implicit memory group were not so
instructed.  In the recognition task (choice old/new responses), the
subjects were shown a series of pictures in  complete form, half of
which were new and half of which had been viewed during the picture
fragment completion task.  Of the old pictures, half were seen once and
half were seen twice in the previous task.  While behavioral
performance in the recognition task (hit rate, false alarms, and
reaction time) did not differ between the groups, there was a
difference in brain activity.  For the explicit group, the ERP
waveforms elicited by seen once and seen twice pictures did not differ,
but both were more positive than the ERP to new pictures (i.e., showed
a repetition effect).  By contrast, the ERP waveforms of the implicit
group revealed an amplitude difference not only between new and old
pictures but between seen once and seen twice pictures.  Moreover, when
the ERPs to old pictures were averaged according to recall performance,
the implicit group's ERPs revealed no difference between recalled and
not recalled pictures.  The explicit group's ERPs, however, revealed
differential brain activity as a function of recall performance.  The
ERP data may reflect differences between the groups in the automatic
and controlled processes thought to underlie recognition memory

Effects of partial sleep deprivation on verbal and spatial memory
David Dalal and Nukte Edguer
Brandon University

This study examined the effects of partial sleep deprivation on verbal
and spatial memory by comparing ten military personnel who had a total
of five hours sleep over a four day period as part of their field
training with ten who were not sleep deprived. A cued recall task
involving nine military orders, each consisting of five pieces of
information and the cue, was used. Orders were presented in a verbal or
a pictorial form three at a time for a duration of one minute.
Following this subjects were presented with the cues associated with
the three orders just presented. Subjects were asked to recall as many
pieces of information from these orders as possible. After the
completion of all three sets of orders subjects completed the revised
Multiple Affect Adjective Check List and a question_ naire designed to
evaluate subjects' general functioning. The results of the study showed
sleep deprived subjects recalled significantly less information, had
higher depression scores and reported to have less concen- tration, to
be less collected and to be more drowsy and tired than the subjects who
were not sleep deprived. Subjects who received verbal orders also
recalled significantly less information, in addition to having signi-
ficantly higher depression and anxiety and lower positive affect scores
than the subjects who received pictorial orders. There was no
significant effect of sleep deprivation on the recall of verbal or
pictorial information.

The concurrent recording of electroencephalography and impedance
cardiography: A methodological study
Kim M. Dalton and Richard J. Davidson
University of Wisconsin - Madison

Three experiments were performed testing the effects of a variety of
impedance cardiograph (ZCG) electrode types and arrangements on
recorded electroencephalography (EEG) using either a monopolar, single
ear reference that was computer re-derived to an average ear reference,
or a physically linked ears reference.  EEG was recorded using Grass
Model 12A5_B and C amplifiers either alone or concurrently with a
Sorba, CIC-1000 impedance cardiograph.  The B amplifiers function
similarly to the C amplifiers when ZCG is recorded using a spot
electrode for the top current inducing electrode of the ZCG; there is
an overall decrease in power density of the EEG and this effect is
dependent on location on the head.  The B and C amplifiers function
differently when the top ZCG spot electrode is replaced by a
mylar-coated neck band electrode and EEG is recorded using a monopolar,
single ear reference.  There tends to be an overall increase in log
power density of the EEG in each frequency band below 60 Hz when B
amplifiers are used, but no effect when C amplifiers are used.  This
effect for the B amplifiers is diminished if the EEG is recorded using
a physically linked ears reference.  Recommendations for the concurrent
recording of EEG and ZCG are discussed.

ERP component differences due to the use of melatonin
Lori Darst, Scott Rammage, Rebecca Thew, and Joel Alexander
Western Oregon State College

Recent evidence suggests that Melatonin (a neurohormonal supplement)
is useful in treating insomnia, fighting jet-lag, and boosting immune
system functions. Given the widespread claims of beneficial effects and
the large increase in usage of this over-the-counter supplement, it
would be useful to assess the psychophysiological effects of
Melatonin.  In the present study Event-Related Potentials (ERPs) were
recorded from 15 channels (+ one eye channel) that were referred to the
nose, with a forehead ground during a standard auditory oddball task. A
total of 24 subjects (12 male and 12 female) participated in a within
subjects, counterbalanced design in which during one session subjects
received 3 mg Melatonin (commonly recommended dose) and during a
different session the same subject received 3 mg of a placebo. Only
subjects within a certain weight range, free of medication, and no
prior use of this drug served as volunteer subjects. Statistical
analysis of the data evinced a significant (p<.001) decrease in the
amplitude of the P300 component across electrode locations when
subjects were under the influence of Melatonin. Given the assumption
that decreases in P300 amplitude indicate lesser engagement of
attentional/memory processes, the results in part support the concern
voiced by some researchers regarding the carefree ingestion of
Melatonin when individuals are engaged in an activity that requires
significant processing of sensory stimuli.

Effects of spacial frequency of a vertically striped rotating drum on
vection-induced motion sickness
Michael S. Davis, Senqi Hu, Alexandrea H. Klose, Eileen M. Zabinsky, 
Stephanie P. Meux, Heather 
A. Jacobson, and Jennifer M. Westfall
Humboldt State University

The present study investigated the effects of differential spacial
frequencies of a vertically striped, horizontally rotating drum on the
observer's frequency of eye nystagmus, perceived velocity of
self-motion, and symptoms of motion sickness. Two experiments were
conducted. In experiment 1, 10 subjects viewed one minute intervals
inside an optokinetic rotating drum covered with five different number
combinations of black and white stripes: 6, 12, 24, 48, and 96. The
results indicated that subjects perceived significantly stronger
circular vection (p < 0.02) and generated significantly higher
frequencies of eye nystagmus (p < 0.0001) when they were viewing the
combination of 24 black and white stripes than when they were viewing
the combinations of 6, 12, 48, or 96 black and white stripes. In
experiment 2, 100 subjects viewed 16 minutes of an optokinetic rotating
drum covered with one of the five different combinations of black and
white stripes: 6, 12, 24, 48, or 96. The results indicated that
subjects in the group of viewing 24 stripe contrasts perceived
significantly stronger circular vection (p < 0.001), reported
significantly more severe symptoms of motion sickness (p < 0.001), and
showed significantly greater ratios of EGG 4-9 cycles per minute
spectral intensity (p < 0.02) than those in the groups of viewing 6,
12, 48, or 96 moving contrasts. These results confirmed that the
severity of vection induced motion sickness is determined by the
frequencies of horizontal eye movements while viewing differential
spatial frequencies of the stripes of the rotating drum.

Non-invasive recording of human salivary activity from surface
electrodes: Logic, method, and application
Christopher Davis, Tom Bauslaugh, and Anne Wintrup
Simon Fraser University

Traditional methods of recording human salivation potentially confound
results by affecting salivary flow.  Common techniques include
allowwing saliva to drain, head down, from the open mouth, collecting
saliva in the mouth for a fixed period while neither swallowing nor
expecorating, and inserting gauze dental rolls between cheek and gum to
absorb salivary production. The location and biological action of human
parotid salivary glands suggest that their activity might be recorded
from the kin surface over the gland.  An electrode placed on the cheek
over the parotid galand, referred to a mostoid electrode revealed
increases in activity in response to lemon juice in the mouth compared
to to a water control.  Contributions from potential extra-parotid
generators in nearby muscles and glands were negligible.  Surface
recorded potentials were correlated salivary production measured using
a modified Lashley Cup.

24-hour profiles of autonomic control over heart rate in exercisers and
Eco de Geus and Lorenz van Doornen
Vrije Universiteit

The cardioprotective effects of regular exercise are partially
attributed to an increase in vagal tone and a decrease in sympathetic
cardiac drive. Improved autonomic balance would be particularly
protective during stress when cardiac sympathetic drive is known to
increase in combination with decreased vagal tone. Assessment of
reactivity of PEP and RSA to a variety of laboratory stressors has not
shown the expected hyporeactivity for exercisers. However, in the
laboratory, short lasting stressors are used with brief pre- and post
task periods. The effect of exercise on cardiac autonomic balance may
show only during prolonged exposure to chronic stress throughout the
day or in the recovery and compensatory reactions during the night.  In
this study 24 hour PEP and RSA profiles of 16 regular exercisers
(8M/8F) were compared to those of 16 sedentary subjects (8M/8F) during
a normal work day. Ambulatory measurement of the thorax impedance was
used to measure PEP. RSA was assessed from the spectral decomposition
of the heart rate variability. Although there were no group differences
in type of activity, posture or body movement on the measurement day,
exercisers showed clearly lower HR throughout the 24 hour period than
non-exercisers (mean group difference at any time of the day: 12.1
bpm). PEP did not show any systematic group differences.  In contrast,
RSA showed a clear group by time-of-day interaction. During the morning
hours, RSA levels were comparable but in the afternoon and the evening
non-exercisers had significantly lower RSA than exercisers. During the
night this difference disappeared. These results could point to a
superior ability of exercisers to maintain vagal tone during fatigue at
the end of a work day. Alternatively, as will be discussed in the
paper, these results could be an artefact created by the inability of
RSA to index vagal tone in subjects with low heart rates.

Topographical analysis of the N100 auditory event-related potential
during the transition to sleep
Duncan de Lugt, Kimberly Cote, William Lee, and Kenneth Campbell
University of Ottawa

A long lasting negative slow wave overlapping the P1-N1-P2
event-related potential (ERP) complex is gradually removed as the
subject moves from a conscious state (wakefulness) to an unconscious
state (sleep). The purpose of the present study is to determine the
topographical distribution of the N100 ERP during the transition from
the waking to sleeping state. Ten subjects were presented with an
auditory oddball task. A standard 1000 Hz (70 dB) tone pip was
presented every 1.5 s. On 4 percent of trials the standard was randomly
changed to a 1500 Hz target. The subject was instructed to press a
hand-held response button upon detection of the target. EEG activity
was recorded from 29 scalp sites. Trials were averaged within
wakefulness, stage 1 and stage 2 sleep. During stage 1 sleep, the
detection rate was approximately 50 percent while during stage 2,
detections were rare (occurring on only 2 percent of trials). A
negative slow wave beginning at approximately 50 ms post-stimulus and
lasting for about 250 ms was gradually removed from the waking waveform
as the subject moved from wakefulness to stage 2 sleep.  This effect
was most marked over fronto-central areas of the scalp. The parallel
change in performance and in the characteristic ERP waveform suggest
that removal of the negative slow wave may reflect a loss of conscious
awareness of the auditory stimulus while leaving the sensory
representation of the stimulus unaffected.

Electrophysiological correlates of perceptual learning in the human
visual system
Aliasgar M. Dhoon and Nancy K. Squires
SUNY at Stony Brook

Several paradigms of perceptual learning have suggested that practice
can trigger long-term experience-dependent changes in the human adult
visual system.  In order to investigate the underlying mechanisms and
localize the brain-structures involved in visual learning, we used a
perceptual learning paradigm, in conjunction with electrophysiological
measures. Participants were trained to recognize the difference beween
vertical and horizontal targets within a diagonally oriented stimulus,
presented for 20ms before a powerful 20ms mask with an SOA of 100ms.
VERPs were recorded from Fz, Cz, Pz, Oz, O1 and O2 sites of the
International 10-20 system while the participants performed the visual
discrimination task.  Results from the current study indicate a fast,
rapidly saturating improvement in performance after the first few
trials, after which performance stabilizes.  In association with this
improvement in performance are significant changes in the the early
VERP components at latencies of 100ms or less over the occipital pole,
suggesting involvement of, and plasticity in the primary visual
cortex.  Gross electrophysiological changes suggest the possible
involvement of ensembles rather than single cells in perceptual
learning.  Results are discussed in terms of possible cortical inputs
to the primary visual cortex.

How different characteristics of emotional slides influence the
electrocortical response
Oliver Diedrich1, Ewald Naumann2, Dieter Bartussek2, and Niels 
1 University of Tuebingen, 2 University of Trier

Several studies have demonstrated that the valence of emotional color
slides influences the late positive complex (LPC) of the ERP:  Pleasant
and unpleasant slides elicit more positive LPC amplitudes than do
neutral slides. However, color slides also differ in other
characteristics that might influence the LPC, too.  In this study,
ratings of slide characteristics like valence, arousal, familiarity,
interest, number of assocations evoked by the slide, complexity, and
slide quality were obtained for 140 slides selected from the
International Affective Picture System (IAPS). The analysis of the
ratings reveals a complex correlation pattern, with all ratings varying
considerably across the slides.  Evoked potentials elicited by color
slides were analyzed according to the different slide characteristics.
All slide characteristics were found to influence the LPC. However,
these influences varied in effect size, temporal extent and topography,
with arousal and interest exhibiting the largest effects.  It is
concluded that (1) color slides vary in different characteristics; (2)
these characteristics are partly confounded; (3) ERPs might be used to
differentiate the slide properties.

Facial EMG and the orienting response.
Ulf Dimberg
Uppsala University

One important question in facial EMG research has been if facial
muscle reactions are associated with pure emotional activity and to
what degree they reflect other different psychological processes. The
present study investigates whether facial EMG reactions reflect
activity that can be interpreted as an orienting response (OR).
Thirty-four subjects were exposed to pictures of different facial
stimuli while facial muscle activity from the corrugator supercilii and
the zygomatic major muscles were measured. Because habituation of
autonomic responses is a distinguishing feature of the OR, skin
conductance (SCR) and heart rate (HR) were simultaneously measured. By
exposing subjects for two consecutive presentations of each of six
different faces it was possible to evaluate to what degree responses
habituated between the first and second presentation (trial) of each
The first presentation of each stimulus evoked an increased facial
EMG response in both muscles which decreased as a function of trials,
F(1,33)=9.83, p<.01. This effect, however, was most evident for the
zygomatic major muscle as indicated by the Trial x Muscle interaction,
F(1,27)= 10.11, p<.01. Similarily, the first trial evoked a large SCR
and a distinct HR decelerative response indicating an OR , which
consistently decreased with repeated presentation for both SCR,
F(1,33)=15.77, p<.001,  and for HR, F(1,27)=7.41, p<.05.  Thus,
particularily the zygomatic major muscle response showed a simultaneous
similar habituation pattern as did the autonomic responses.  These
results indicate that the facial EMG response to the first presentation
of a stimulus is mixed up with a fast habituating OR.

Event-related potentials in speech-related tasks in aphasic patients
and controls.
Christian Dobel1, Elvira Zobel1, Brigitte Rockstroh1, Rudolf Cohen1, and 
Paul Walter Schvnle2
1 University of Konstanz, 2 Rehabilitation Center Kliniken Schmieder 

Reorganization of cortical processes related to speech comprehension
and production in aphasia were examined by means of the topographical
pattern of event-related brain potentials.  ERPs were recorded from 19
scalp locations from (so far) 15 patients with aphasia, 9 patients with
diffuse brain lesions, and 9 healthy Ss.  Four experiments were
designed to activate speech-related processes (based on
neuropsychological tests sensitive to dysfunctions in aphasia),
involving concept formation, word recognition, access to the syntactic
word, and phonological encoding.  In all experiments, visual stimuli
were arranged within two-stimulus reaction time tasks comprising 1-s
S1, and 2-s interstimulus intervals, while S2-presentation was
terminated by the subject's response (left-hand button presses).  The
sensitivity of tasks for aphasia-specific dysfunctions was verified by
group differences in response latencies and errors. In all tasks ERPs
were characterised by late positive complex following S-onset, and a
CNV prior to the S2.  CNV scalp distribution varied between experiments
and groups, with more frontal and laeralised distribution in aphasics
than in the two control groups.  When the location of the maximum in
the scalp distribution of the CNV was established for every subject and
task, aphasics exhibited a more pronounced interindividual variability
and more lateralised location of maxima compared to controls.  Results
suggest that ERPs can be used to examine cortical processes related to
dysfunctions and/or substitution in aphasia; furthermore,
interindividual variability may be a useful measure aphasia;
furthermore, interindividual variability may be a useful measure
differentiating between groups, when averages across groups would
provide an inadequate impression.  Research was supported by the German
ministry of education and research (BMBF).

EEG-defined left versus right frontally activated groups differ in
metabolic asymmetry in the amygdalae
Isa V. Dolski, Jessica R. Malmstadt, Stacey M. Schaefer, Christine L. 
Larson, Heather C. 
Abercrombie, R. Terry Ward, Patrick A. Turski, Scott B. Perlman, James 
E. Holden, and Richard J. 
University of Wisconsin-Madison

Human and animal research suggests a role of the amygdala in negative
affect, but amygdala asymmetry per se, has not been tested.  Anterior
brain activation asymmetry as measured by EEG has been associated with
differences in positive and negative emotional reactivity.  In order to
determine whether amygdalar metabolic rates differ in those who show
patterns of extreme left or right anterior brain EEG asymmetry, 18-F
fluorodeoxyglucose Positron Emission Tomography (FDG-PET) was measured
in right-handed subjects who were selected from a large sample in which
resting EEG was collected on two occasions.  Based on average
midfrontal alpha power asymmetry (F4-F3), five extreme right frontally
activated (RFA) (-.7 SD from the mean) and four extreme left frontally
activated (LFA) (+.7 SD) subjects were selected.  Each subject's
Magnetic Resonance Image (MRI) was co-registered to their PET scan.
The left and right amygdala were drawn on the MRI for each subject, and
regional cerebral glucose metabolism (rCMR) was extracted for these
regions from the matching PET scan. Mean rCMR was residualized for
global metabolic rate. There was a significant main effect for Group
(RFA vs. LFA), such that those with RFA had higher amygdala metabolism
than those with LFA.  A significant interaction was found between Group
and Hemisphere (right vs.  left amygdala metabolism), such that
subjects with RFA showed significantly more left than right amygdala
metabolism. These data show that amygdalar metabolism differs in
subjects with extreme patterns of frontal EEG asymmetry.

Changes in heart period and RSA associated with quiet versus active
sleep state in full-term and preterm infants.
Jane A. Doussard-Roosevelt, Bonita D. McClenny, Cynthia A. Stifter, and 
Stephen W. Porges
University of Maryland

Even at full-term chronological age, the sleep of preterm infants is
not as well organized as that of full-term infants (Dreyfus-Brisac,
1974). Given that RSA amplitude (Vna) is related to behavioral state
regulation, we expect full-term infants, but perhaps not preterm
infants, to exhibit RSA differences during quiet versus active sleep.
Heart period and RSA (Vna) were measured during quiet and active sleep
in 17 preterm (<1500 g) and 18 full-term infants.  Preterm infants were
tested at approximately term (x = 39 wks); full-term infants were
tested at least 24 hours after delivery.  We hypothesize that
behaviorally coded sleep states are associated with changes in RSA in
the full-term but not the preterm infant.  State was coded every 15
seconds. Three minute ECG segments during quiet sleep and during active
sleep were analyzed.  Full-term infants had lower heart rate and higher
amplitude RSA than their preterm counterparts tested at term.
Within-group repeated measures analyses indicated longer heart period
during quiet sleep in both the preterm (F(1,16) = 9.41, p < .01) and
the full-term (F(1,17) = 12.47, p < .01) groups.  RSA amplitude
differences between states were found only in the full-term group
(F(1,17) = 21.67, p < .001) with higher RSA amplitude during quiet
sleep.  The preterm group showed no state-related RSA differences
(F(1,16) = .27, ns).  Findings are discussed in terms of the maturation
of vagal control of the heart and the relation between vagal control
and behavioral regulation.

Revised combined oddball and matching-to-sample procedure for detection
of simulated malingering of cognitive deficit with P300
J.W. Ellwanger, J.P. Rosenfeld, R. Bermann, K. Nolan, and J. Sweet
Northwestern University, Evanston Hospital

We have previously introduced a procedure, the "P3-MDMT" (Rosenfeld
et. al., 1994, 1995, 1996) in which  a presented sample 3-digit number
is followed after 3 seconds by a single test number which either
matches or fails to match the sample .  P300s to the rare matches
(P=.17) were significantly greater (group effect) than P300s to the
mismatches (P=.83) in subjects with behavioral  hit rates averaging
about 50%, however, the best  P300-based individual detection rate
achieved was <70%, a disappointing level.  In the present paradigm
modification, the sample number is followed by a series of 9 test
numbers, only one of which is a match.  P300s to the match (P=.11) and
mismatch (P=.89) were separately stored and processed.
Behavioral hit rates in 2 experiments varied from about 50% to
70%, which were the respective target hit rates achieved by
manipulation of instructions.  In both experiments, P<.001 for group
comparisons of match- versus mismatch-evoked P300 measured from
baseline to peak.  Individual, bootstrap analyses on either
peak-to-peak  or baseline-to-peak P300
amplitudes, match versus mismatch, utilizing single sweeps, (as in
Rosenfeld et. al., 1996) allowed correct classification of 16 of 19
subjects in both experiments, an 84% hit rate.

Psychophysiological changes related to mental absorption in a story
listening context.
Lynn S. Ferrante, Mark Sanders, Mary M. Assenat, and J. Richard Jennings
University of Pittsburgh

Absorption is a state of consciousness defined by a unitary focus on
a task to the exclusion of visceral afferents and unrelated
environmental input.  Our aim was to interpret the behavioral
correlates of this state to determine if absorption involves attentive
engagement in a story or an active participation with story elements.
A stronger investment of attention would result in decreased heart rate
and increased cardiac variability while a greater degree of activation
as a result of story participation would exhibit the opposite results.
Twenty-four subjects listened to an absorbing and a control story
session in which average heart rate, cardiac variability, a time
estimation measure, and a secondary task reaction time were assessed.
The subjects' self-reports rated specific time periods within each
session as being least and most absorbing as well as their overall
level of absorption in the story.
The subjective measures of time estimation  and self-reports as
well the reaction time significantly supported our induction of  an
absorbed state. Increased heart rate during the high absorption
periods (relative to the low absorption periods) occurred only during
the absorbing session (Session x Absorption Rating,
F(1,23)=10.13,p<.005).  Active participation with the story was
characterized by decreased cardiac variability during absorbing time
periods, but only during the absorbing manipulation
(F(1,17)=6.07,p<.05). Overall, our attemptsto induce absorption and
measure psychophysiological correlates of this state have been
successful. Absorption appears to involve activation via involvement in 
the story as opposed to attentive engagement in the story.

Extraversion, neuroticism and brain function: A PET study of
Hakan Fischer 1, Gustav Wik 2,3, and Mats Fredrikson 1
1 Uppsala University, Uppsala, Sweden, 2 Karolinska Institute and 
Hospital, Stockholm, Sweden, 
3University of Muenster, Muenster, Germany.

The personality dimensions Extraversion and Neuroticism seem
associated with differences in central nervous system function. We used
positron emission tomographic (PET) measures of regional cerebral blood
flow (rCBF) to investigate central neural differences in Extraversion
and Neuroticism, as determined by the Swedish version of the NEO PI-R
personality inventory. A median-split defined extraverts from
introverts and relatively more and less neurotic subjects. The relative
rCBF in the caudate nucleus,putamen and the secondary visual cortex was
higher in intro- than extraverts. In intro- but not extraverts activity
in the putamen was left-lateralized. These areas have high
concentrations of dopamine terminals, implicating a dopaminergic basis
for individual differences in Extraversion. As a function of
Extraversion rCBF did not differ in the prefrontal, orbitofrontal,
temporopolar, cingulate, primary visual cortex, the thalamus and the
hypothalamus. Thus, individual differences in Extraversion correlate to
subcortical rather than cortical brain regions. No rCBF differences
were related to Neuroticism. Because introverted subjects displayed an
increased neuronal activity in brain regions previously associated with
learning, motor and vigilance control, and since those behaviors in
part define Introversion, a subcortical, possibly dopaminergic, rather
than a cortical correlate of the personality dimension Extraversion is

Experimental studies on the placebo response: Psychophysiological
effects of information about drug action.
Magne Arve Flaten
University of Troms

It was investigated whether information given to a subject about the
action of a drug that the subject ingested, could elicit
psychophysiological and subjective responses. Forty- eight healthy
volunteers (19-33 years) received information that the substance they
ingested was a drug that acted as either a relaxant (Relaxant group),
as a stimulant (Stimulant group), or that was an inactive substance
(Control group). All subjects received double-blind administrations of
an inactive substance (lactose), dissolved in water. Dependent
variables were skin conductance and startle reflexes elicited by
tactile and auditory stimuli, and subjective measures of stress and
arousal.  Recordings were performed before administration of the
substance.  This indicates that the knowledge the experimentator had of
which information the subject should receive, modulated the response of
the subject accordingly. This lends support to previous research, and
indicates that subtle cues may influence the responsiveness of the

The separate contributions of the tactile and acoustic components of
airpuffs to the blink reflex
Magne Arve Flaten1 and Terry D. Blumenthal2
1University of Troms, Norway, 2Wake Forest University, NC

Airpuffs are often used as blink-reflex eliciting stimuli in classical
conditioning and reflex modification studies. When an airpuff leaves
the nozzle it makes an abrupt noise of sufficient intensity to be
reflex eliciting. The present study investigated the separate
contributions of the acoustic and tactile components of the airpuff
used to elicit the blink-reflex in 18 human subjects (eight females,
mean age 29.7 years). Two within-subjects variables were presented in
counterbalanced order: airpuff intensity (1.5, 3, and 4.5 psi)and
stimulus condition (airpuff plus noise, with the airpuff directed to
the temple; airpuff alone, with the airpuff directed to the temple, but
the subject was wearing sound-attenuating headphones and earplugs;
noise alone, with the airpuff directed away from the temple, and
without the subject wearing headphones or earplugs). Fifteen trials
were presented at each airpuff intensity in each stimulus condition.
The results showed that blink reflexes to noise were significantly
smaller than to airpuffs. Reflexes to airpuff plus noise were
significantly larger and faster compared to reflexes to the airpuff
alone. This indicates that the noise has an excitatory influence on the
airpuff-elicited blink reflex. This suggests that the noise component
of the airpuff should be reduced in studies of reflex modification,
since the "airpuff alone" is often used as the control condition, which
may be elevated by the noise.

Reorganization of motor and somatosensory cortex is related to phantom
limb pain but not to non-painful phantom phenomena.
Herta Flor1, Niels Birbaumer2, 1Anke Karl, 1Sabine Gosser, 1Werner 
Muehlnickel, and 2Werner 
1Humboldt-University, Berlin 2University of Tuebingen

Phantom limb pain and non-painful phantom phenomena (such as facial
remapping and telescoping) occur in about 80% of amputees.  The causes
for these often very disturbing sensations are still unknown.  We have
previously reported a close association between phantom limb pain and
reorganization of somatosensory cortex. The purpose of this study was
(a) to extend these findings to motor cortex and (b) to examine in
greater detail the relationship of non-painful phantom phenomena and
reorganization of somatosensory and motor cortex.  Study 1: 5
unilateral arm amputees with phantom limb pain and 5 without phantom
limb pain participated in transcranial magnetic stimulation.  EMG of
the biceps and zygomaticus muscle as well as ratings of subjective
sensations created by the magnetic stimulation served as dependent
variables.  Study 2: 13 unilateral arm amputees with phantom limb pain
and 7 without phantom limb pain received 1000 pneumatic stimuli to the
1st and 5th digit of the intact hand and the corner of the mouth of the
intact and the amputation side.  EEG was assessed from 60 channels
centered over the postcentral gyrus, magnetic resonance images were
taken to assess anatomical structures.  Equivalent current dipoles were
computed for the four locations and cortical reorganization was
assessed by computing the distance between two lip representations.
The studies showed significant reorganization of the motor and
somatosensory cortices only in the phantom limb pain patients. Non-
painful phantom phenomena were unrelated to reorganization. Different
processes may contribute to phantom limb pain and non-painful phantom
phenomena.  Supported by the DFG (Fl 156/15, 16; Bi 195/24).

Sensitisation to fear-relevant stimuli after masked conditioning to
directed threat of cultural and biological origin.
Anders Flykt, Francisco Esteves, and Arne Ohman
Karolinska Institute and Uppsala University

A series of conditioning experiments was conducted to test whether it
is possible to establish an aversive differential skin conductance
response to backwardly masked cultural threat stimuli. In experiment
one group one was conditioned by white-noise to a masked picture of a
gun directed toward the subject whereas a masked picture of a gun
directed to the side served as control stimulus. The other group had a
masked picture of a directed snake followed by electric shock and a
masked picture of an undirected snake as the control stimulus. To test
the hypothesis that what looked like a conditioning effect was a
sensitisation effect, a second experiment was performed.  One half of
the subjects had pictures of snakes as stimuli and the other half had
pictures of guns. Each stimulus condition had three sub-groups: a
replication of the first experiment, the reversed contingence and the
third group did not have any  target stimuli before the masking
stimulus. The result revealed a sensitisation effect to both biological
(i.e. snakes) and cultural threat (i.e. guns). Stimuli directed towards
the subjects resulted in an elevated SC responding, regardless of
treatment. To test if direction was the reason for the sensitisation
effect, a third experiment was conducted. Two conditioning groups and
two sensitisation groups were run. Pictures of fingers pointing in
different directions were used as stimuli. The only differential
responding effect was in the unmasked conditioning group.
The results suggests that the differential SC responding results
from attentive appraisal of the combination of threat and

Event-related brain potentials (ERPs) measure orthographic and
phonological influences during silent reading involving homophones and
Kelly A.K. Forbes and John F. Connolly
Dalhousie University

Traditionally, the role of phonological representations during reading
has been studied using behavioural measures and isolated words.  More
recently, event-related brain potentials (ERPs) have demonstrated that
phonological codes are actively involved in word identification.  The
N400 component occurs reliably to semantic incongruities. An earlier
component, the N270, appears to be sensitive to deviations in the
orthographic representation of the expected terminal words.  This study
manipulated the endings for highly contextually constrained sentences.
Endings were either semantically congruous or incongruous. Incongruous
endings were either orthographically similar (OS) or orthographically
dissimilar (OD) to the expected terminal words.  OS endings were either
phonologically expected words (i.e., homophone foils) or were
phonologically unexpected words and non-words.  OD endings were
phonologically unexpected words and non-words.  All non-words were
pronounceable but did not share phonological representation with any
legal English words. A N400 response was not expected to occur to
homophone foils if phonological recoding occurred but was expected to
occur to all other semantically incongruous endings.  However, a N400
was expected to occur to all semantically incongruous endings,
including homophone foils, if phonological recoding did not occur.  A
N270 was expected to be elicited only by OD words and non-words because
their orthographic representation deviated from the expected
orthography.  Results demonstrated that phonological recoding occurred
and indicated that phonological and orthographic influences may operate
in parallel.  Findings are discussed in terms of current reading models
that propose an active role for phonological representations during the
word identification process.

Pre-pulse inhibition of ERP components elicited by startling noises
Judith M. Ford, Walton T. Roth, Clarine M Bell, Yafeng Li, and Shamini 
Stanford University and VAPA Health Care System, Palo Alto

	Startling noises elicit blinks which are inhibited when
preceded by a stimulus (pre-pulse).  The N1 and P300 components of the
event-related brain potential (ERP) are also elicited by startling
noises.  We asked whether N1 and P300 behave like startle blinks and
are inhibited by a pre-pulse.  Eight subjects (18-54 years old) were
studied who showed pre- pulse inhibition of startle blink.  They were
presented with intense broadband noises, one third of which were
preceded 120 ms by a 1000 Hz, 25 ms, 85 dB SPL tone (pre-pulse, 120PP),
one third of which were preceded 600 ms by the pre- pulse (600PP), and
one third of which were not preceded by a pre-pulse.  The inter-trial
interval varied between 13, 15, and 17 s.  N1 was measured at Cz, P300
at Fz, Cz and Pz, and blinks from electrodes placed above and below the
right eye.  Blink artifacts were mathematically removed from ERPs,
allowing the measurement of N1.  In the 120PP condition, the
noise-elicited N1 component was overlapped by the ERP to the
pre-pulse.  This overlap was removed by subtracting the tone- elicited
ERP in the 600PP condition from the tone-elicited ERP to the 120PP.  As
expected, 120PP significantly reduced startle blink amplitude, but
600PP did not.  Both 120PP and 600PP significantly reduced N1 and P300
amplitudes.  These data suggest that blink and ERP components elicited
by startling noises are affected differently by a pre-pulse.

Nociceptive flexion reflex (RIII) threshold is inversely related to
resting systolic blood pressure: Evidence of hypoalgesia in
normotensive men and women.
Christopher France, Gary Page, Valerie Bonk, Michelle Henninger, Kay 
Stewart, Catherine Scott, and 
Jennifer Polewchak
Ohio University

Previous studies have demonstrated that systolic blood pressure (SBP)
is inversely related to subjective pain ratings. Although these
findings have been replicated across laboratories, limitations include
a reliance on verbal reports of pain and a primary focus on men. We
addressed these limitations by examining the nociceptive flexion reflex
(NFR) in men and women. The NFR is a protective withdrawal reflex
characterized by biceps femoris muscle contraction following noxious
sural nerve stimulation.  Numerous studies have demonstrated that the
long latency component of the NFR, known as RIII, is associated with
stimulation of nociceptive afferents and is highly correlated with
perceived pain threshold. Resting SBP was recorded from 44 healthy men
and women using an IBS-700A automated sphygmomanometer. On a separate
day, participants received a series of electrical stimulations applied
over the sural nerve using an "ascending staircase" method of
increasing and decreasing intensities. NFR threshold was defined as the
average mA level required to elicit RIII on an electromyogram recorded
from the ipsilateral biceps femoris.  Although NFR threshold was
significantly higher for men [F(1,43)=10.3, p<.01], a stepwise
regression using both gender and SBP as potential predictors was
completed with only SBP as a significant predictor of NFR threshold
(r=.52, p<.001).  When gender was forced into the equation on the first
step, the observed partial correlations were r=.40 (p=.01) for SBP and
r=.23 (p=.17) for gender. These findings suggest that resting SBP is
inversely related to pain perception in men and women.

Brain mechanisms in associative and non-associative processes in
classical conditioning
Mats Fredrikson1, Hakan Fischer1, and Gustav Wik2
1 Uppsala University 2 Karolinska Institute

Using positron emission tomography we contrasted regional cerebral
blood flow (rCBF) during acquisition and extinction of classical
conditioning. Acquisition was achieved by pairing a visual snake
stimulus with an unconditioned electric shock delivered to the right
hand. Under stimulus pairing, in contrast to subsequent snake alone
presentations during extinction, rCBF was higher in the vermis and
bilaterally in the thalamus, in the left hippocampus, the left
perirhinal and entorhinal cortices, the right nucleus accumbens, the
left inferior parietal and occipitotemporal corticies. The rCBF was
lower in the left secondary visual cortex, the right dorsolateral
prefrontal cortex, the left posterior cingulate cortex, bilaterally in
the anterior cingulate cortex, in the left primary motor and right
premotor cortex. Thus, even though sharing similarities  the functional
organization of emotional learning during acquisition and extinction is
different. The general pattern of blood flow changes during acquisition
resemble that observed during phobic anxiety provocation and support
that human fear-conditioning in part may serve as a laboratory analogue
to phobic anxiety.

Autonomic and affective characteristics of shock avoidance and cold
face stress in college females
Bruce H. Friedman1 and Julian F. Thayer2
1Washington University in St Louis, 2University of Missouri-Columbia

Shock avoidance (SA) and cold face stress (CFS) are widely-used
laboratory tasks that elicit distinct autonomic cardiovascular patterns
(Friedman, Thayer, & Tyrrell, in press, Clinical Autonomic Research).
However, the affective components of these manipulations are not
well-known. A study was conducted to examine self-reported emotional
responses to these tasks. 17 female university students engaged in a
reaction-time SA task and CFS (induced via a plastic bag filled with
cold water, avge. temp 8 deg. C, held on the forehead) for four-minutes
periods while ECG was recorded. Following each period, a 13-item
emotion adjective questionnaire with a five-point Likert scale rating
for each item was completed by each participant. ECG spectral power was
calculated in high (HF; .18-.35 Hz) and low (LF;
.04-.15 Hz) frequency bands with an autoregressive algorithm. The ratio
of LF/HF was derived as an index of sympathovagal balance, with higher
ratios corresponding to greater relative sympathetic activity.
Contrast analysis revealed that SA evoked higher LF/HF than CFS
(SA=1.2, CFS=.3, t(15)=2.6, p<.01), and was characterized by greater
levels of interest, excitement, and activation (all p<.05). CFS was
associated with more relaxation, serenity, and pleasantness (all
p<.05). These data are consistent with literature which suggests that
autonomic responses are more sensitive to the arousal rather than the
valence dimension of affective space (see Patrick, 1994,
Psychophysiology, 319-329). It is also possible that these tasks which
are generally viewed as aversive may actually evoke positive emotional

Functional neuroanatomy of classically conditioned electrodermal
Tomas Furmark1, Hakan Fischer1, Gustav Wik2, and Mats Fredrikson1
1Uppsala University, 2Karolinska Institute

The neuroanatomical correlates of electrodermal activity (EDA) are not
well characterized. Studies on conditioned EDA and measures of central
neural activity in humans are lacking. Lesion and anatomical imaging
studies suggest that the prefrontal, right inferior parietal and left
temporal areas, as well as the anterior cingulate gyrus and pons, might
be important in mediating electrodermal orienting responses. To
evaluate the brain mechanisms underlying conditioned EDA the present
study correlated conditioned regional cerebral blood flow (rCBF) with
measures of EDA in eight healthy subjects. The rCBF and EDA-data was
obtained before and after Pavlovian fear-conditioning using positron
emission tomography and [15O]-butanol as tracer. The regions of
interest were chosen in accordance with earlier findings on animals and
humans.  Conditioned rCBF in and [15O]-butanol as tracer. The regions
of interest were chosen in accordance with earlier findings on animals
and humans. Conditioned rCBF in the anterior cingulate cortex (Brodmann
areas 24, 33) showed a positive correlation with conditioned
electrodermal fluctuations (n = 8; r= .76, p =.02).
Borderline significant positive correlations were obtained between
conditioned EDA and rCBF in the parietal cortex (Brodmann areas 39,
40; n =8; r = .62, p = .05) and in the amygdalae (n=6; r = .62, p =
.10). Because the cingulate cortex and the amygdalae are associated
with fear-conditioning in animals and the parietal cortex with
orienting responses in humans, data support that conditioned EDA
reflect affective and attentive processes regulated by limbic and
cortical areas.

Right lateralized P300s in response to emotional stimuli
Wendi Gardner1, John Cacioppo2, and Gary Berntson2
1Northwestern University, 2The Ohio State University

We recently reported the finding that P300s evoked during attitudinal
or evaluative (positive/negative) categorizations are significantly
right lateralized in comparison to those evoked to identical stimuli
during a non-evaluative task (Cacioppo, Crites, Gardner, & Berntson,
1996). One possible explanation for this right lateralization is the
possible overlap between emotional and attitudinal processing.  Because
affect is considered a fundamental feature of evaluation, we examined
whether increasing the affective nature of the stimuli was sufficient
to produce increased right lateralization of the P300 during an
evaluative task. 20 subjects evaluated 4 groups of photographic slides
(positive low emotional, positive high emotional, negative low
emotional, negative high emotional) that were matched on extremity but
differed in affective intensity.  All slides were presented within a
series of positive slides.  As expected, a larger P300 was evoked to
evaluatively rare (negative) slides than to evaluatively frequent
(positive) slides within the positive context.  In addition, the P300s
evoked during this evaluative task were significantly right
lateralized.  Finally, this lateralization interacted with the
affective nature of the stimuli such that highly emotional slides (both
positive and negative) produced more right lateralized P300s than did
the less emotional slides.  Results imply that the neural mechanisms
underlying attitudinal and emotional processing may partially overlap.

Trait aggression and cardiovascular functioning: An examination of
Timothy S. Garvey and Bert N. Uchino
University of Utah

Trait aggression has been linked to increased mortality from
cardiovascular disorders.  However, less data exist on the potential
mechanisms responsible for such epidemiological findings. In the
present study, we examined the relationship between trait aggression
(Buss & Perry, 1992) and resting measures of systolic blood pressure
(SBP), diastolic blood pressure (DBP), and heart rate (HR) in 37 men
and 36 women.  In addition, we evaluated the mechanisms postulated by
the psychosocial vulnerability and health behavior models in accounting
for the association between trait aggression and health (see Smith,
	Results revealed that trait aggression predicted a higher
resting HR (r = .31, p < .01).  Subsequent analyses revealed
that statistically controlling for social support (ISEL; Cohen et al.,
1985) as a mediator did not influence this association (r = .29, p <
.01).  In addition, separate statistical controls for body mass,
exercise habits, sleep habits, and caffeine intake did not alter the
association between trait aggression and resting HR.  Due to the
potential importance of "hyperkinetic" states (e.g., increased HR) in
predicting subsequent elevations in peripheral resistance, these data
suggest that trait aggression may be an important personologic factor
influencing long-term health.  Future research is needed to identify
more specific social, psychological, and behavioral mechanisms
potentially responsible for these associations.

Cognitive olfactory event-related brain potentials, neuropsychological
performance, and olfactory thresholds in the young and elderly
Mark W. Geisler1,2, Charlie D. Morgan3, James W. Covington2, Spencer 
Wetter2, Mario Dulay2, 
Abraham Galvan2, and Claire Murphy1,2
1University of California School of Medicine, San Diego 2San Diego State 
University, 3SDSU/UCSD 
Joint Doctoral Program in Clinical Psychology

The P3 event-related brain potential (ERP) reflects neuroelectric
activity relating to cognitive processing speed and allocation of
attention.  The present study assessed P3 olfactory event-related
potentials (OERP) elicited in a single stimulus paradigm with
neuropsychological performance and olfactory thresholds in the young
and elderly.  OERPs were recorded monopolarly at the Fz, Cz, and Pz
electrode sites in 16 young adults (8 M, 8 F) and 16 elderly adults (8
M, 8 F) with inter-stimulus intervals (ISIs) of 45, 60 and 90 seconds
using amyl acetate, geraniol, and phenylethyl alcohol as odorants. P3
peak amplitude and latency were compared with olfactory thresholds and
cognitive performance. Odor thresholds were measured using a
two-alternative forced choice ascending method of limits for all three
odorants. Neuropsychological performance was assessed using the
California Verbal Learning Test (CVLT) and the Trail Making Test
(TMT).  Cognitive estimates of odor magnitudes elicited highly
reproducible P3s.  Results indicate that late cognitive OERP components
change with age and are associated with neuropsychological performance
changes on all three ISIs and all three odors. Specifically, an
increase in  P3 amplitude and a decrease in P3 latency was associated
with an increase in the number of items recalled on the CVLT. Faster
completion of the TMT was correlated with increased P3 amplitudes and
decreased P3 latencies. No age or gender differences in odor threshold
were found.  Findings suggest OERPs in conjunction with
neuropsychological measures can be used to assess olfactory dysfunction
and cognitive dementia associated with aging.  Supported by NIH grant#
DC02064 (CM) and training grant# DC00032 (MWG).

Stress-induced blood pressure measurement predicts left ventricular
mass index three years later in borderline hypertensive men.
Anastasia Georgiades1, Carola Lemne2, Kaj Lindvall3, Ulf de Faire2,4, 
and Mats Fredrikson1.
1Uppsala University, 2Karolinska Hospital, 3Huddinge Hospital, 
4Karolinska Institute

Exaggerated cardiovascular reactivity has been implicated in the
development of left ventricular hypertrophy. The aim of the present
study was to investigate the association between stress-induced blood
pressure levels and reactivity in the laboratory and left ventricular
mass index (LVMI) over a three year period. 66 middle-aged men (mean
age 50 years) with borderline hypertension (diastolic blood pressure of
85 to 94 mmHg) participated on two occasions separated by 3 years. The
stress reactivity assessment included a pretask resting period and two
stress tests, mental arithmetic and isometric muscle contraction. Left
ventricular wall thickness was assessed by M-mode echocardiography. A
stepwise multiple regression analysis with LVMI at the three year
follow-up as dependent variable was applied. Initial baseline LVMI was
used as forced independent variable. Casual blood pressure levels, age,
body mass index and aggravated average stress-induced blood pressure
levels and reactivity measures were entered stepwise as independent
variables. Initial baseline LVMI explained 24% of the variance in LVMI
three years later. Aggravated averaged mean blood pressure reactivity
in the laboratory was the strongest independent variable within the
regression analysis and added 15% to the prediction of LVMI three years
later over and above the variance explained by baseline LVMI. In the
present study, blood pressure reactivity was the strongest independent
predictor of LVMI three years later. This supports the assumption that
exaggerated blood pressure reactivity can be a risk factor for the
development of left ventricular hypertrophy.

No change in parasympathetic tone during embarrassment and blushing in
social phobics
Alexander L. Gerlach, Frank H. Wilhelm, and Walton T. Roth
Stanford University School of Medicine and VAPA Health Care System, Palo 

The few studies on the psychophysiology of embarrassment have
suggested possible involvement of parasympathetic activation. However,
blushing, the hallmark of embarrassment and a prominent symptom in
social phobia, is more likely to be produced by cervical sympathetic
outflow.  Hitherto there has been no evidence of parasympathetic
innervation of the facial blood vessels. We report here preliminary
results from a sample of 26 social phobics,  17 of whom presented with
the chief complaint of fear of blushing.  The phobic watched, together
with a 2-person audience, a previously made videotape of the phobic
singing a children's song. While RSA during this embarrassing task did
not indicate heightened parasympathetic tone (F(2,50)=0.73; p=0.49),
increased heart rate marked sympathetic activation (F(2,50)=8.92;
p<0.0005).  A photoplethysmographic measure of blushing showed an
increase of superficial facial blood volume at the beginning of the
embarrassing task  (F(1,24)=18.39; p<0.0003).  Thus, our data do not
support the notion that parasympathetic activation plays a significant
role in social phobia. A group of non-social phobic individuals is
currently being assessed with the same paradigm.

Ambulatory monitoring of EGG during chemotherapy-induced nausea
Peter J. Gianaros1, Gary R. Morrow2, Jane T. Hickok2, and Robert M. 
1Pennsylvania State University, 2University of Rochester 

Prior research in our laboratory has demonstrated that gastric
tachyarrhythmia, abnormal gastric myoelectric activity, accompanies the
onset of nausea during motion sickness. Nausea and vomiting are two of
the most severe side effects of cancer chemotherapy; however, little is
known about the physiological changes mediating these symptoms. The
purpose of the present study was to investigate changes in gastric
myoelectric activity during chemotherapy-induced nausea. Eleven female
chemotherapy patients from the University of Rochester Cancer Center
were studied. All were being treated for a variety of histologically
determined malignancies (8 breast cancer, 2 ovarian cancer, 1 bladder
cancer) and ranged in age from 34-66 years old with a median age of 48
years. All patients received standard clinical dosages of both
anti-emetic drugs (e.g., Decadron and  Kytril) and chemotherapy agents
(e.g., Cisplatin and Taxol). A UFI Biolog Ambulatory Data Recording
System was used to monitor continuously electrogastrograms for 24 h
following the patients' first chemotherapy treatment. Onset to nausea
ranged from 3-9 h following treatment. Running spectral analyses were
performed on 15 min intervals at 2 h and at 15 min prior to reported
nausea.  Mean spectral percentage of gastric tachyarrhythmia was
greater during the 15 min period prior to reported nausea than during
the 15 min interval at 2 h prior to nausea. These results suggest that
gastric tachyarrhythmia is a sensitive index of chemotherapy-induced
nausea, and that the changes in gastric myoelectric activity seem to be
similar during the nausea of motion sickness and chemotherapy.

Effects of tobacco abstinence on frontal theta and other EEG spectral
components during a visual vigilance task
David G. Gilbert
Southern Illinois University

EEG was recorded from 19 sites in female smokers on two occasions
while participants performed a visual rapid information processing
(VRIP) task. On one of the occasions participants were not smoking
deprived and on the other they were approximately 15 hours deprived.
The VRIP task required a button press whenever three consecutive odd or
three consecutive even digits appeared on a computer monitor.  Single
digits appeared at the rate of 116/min for the 16 min task.  Smoking
abstinence, relative to nonabstinence, resulted in a decrease in
frontal (Fz, F3, F4) high-frequency theta power (6.0-7.5 Hz).  Frontal
midline theta is characteristic of vigilance and other tasks involving
attention.  Vigilance performance was also decreased in the abstinence
relative to the nonabstinence condition.  Given that frontal theta
during vigilance tasks is positively related to dopaminergic activity
and inversely associated with platelet MAO B activity, it is
hypothesized that the abstinence-induced reductions in frontal theta
power were in part a result of abstinence-related increases in MAO B
activity that decreased dopaminergic activity.  This model is discussed
in terms of recent findings suggesting that nicotine acts as a MAO
inhibitor and may thereby have antidepressant effects, especially in
those disposed to depression.  This model may help to explain earlier
findings from our lab showing depression-prone smokers to exhibit
larger decrements in vigilance and larger increments in depressive
affect when they quit smoking.

A study of the ability to precisely control the contraction of muscles
in  different regions of the face
Eric Girard1, Louis G. Tassinary2, Arvid Kappas1, and Daniel Bontempo2,
1Laval University, 2Texas A&M University

The polysemous nature of facial expressions is widely recognized.
Many investigators have sought ways to use the topography of facial
actions to disambiguate the meaning of a particular expression, though
surprisingly little quantitative research has specifically examined the
ability of naive subjects to control the topography of their facial
actions.  Instead, emotion researchers have often drawn conclusions
about the controllability of different regions of the face based solely
on clinical neuroanatomical data.  This is misleading because the
effects of the documented differences in the lateralization of the
neural control of the upper and lower part of the face are only clearly
evident in cases of obvious neuropathology (cf. Fridlund, 1994).
Fourteen participants were asked to pose three simple expressions
(frown, smile, and lip press) in three different ways.  To alter the
dynamics of the expressions in predictable ways, participants were
required to track the changing frequency of a 2.5 s tone.  The rate of
change was altered according to the function Y = X^b, where Y =
frequency, X = time, and b was set equal to either 2.5, 1 or .4.
Electromyographic activity from the brow, cheek and mouth regions was
recorded.  Overall, participants were able to more precisely track the
tone with either a frown (r=.651) or a lip press (r=.698) than with a
smile (r=.383), F(2,26)=9.72.  In addition, they were able to more
precisely track the tone profiles that changed either linearly (b=1;
r=.611)  or began slowly (b=2.5; r=.633), rather than one which began
abrubtly (b=.4; r=.488), F(2,26) = 2.06.

Startle reflex modulation during unconscious processing of fear evoking
slides in animal phobics
Jutta Globisch, Almut I. Weike, and Alfons O. Hamm
University of Greifswald

Animal Phobics show startle response potentiation during viewing of
their feared material. Animal data suggest that this potentiation is
mediated by subcortical (mainly amygdaloid) structures. The present
study was designed to examine whether Animal Phobics would also show
potentiation of their blink responses to masked stimuli, preventing
elaborate processing of thesematerials.  30 subjects (14 Animal
Phobics) were selected according to their scores in the SPQ and SNAQ.
Feared (spiders or snakes), neutral and pleasant slides were presented
for 20 ms. To prevent afterimages, a visual mask followed the emotional
slide for 130 ms. Acoustic startle probes (50 ms white noise bursts of
105 dB(A)) were presented 120, 300, 800, 1300 and 3800 ms after the
onset of the emotional stimulus. Startle response magnitudes as well as
electrodermal and cardiovascular responses were recorded. Subjects were
also administered the SAM and a forced choice recognition test. Blink
responses to probe stimuli were not affected by the emotional contents
of the different target stimuli. Animal Phobics did not show any fear
potentiated startle response if pictures of spiders or snakes were
backwardly masked after 20 ms.  Autonomic responses supported these
results. Possible reasons for the failure to find fear potentiated
startle effects in this experiment will be discussed.  This research
was supported by grants Ha1593/6-2 and Ha1593/10-1 from the Deutsche
Forschungsgemeinschaft (DFG) to Alfons Hamm.

Reactivity, rumination, and recovery: Emotional components of
cardiovascular responses.
Laura M. Glynn1, Nicholas Christenfeld1, and William Gerin2
1University of California at San Diego, 2Cornell University Medical 
Center at The New York Hospital

We examined the effects of the level of reactivity, the emotionality
of the stressor, and rumination on cardiovascular recovery.
Seventy-two subjects performed one of four tasks, with blood pressure
continuously monitored during a baseline, stress, recovery, and
rumination period.  Of the two tasks which produced high reactivity,
one was emotional (serial subtraction with harassment, 34 mmHg SBP
reactivity) and one not (walking in place, 29 mmHg).  Similarly, of the
two producing low reactivity, one was emotional (shock threat, 18 mmHg)
and one not (cold pressor, 15 mmHg).  Subjects did rate the emotional
tasks as significantly more emotionally involving than the
non-emotional tasks.
At the end of a 10 minute recovery period, the groups which had
performed one of the emotional tasks displayed identical and elevated
levels of systolic blood pressure (15 mmHg), but the non-emotional
groups had recovered to within 5 mmHg of baseline.  Recovery depends on
the emotional component of the task (or possibly active coping), and
not on the reactivity that the task originally created.  All subjects
then performed a distracting task for six minutes at the end of which
the two emotional groups were also within 5 mmHg of baseline.  However,
during the following rumination period, the emotional groups drifted
back to 13 mmHg above baseline, and the non-emotional groups remained
at baseline.  These results suggest that the slow recovery of the
emotional groups is due to continued thinking about the tasks, since
when this is prevented they recover, and when it is allowed again, they
return to an elevated level.

Context and startle: Fear-potentiated startle to threat cues in the
Christian Grillon
Yale University School of Medicine

Startle is increased by negative affects. For example, anticipating
unpleasant shocks potentiates startle (fear-potentiated startle).
Recent findings indicate that startle is sensitive to lighting
conditions in humans and animals. In humans, a diurnal species, startle
is increased in the dark. In the rat, a nocturnal species, startle is
facilitated by bright lights. This later effect is blocked by
anxiolytic drugs, suggesting that changes in illumination produce an
affective reaction.
	One interpretation for the darkness facilitation of startle in
humans is that anticipating aversive startling stimuli is more
anxiogenic in the dark than in the light.  This interpretation predicts
that anticipating more unpleasant stimuli, such as electric shocks,
might produce greater fear-potentiated startle in the dark than in the
light. The present study tested this hypothesis.
	Startle stimuli were delivered in alternating periods of
darkness and light.  During each period, subjects were
presented with threat and safe conditions. They could receive a shock
in the threat but not in the safe condition.
	Startle was facilitated by darkness and potentiated in the
threat condition. However, the degree of potentiation to the
threat signal was similar in the dark and in the light. These results
do not support the hypothesis that darkness increased the anxiogenic
effect of aversive stimulation.  However, they indicate that the effect
of darkness and threat of shock on startle are additive.  The results
will be discussed with references to animal research which suggests
that additivity is suggestive of independence (i.e., different neural
substrates) between two processes.

Context and startle: Darkness facilitates startle amplitude
Christian Grillon, Mark Pellowski, and Kathleen Merikangas
Yale University School of Medicine

The startle reflex is affected by negative emotional states. It is
potentiated during shock anticipation, during conditioned fear, and
during the processing of unpleasant stimuli. The present study examined
the impact of darkness on startle. Given anecdotal  evidence suggesting
that humans, and more specifically children, are afraid of the dark, it
was hypothesized that darkness would increase startle.  This hypothesis
was also suggested by recent animal studies that have indicated that in
the rat, a nocturnal species, bright lights increase startle (Walker &
Davis, submitted). The basic experiment consisted of delivering
acousting startle stimuli in alternating periods (2 min each) of
darkness and light. An initial study also investigated the effect of
darkness on prepulse inhibition (PPI) using an acoustic prepulse.
Darkness  was found to significantly increase startle without
affecting  PPI.  The PPI result suggests that the facilitation of
startle in the dark was not due increased attention to the auditory
modality.  Increased attention to the auditory modality would be
expected to increased PPI.
	A second study examined the effect of darkness on startle in
various age group: 8-12 years old, 13-18 years old, 19-30 years
old, and 31 to 60 years old.  Results showed that the amplitude of
startle was facilitated in the 4 groups.
	These results, which suggest that an emotional process is
responsible for the facilitation of startle in the dark, will
be discussed with respect to studies assessing the effect of light/dark
in the rat currently underway in Dr. Davis' laboratory.

Context and startle: Effect of explicit and contextual cue conditioning
following paired vs. unpaired training
Christian Grillon
Yale University School of Medicine

Animal studies suggest that fear to explicit cues (e.g., a light that
has repeatedly been paired with a shock) and contextual fear (e.g., the
experimental cage) are separate processes mediated by different brain
structures.  The present study examined conditioned fear to explicit
and contextual cues with the startle reflex methodology in human
subjects.  Two groups of subjects participated in conditioning
experiments on two sessions separated by 4-5 days.  Each session
consisted of a) a startle habituation period, b) a pre-conditioning
phase, c) a conditioning phase, and d) a post-conditioning extinction
test. The unconditioned stimulus (US), an electric shock, was paired
with the conditioned stimulus (CS) in the "paired" group, but not in
the "unpaired" group, in the conditioning phase.  In the paired group,
the CS potentiated startle (fear-potentiated startle) in the
post-conditioning phase.  This conditioned response was fully retained
over the retention interval. There was no substantial change in
baseline startle in session 2 (startle delivered in the absence of
CS).  By contrast, startle was not potentiated by the CS in the
unpaired group, but baseline startle was increased from session 1 to
session 2. Hence, as predicted by conditioning theories, paired
presentations of a CS and an aversive US results in conditioned fear to
the CS but little contextual fear, whereas unpaired presentations of  a
CS and US leads to poor explicit cue conditioning, but substantial
contextual fear.

An experimental study of two forms of emotion regulation
James J. Gross
Stanford University

To empirically test the distinction between antecedent-focused and
response-focused emotion regulation, Ss (60 men and 60 women) watched a
short disgust-eliciting film while their behavioral, subjective, and
physiological responses were recorded.  Ss either were told to watch
the film (Watch condition), to think about the film in such a way that
they felt nothing (Reappraisal condition), or to behave in such a way
so that someone watching them would not know they were feeling anything
(Suppression condition).  Compared with watch instructions, both
reappraisal and suppression instructions reduced expressive behavior.
However, only reappraisal led to decreased disgust self-reports, and
only suppression led to increased signs of sympathetic nervous system
activation (including greater increases in skin conductance, greater
decreases in finger pulse amplitude, and greater decreases in finger
temperature). These results suggest the value of distinguishing among
various forms of emotion regulation, each of which may have somewhat
different consequences for adaptive functioning.

Cardiovascular autonomic characterization of the simulated
public-speaking task
Paul Grossman1 and J. Andrew Taylor2
1Lown Cardiovascular Research Laboratory, Harvard School of Public 
Health, and 2Division of 
Aging, Harvard Medical School

This study examines the autonomic response constituents of the often
used "simulated public-speaking" stressor. Cardiovascular autonomic
reactions to this task were compared with responses to simply counting
aloud at normal speaking volume and cadence (with order of tasks
balanced). Respiration, ECG and Finapres finger blood pressure were
continuously monitored, and peroneal sympathetic nerve activity (PSNA)
was recorded via microneurography among 11 healthy adults under 40
years of age. Spectral analysis of R-R interval and systolic blood
pressure was performed to evaluate periodic components of cardiac and
blood pressure variability. Cardiac vagal control was estimated from
respiratory sinus arrhythmia individually adjusted for respiration rate
and tidal volume. Baroreflex sensitivity, an index of vagal
responsivity, was derived from transfer function analysis of
blood-pressure and R-R signals.  Mayer-Wave blood-pressure
oscillations, reflecting sympathetic vasomotor activity, were defined
systolic spectral power (.07-.12 Hz). PSNA was quantified by burst
frequency and integrated activity. Simulated public speaking elicited
greater mean hemodynamic reactions than counting aloud (heart rate,
23.3 vs. 7.9 bpm; systolic pressure, 29.5 vs. 5.9 mmHg; diastolic
pressure, 21.1 vs. 7.7 mmHg; p's<0.005). Simulated public speaking
evoked large reductions in respiratory sinus arrhythmia and baroreflex
sensitivity from baseline (p's<0.004), although baseline-to-counting
responses were not significant. Mayer-Wave variability and PSNA
increased from baseline to public speaking (p's<0.0001 and <0.03,
respectively), but not from baseline to counting. These findings
elucidate the specific involvement of vagal, alpha- and beta-adrenergic
mechanisms in cardiovascular responses to simulated public speaking.

ERP variations related to time course manipulations during presentation
of consonant sequences
Kerstin Grune1, Herbert Hagendorf2, and Anna-Marie Metz1
1University Potsdam, 2Humboldt University Berlin

In recent experiments with a pseudo-random sequence of 7 consonants
to recall after trial (stimulus onset asynchrony (SOA): 1000 ms) a
reverse relationship between P300 amplitude and letter presentation
position was shown.  It was assumed that this relationship reflects a
competition between encoding and retention processes in this time
regime.  In order to disentangle processes of encoding and rehearsal in
this study the SOA between 6 consonants was prolonged to 2800 ms. The
performance level was higher compared to a control task with a SOA of
1000 ms between the letters.  This could be a first hint to lower
processing load in the long SOA condition. ERP data support the
hypothesis of a resource competition between encoding and retention
processes in the short SOA condition which is not present during the
long retention intervals: A significant relationship between P300
amplitude and letter presentation position was found only in the task
with short SOA.  The negative slow waves following P300 during the 2800
ms SOA were largest at parietal electrode sites. They are probably
associated with rehearsal processes.  Positive shifts more pronounced
in the control condition were found analyzing the brain activity across
the full presentation interval of trial.  This result is compatible to
our hypothesis of resource competition in the short SOA task.

Opposite patterns of P300 asymmetry in schizophrenia are syndrome
John Gruzelier, Jochen Kaiser, Alexandra Richardson, David Liddiard, 
Soraj Cheema, Bassant Puri, 
and Christopher McEvedy
Charing Cross and Westminster Medical School

In schizophrenia reduction of the P300 amplitude is a robust and
highly replicable finding. A series of reports further suggest that the
P300 reduction is consistently lateralised to the left temporal
region.  However, attempts at replication have yielded inconsistent
results, or evidence of asymmetries in either direction in individual
patients. Our previous investigations of functional asymmetry, which
began with bilateral electrodermal orienting responses, have shown that
opposite patterns of asymmetry in schizophrenia are associated with
different clinical syndromes which we describe as Active versus
Withdrawn.  We therefore predicted that that schizophrenic patients
would shown opposite asymmetries of the P300 according to their Active
or Withdrawn syndrome profile. A standard oddball detection task was
used to elicit auditory evoked potentials from 28 derivations in 18
dextral DSM-IV schizophrenic patients . On blind clinical rating 8 were
withdrawn and 10 had predominantly Active symptomatology. Both patient
groups had a congruent P3 maxima at Pz or P4, attesting to the
reliability of the task recording. Patients belonging to the Active
syndrome had lower overall P3 amplitude (p<0.05). Lateral asymmetries
in P3, N2-P3 and N1 were observed in the posterior temporal leads.
Those with an Active syndrome showed a left lateralised reduction in P3
and those with a Withdrawn syndrome showed a lower P3 amplitude on the
right. These findings provide some clarification of hitherto
inconsistent findings, and give further endorsement to a syndromal
approach to laterality research in schizophrenia.

Differentiation in schizophrenia of orbito-frontal from dorsolateral
functions in relation to bilateral electrodermal responses and
John Gruzelier, Julian Green and Andrew Nagy
Charing Cross and Westminster medical School

The orbitofrontal cortex in contrast to dorsolateral prefrontal (DLPF)
functions has received little attention in schizophrenia. Its role in
emotion, disinhibited behaviour and limbic/ ANS modulation suggests
relevance to the disorder and its psychophysiology. Here we examined
bilateral electrodermal orienting responses during a standard tone
habituation sequence, a procedure which previously delineated Active
and Withdrawn syndromes (Gruzelier and Manchanda, 1982). The 30 DSMIV
schizophrenic patients were administered the WCST (a popular DLPF
measure), and visual discrimination reversal (VDR), a putative OBF
task. Syndromes were assessed with a Schizophrenia Syndrome Scale
(Gruzelier) and the PANSS (Kay), rated blind to the laboratory
measures. Schizophrenic patients were deficient on both the frontal
tasks compared with normal controls matched for IQ, age and sex.  While
VDR learning was associated with WCST perseverative errors it was
unrelated to other WCST measures. The various VDR measures correlated
with EDA responsiveness (e.g., reversals:- THab r=0.  61,p<0.017; NSRs
r=0.48,p<0.03), such that the poorer the performance the lower the
reactivity.  High reactivity and better VDR performance were both
correlated with anxiety, tension, active social avoidance and
overvalued ideas, commonly associated with OBF function. Poor WCST
performance in contrast was associated with thought disorder
(r=0.42,p<0.02). A left>right EDR asymmetry replicated previous
associations with the Active syndrome including thought disorder,
hallucinatory behaviour, and perseveration.  The left>right asymmetry
was also associated with WCST perseverative errors. Distinguishing
between the frontal systems holds promise in elucidating
psychophysiological findings in schizophrenia and the
substrate of different syndromes.

Lateral asymmetry of slow potentials: Learned control and individual
John Gruzelier1, Elinor Hardman1, Kate Cheesman1, Ceri Jones1, David 
Liddiard1, Hans 
Schleichert2, and Niels Birbaumer2.
1Charing Cross and Westminster Medical School, 2Psychological 

While learned control of slow potentials is well documented, we know
of no reports of the control of lateral asymmetries generated from
bipolar recordings. Two experiments are presented. Following work at
Tubingen feedback took the form of a small rocket, placed centrally on
a VDU screen and moved in accord with discriminant stimuli. Here rocket
ascent (A-trial) indicated a leftward hemispheric shift and descent (B-
trial) a rightward shift. Subjects participated in three sessions of 60
psuedo-randomly ordered eight second trials in three blocks with a four
second intertrial interval. In experiment I using bipolar recording
from C3-C4 subjects (N=15) were assisted with a strategy to concentrate
on the contralateral arm. Subjects were surprisingly successful in
learning control as shown in a MANOVA (Session, Block, Stimulus A/B),
in which there was a main effect of Stimulus (F(1,14)=14.38, p<0.002)
without effects of Session or Block, or interactions. On A trials there
was a leftward shift (-0.69) and on B trials a rightward shift
(+1.52).  In experiment II recording was from F3 and F4 and subjects
(N=16) were divided into a strategy (positive versus negative emotion)
and no-strategy group. Asymmetry control was achieved in the third
block of trials in all sessions, where there was a main effect of
Stimulus (p<0.001).  Individual differences in attention and schizotypy
were associated with control of asymmetry:- Extremely Focused Attention
(Crawford), Absorption (Tellegen and Atkinson), Unreality (Gruzelier),
as was a rightward shift.  Neurophysiological and clinical implications
will be discussed. (Supported by a NARSAD Senior Investigator Award).

Prepulse modulation before and after pallidotomy in patients with
Parkinson's Disease
Paul Haerich1, Amy D. Clegg1, and Robert P. Iacono2
1Loma Linda University, 2School of Medicine, Loma Linda University

Parkinson's Disease is a neurodegenerative disorder in which the
nigral dopamine input to the striatum progressively declines resulting
in, inter alia, increased activity among GABAergic cells in the
posteroventral pallidum.  These pallidal neurons produce two major
projections:  the first to the thalamus which is associated with tremor
and dyskinesias, the other to the pedunculopontine tegmental nucleus
which is thought to play a role in the stiffness and bradykinesia
characteristic of Parkinson's Disease. Pallidotomy is a neurosurgical
technique in which these cells are lesioned in order to provide
reduction in the severity of parkinsonian symptoms including tremor,
rigidity, and akinesia.  Individuals undergoing therapeutic pallidotomy
were tested before (1 or 2 days) and after (5 - 8 days) surgery.
Reflex blinks were elicited with airpuffs (20 kPa, 100 ms duration)
directed 1 cm laterally to the outer canthus of the left eye.  A
vibrotactile stimulus affixed to the middle finger of the right hand
was used as a prepulse.  Each test session consisted of 60 airpuffs (12
blocks of 5 trials).  One trial in each block presented the airpuff
alone; in the others the vibrotactile stimulus was presented 200, 350,
1000, or 2000 ms prior to and terminated with the onset of airpuff.
Results suggest that following pallidotomy at both short and long lead
intervals (a) prepulse modulation of blink amplitude is significantly
reduced (p = .05), while (b) the probability of a reflex blink is
increased (p < .02).  These results confirm in humans observations in
which lesions of the striato-pallidal-pedunculopontine circuitry
decrease prepulse modulation in laboratory rats.

Anterior EEG asymmetry and facial EMG as evidence that affect is
involved in the mere exposure effect
Eddie Harmon-Jones1 and John J. B. Allen2
1University of Texas -- Arlington, and 2University of Arizona

Explanations of the mere exposure effect assume that familiar stimuli
evoke more positive and/or less negative affect than unfamiliar
stimuli.  In an experiment designed to test this assumption,
participants repeatedly viewed photographs of women's faces and then
viewed these women again (familiar) and novel women (unfamiliar) while
EMG activity in the zygomatic and corrugator muscle regions was
recorded.  Familiar stimuli were rated more positively than unfamiliar
stimuli (mere exposure effect), and they evoked more EMG activity in
the zygomatic region.  We also hypothesized that individual differences
in motivation and affect would interact with reactions to merely
exposed stimuli.  To test these hypotheses, at baseline we assessed
participants' self-reported affect, and recorded 4 min of resting EEG
at F3, F4, P3, and P4.  Asymmetrical EEG activity scores were computed
by subtracting log alpha power at F3 from log alpha power at F4, so
that higher scores reflect greater left than right activity.  Anterior
EEG activity was recorded because much research has found that the
anterior regions of the left and right hemispheres are specialized for
expression and experience of approach and withdrawal motivation,
respectively.  Individuals with relatively less left anterior
activation evidenced a greater mere exposure effect.  Individuals
reporting less positive affect and individuals reporting more negative
affect at baseline evidenced more zygomatic muscle region activity to
the familiar than to the unfamiliar.  These results implicate affect in
the mere exposure effect, and suggest mere exposure increases positive
affect rather than decreases negative affect.

Anger and prefrontal brain activity: EEG asymmetry consistent with
approach motivation despite negative affective valence
Eddie Harmon-Jones1, John J. B. Allen2, and Ernest S. Barratt3
1University of Texas -- Arlington, 2University of Arizona, 3University 
of Texas Medical Branch at 

The anterior regions of the left and right hemispheres have been
posited to be specialized for expression and experience of approach and
withdrawal processes, respectively.  Much of the evidence supporting
this hypothesis has been obtained using EEG activity, particularly
power in the alpha band (8-13 Hz), which has been found to relate
inversely to activity.  Moreover, anterior asymmetrical activity has
been posited to function as a diathesis that predisposes persons to
respond with approach or withdrawal processes given appropriate
situations.  Much research has supported these hypotheses.  However, in
most of this research, motivational direction has been confounded with
affective valence, such that, for example, approach motivation has been
associated with positive affect.  In the present research, we tested
the idea that dispositional anger, an approach-related motivational
tendency with negative valence, would be associated with increased left
anterior activity.  Adolescents' EEG during 3 min of eyes-open and 3
min of eyes- closed resting periods was recorded.  Frontal and parietal
asymmetrical activity (log F4 - log F3, log P4 - log P3, so that higher
scores reflect greater left than right activity) were examined in
relation to scores on the Buss and Perry (1992, Journal of Personality
and Social Psychology, 452- 459) aggression questionnaire.  The
anterior asymmetry correlated positively and significantly with trait
anger, suggesting that increased relative left frontal activity was
associated with increased trait anger.  No significant relationships
with parietal asymmetry occurred.  These results suggest that the
anterior asymmetry varies as a function of motivational direction
rather than affective valence.

Implicit memory bias for threat: A state manipulation
Lesley K. Harrison and Graham Turpin
University of Sheffield

Previous research has revealed equivocal findings concerning a threat
bias in implicit memory. This study sought to demonstrate such an
effect by exposing high and low trait anxious students to a state
anxiety manipulation using examinations. Hypotheses proposed that
elevated state anxiety would increase recall for threatening
information. In addition, psychophysiological responding to threat and
non-threat words during the priming phase would relate to performance
on a subsequent implicit memory task.  Subjects (N=40) were allocated
to two equal groups; each tested during an exam and non-exam period.
All subjects viewed words (20 threat, 20 non-threat), presented for a
duration of 3s, with an ITI of 25s to 35s. A filler task and word-stem
completion task followed the priming phase. Heart rate, skin
conductance and respiration were measured throughout the experiment
using a Biopac system. High and low trait anxious subjects were equally
allocated to each group.  State anxiety was significantly elevated
during the exam test session (p<.002). Analysis of performance data
revealed significant effects for priming (p<.0001) and threat
(p<.0001). Higher order interactions between trait anxiety, threat and
exam status were also obtained (p<.0001 to p<.02). Repeated measures
ANOVA for heart rate indicated a stable response profile (p<.0001) and
significant interactions with seconds were obtained between exam
status, threat, wordlist (p<.04) and also between threat, trial blocks,
trait, wordlist and exam order (p<.02). The results are discussed in
relation to current models of anxiety and implicit memory.

The association of age and aerobic activity history with 40-Hz EEG
activity during cognitive challenge
B. Hatfield, D. Santa Maria, T. Spalding, C. Blanchard, A. Haufler, T. 
Hung, S. Kerick, P. Lockwood, 
L. McAllister, P. Saarela, R. Apparies, J. Lanter, and K. McDowell
University of Maryland at College Park

EEG 40-Hz activity was obtained from four fitness-age groups during
eyes-closed rest (EC), mathematical (MA) and visuospatial (VS)
processing to determine the interactive effect of age and aerobic
fitness upon neurocognitive effort.  Because fitness has been shown to
reduce age-related declines in CNS function, we proposed that fitness
would be associated with reduced spectral power, especially for older
subjects relative to age-matched sedentary subjects.  Subjects (88
males and females) were classified into four groups consisting of 22
low-fit, younger (M=23.2 yrs, mean daily aerobic expenditure=48.9 Kcal,
as qualified by an intensity to promote fitness), 21 high-fit, younger
(M=22.3 yrs, M=964.0 Kcal), 22 low-fit, older (M=69.7, M=36.0 Kcal),
and 23 high-fit, older subjects (M=65.7 yrs, M=551.6 Kcal) and were
monitored for 3-min periods during each task while EEG was recorded
from T3 and T4, referenced to linked ears, with low- and high-pass
filter settings of 1 and 100 Hz, respectively, amplified 100K, and
sampled at 240 Hz, continuously.  EOG-corrected epochs of 512 points
were subjected to spectral analysis to examine 36-44 Hz average band
power (i.e., log transformed).  A 2x2x2x2 (age x fitness x gender x
site) ANOVA revealed significant age-x-fitness interactions for both MA
and VS, such that the low-fit, older subjects exhibited greater power
than the three other groups, while no such effect was observed during
EC.  We conclude that a lack of fitness in older subjects is associated
with increased effort during neurocognitive processing relative to
younger and older physically fit individuals which is revealed during

Cardiovascular reactivity to sensory intake rather than working memory
tasks differentiates offspring of normotensives and hypertensives
Louise C. Hawkley1, John M. Ernst1, Gary G. Berntson1, Stephen M. 
Kosslyn2, and John T. 
1Ohio State University, 2Harvard University

Studies of cardiovascular reactivity in offspring of normotensive and
hypertensive parents have emphasized tasks that require mental
reasoning and have produced mixed results. In the present research, we
administered the Harvard cognitive (deductive reasoning, vigilance,
image rotation, and attention shifting) battery to normotensive
undergraduate offspring of normotensive (n = 12) or hypertensive
parents (n = 13).  Results revealed a significant Group x Task
interaction.  As in prior research, the offspring of normotensive
parents showed increased heart rate to the reasoning task, mild heart
rate deceleration to the tasks that emphasized sensory intake
(vigilance, attention shifting), and no effect for the task with both
components (image rotation).  Offspring of hypertensive parents also
showed tachycardia to the reasoning task and no effect for the image
rotation task, but showed tachycardia rather than bradycardia to the
challenging attentional tasks.  These results are potentially important
for several reasons.  First, the autonomic responses to attentional
tasks may be a factor in the etiology of hypertension, or at least,
serve as a marker for the development of hypertension.  Second, these
results suggest an alteration in the autonomic control at the brainstem
and possible changes in more rostral brain regions.  Third, it may help
account for the inconsistencies in the literature, which have tended to
rely on tasks that elevate heart rate in subjects with and without a
family history of hypertension.  This research suggests that the groups
may be more clearly discriminated using challenges that alter vagal
control and reduce heart rate in normals.

Impaired attentional modulation of startle in unmedicated
Erin A. Hazlett, Monte S. Buchsbaum, M. Mehmet Haznedar, Melissa Biren, 
and David B.Schnur
Mount Sinai School of Medicine

	Previous research has shown that in healthy normal individuals,
the startle reflex (SR) elicited by abrupt stimuli can be modified by
attention to nonstartling stimuli that shortly precede the
startle-eliciting stimulus. In contrast, medicated schizophrenic
patients fail to show both short and long lead-interval attentional
modulation of the SR. The present study examined SR modulation in
unmedicated schizophrenic patients (n=13) and age- and sex-matched
normal individuals (n=12). Participants performed a selective attention
task involving presentation of to-be-attended and to-be-ignored tones
which served as prepulses. Acoustic startle probes were presented at
lead-intervals of 120, 240, and 4500 ms following onset of tones and
occasionally during the intertone interval.  In addition, probes were
presented at either the 120 or 4500 ms lead-interval following onset of
novel tones.  The controls showed significantly greater SR modulation
during attended tones compared to ignored tones at the 120 and 4500 ms
lead-intervals. In contrast, the schizophrenics failed to show
differential SR modulation.  Further, in controls, SR modulation during
the novel tones was in-between that of the attended and ignored tones
and did not differ significantly from either, whereas, the
schizophrenics showed greater SR modulation during the novel tones at
4500 ms than during either the attended or ignored tones.  Correlations
between SR modulation and simultaneously obtained regional brain
glucose metabolism from postiron emission tomography scans will also be

The effects of the presence of another person on male subjects'
cardiovascular response to stress
Suzanne G. Helfer and Margret A. Appel
Ohio University

Epidemiological studies indicate that adequate levels of social
support are related to lower rates of cardiovascular and all-cause
mortality.  One mechanism of action of social support may be
attenuation of the cardiovascular stress response while in the presence
of a friend.  A laboratory model of social support has been devised in
which subjects have a friend with them while they perform a stressful
task.  In several previous studies, subjects who had a friend or
supportive confederate present had attenuated cardiovascular reactivity
to stress compared to subjects who were alone.  Much of the research on
this phenomenon has used female subjects.  The purpose of the present
study was to explore the reaction of male subjects to the social
support manipulation.  Sixty-three male undergraduates performed serial
subtraction in the presence of a friend, a stranger, or alone.  Heart
rate, systolic blood pressure, and diastolic blood pressure were
monitored.  One-way ANOVAs were performed to explore group differences
in response to the task.  Despite the fact that the task elicited
cardiovascular and emotional reactivity, no significant differences
were found among the groups as a function of the social support
manipulation for any cardiovascular measure.  This study constrasts to
previous studies in our laboratory using similar manipulations where we
found that females respond to the presence of another person with
attenuated cardiovascular reactivity. The present study suggests that
in males, on the other hand, the reactivity attenuating effect of the
presence of another person may be weak or absent.

Does watching happy expressions make us happy? Facial mimicry and
emotional contagion to dynamic emotional facial expressions.
Ursula Hess and Sylvie Blairy
University of Quebec at Montreal

In recent years, evidence has accumulated suggesting that individuals
exposed to the emotional facial expressions of others tend (a) under
certain circumstances to show congruent facial reactions (facial
mimicry), and (b) experience congruent emotional states (emotional
contagion).  However, most research in this context has employed highly
prototypical, non dynamic, posed stimuli with high recognition rates.
The present study had the goal to assess whether individuals asked to
decode a series of video clips showing more ecologically valid,
spontaneous, idiosyncratic, emotional facial expressions which were
recorded during an emotional imagery task, would also show evidence of
facial mimicry and emotional contagion.  Twenty subjects were asked to
decode 4 expressions each of happiness, anger, disgust, and sadness.
Dependent measures were IEMG of the Corrugator Supercilii, Orbicularis
Oculi, and Levator Labii Alesque Nasii sites as well as self-reported
levels of happiness, anger, disgust, and sadness which were obtained
under a pretext following the judgment of one expression of each
emotion.  While the overall level of rating accuracy is low, the
results nonetheless suggest that individuals mimic facial expressions
of happiness.  For these expressions significantly lower levels of
Corrugator S. activity as well as significantly higher levels of O.
Oculi activity were found.  Further, participants indicated
significantly higher levels of happiness after having been exposed to
an expressions of happiness than after being exposed to any other
expression, as well as significantly higher levels of sadness after
having been exposed to an expression of sadness then after having been
exposed to any other expression.

Stimulus movement and peripheral stimulus localization by 8-, 14-, 20-,
and 26-week-old infants
Julie M. Hicks and John E. Richards
University of South Carolina

Peripheral stimulus detection in infants occurs infrequently during
sustained attention to a central stimulus, and occurs more frequently
when no central stimulus is present, or during attention termination
(Richards, in press).  Dynamic peripheral stimuli may be localized more
quickly and more often than static peripheral stimuli because brain
cells/systems in the peripheral visual field are more sensitive to
movement than form.  The current study examined the effect of stimulus
movement on localization probability and latency during attention and
inattention.  Forty infants, 10 each at 8, 14, 20, and 26 weeks of age
were presented with a central stimulus.  Then, a peripheral stimulus
was presented (static or dynamic checkerboard) during various
heart-rate defined attention phases.  Control trials, which consisted
of the presentation of the central stimulus alone, were used to
estimate false alarm and correct rejection rates.  The dynamic stimulus
was localized more frequently than the static stimulus when the central
and peripheral stimulus were present simultaneously and during
sustained attention.  There was no difference in localization
probability during attention termination.  The latency results showed
that the static stimulus was localized more quickly than the dynamic
stimulus. Older infants localized the peripheral stimulus faster than
younger infants.  The results suggest that a dynamic peripheral
stimulus may partially overcome the effects of attention to a central

Mismatch negativity in schizophrenia using 64 channel geodesic sensor
Yoshio Hirayasu, Brian F. O'Donnell, Geoffrey Potts, James Levitt, 
Hajime Arakaki, Sare J. Akdag, 
and Robert W. McCarley
Harvard Medical School

The auditory mismatch negativity (MMN) is an ERP related to the
detection of novel stimuli, which occurs at short latency (100 - 200
ms), and does not depend on active attention. Reduction of MMN
amplitude has been reported in schizophrenics. MMN has been localized
within, or near, Heschl's gyrus in the superior temporal gyrus, which
often shows anatomic abnormalities in schizophrenia. In this study, MMN
were recorded with a 64 channel Geodesic Sensor Net for better spatial
localization to test whether patients show reduced MMN amplitude or an
asymmetrical distribution. We evaluated MMN in right-handed, male
patients with DSM-III-R diagnosed schizophrenia (N =3D 10) and male
controls (N =3D 6). Deviant stimuli (1.2 K Hz,5%), and standard stimuli
(1.0 KHz) were presented with a mean inter-stimulus interval of 300
msec, and 100 msec duration. The subjects were instructed to ignore
presented stimuli. For MMN, difference waveforms were constructed by
subtracting the averaged response to standard from the averaged
response to the deviants. Amplitude was measured from a 50 msec
integration window between 120 - 170 msec. MMN amplitude at Fz in
schizophrenics (-1.50=B12.43 =B5V) was significantly smaller than in
controls (-6.12=B14.18 =B5V) [F(1,14)=3D7.99, P=3D0.014]. Significant
amplitude reduction in schizophrenics was also found at Cz and Pz, but
voltages at left and right temporal sites did not show a significant
asymmetry between groups. These findings suggest that the early stages
of information processing at the level of the auditory cortex are
impaired in schizophrenia.

Smoking and heart rate
Michael E. Houlihan, Walter S. Pritchard, and John H. Robinson
Bowman Gray School of Medicine and R. J. Reynolds Tobacco Co.

Smoking following overnight abstention is reported to reliably
increase heart rate (HR), an effect due to nicotine absorption.  The
effect of subsequent cigarettes on HR is much less, an indication of
tachyphylaxis (acute tolerance).  Additional support for tachyphylaxis
of the HR response to nicotine has come from studies using a variety of
non-smoking methods of delivering nicotine (iv, sc, gum, nasal spray,
transdermal patch) to both smokers and nonsmokers.  However, smoking/HR
studies to date have not been conducted double blind. Instead, control
conditions have included non-smoking or some type of "sham" smoking
(puffing on an unlit cigarette or a straw).  We observed HR changes in
overnight abstaining subjects following smoking of the first, second
and third cigarette of the day (40 min. between each cigarette) on two
occasions.  Our control condition involved replacing the second
cigarette with a cigarette with very low nicotine yield (0.05-mg)
compared with the regular cigarette with a nicotine yield of 1.1-mg.
The increase in HR after each of the three regular cigarettes was 14, 9
and 8 beats per min. (BPM), indicating some tachyphylaxis.  The rise in
HR during the session with the second cigarette as a control was 16, 1
and 12 BPM, indicating that a period greater than 80 min. would be
needed before the HR response to smoking was comparable to that
observed for the first cigarette of the day.

Effects of quantified cigarette-smoke delivery on EEG linear and
nonlinear dynamics during a stimulus-response compatibility task
Michael E. Houlihan1, Walter S. Pritchard2,1, John H. Robinson2,1
1Bowman Gray School of Medicine, 2R. J. Reynolds Tobacco Co.

Under both resting as well as relatively easy task conditions (the
latter visual and auditory go/no-go oddballs), we have found that
smoking has a flexible effect on EEG dynamics as indexed by dimensional
complexity (DCx, estimated correlation dimension).  Specifically,
smoking [1] raises DCx when baseline levels are low, [2] does not
affect it when baseline levels are intermediate, and [3] lowers it when
baseline levels are high.  Fifteen male subjects participated in two
counterbalanced, double-blind sessions, during which they performed a
rather difficult stimulus-response compatibility RT task before and
twice after smoking using a Quantified Smoke Delivery System (12 2-s,
35 cc puffs).  In one session, the cigarette had a nicotine yield of
1.1 mg and in the other a 0.05-mg yield.  The task crossed
unmasked/masked (visual +snow+) stimuli with a stimulus location (left
or right of screen) either compatible or incompatible with a left- or
right-hand button press.  EEG was recorded during task performance from
Fz, Cz, Pz, Oz, T5 and T6.  Quantified smoking of the 1.1-mg cigarette
(relative to the 0.05-mg control) increased accuracy without affecting
RT.  Rather than having the flexible effect discussed above, the 1.1-mg
cigarette globally decreased DCx.  It also globally increased delta
power (no other spectral-band effects were obtained).  These unexpected
results indicate, under more difficult task conditions, quantified
smoking of a nicotine-yielding cigarette can improve accuracy by moving
cortical dynamics into a more simplified state characterized by a
longer EEG autocorrelation time.  Finally, surrogate-data testing
(Gaussian amplitude-adjustment procedure) indicated significant
nonlinearity that was not affected by quantified smoking of the 1.1-mg

Retention of acquired adaptation to vection-induced motion sickness
Senqi Hu, Todd S. Hoffman, Katherine M. Glaser, and Tamera M. Stanton
Humboldt State University

This study investigated the length of the retention of adaptation to
vection-induced motion sickness.  Thirteen subjects who had a history
of motion sickness participated in the study.  The subjects repeatedly
viewed an optokinetic rotating drum for 16 minutes in two day intervals
until they had no feelings of nausea.  Subjects' ratings of nausea and
electrogastrograms (EGGs) at 4-9 cycles per minute (cpm) were obtained
for each session.  The subjects were reexposed to the optokinetic drum
after waiting one of the following waiting periods: one month (2
subjects), three months (5 subjects), six months (3 subjects), and 12
months (3 subjects).  Repeated measures of analysis of variance on
nausea ratings for the session of reexposure to the optokinetic drum
revealed significant differences among subjects of the four waiting
periods, F(3, 9) = 16.05, P < 0.01.  The subjects who were reexposed to
the optokinetic drum after waiting 12 months had significantly higher
mean nausea ratings than those after waiting 1 or 3 months.  One-way
analysis of variance on mean ratios of EGG 4-9 cpm spectral intensity
for the session of reexposure to the optokinetic drum indicated
significant differences among the four waiting periods, F(3, 9) = 5.61,
P < 0.02.  The subjects who were reexposed to the optokinetic rotation
after waiting 12 months had significantly higher mean ratios of EGG 4-9
cpm spectral intensity than those after waiting 1 or 3 months.  These
results indicated that acquired adaptation to vection-induced motion
sickness was retained for up to three months.

Psychophysiology research and career development:  Funding
opportunities at NIMH"

Lynne C. Huffman
National Institute of Mental Health)

Abstract:  The National Institute of Mental Health actively supports
research pertinent to psychophysiology as it is related to basic
behavioral and social processes as well as clinical issues.  Support is
available in the form of grants for research, training, and career
development, and several types of grants are specifically directed
toward junior investigators.  This poster presentation will provide the
opportunity to learn more about these grants as well as obtain practical
information that will maximize chances of funding.

The perseveration of response preparation during memory scanning
Aaron B. Ilan1 and Jeff Miller2
1University of California, San Diego, 2University of Otago, Dunedin, New 

Previous data (Ilan & Miller, in preparation) suggest that motoric
response preparation cannot be initiated in parallel with memory
scanning, but leave open the possibility that it may continue in
parallel with memory scanning once it has been initiated.  Thus, the
two experiments presented here investigated this distinction by
examining whether motor preparation can be maintained or even
perseverated while memory scanning is in progress.
	Lateralized readiness potentials (LRPs) were computed in a
modified memory scanning task.  Each trial began with the presentation
of a colored precue that indicated which hand might be needed to
respond.  The memory set membership of a letter appearing 1s later
determined whether the primed response should be made or withheld.  The
results indicated that motor preparation, as is indexed by the LRP, was
initiated in response to the precue before the letter appeared.  More
importantly, LRP amplitude continued to increase after the letter was
presented, suggesting that once it had been initiated, motor
preparation was actively perseverated in parallel with memory
	Experiment 2 investigated whether the rate at which LRP
amplitude increases depends on the difficulty of the concurrent
task.  The S-R mapping in the memory scanning task was changed from
varied to consistent to make the task less cognitively demanding.  LRPs
were also computed in an easier, more perceptual, size discrimination
task.  The findings of Experiment 1 were replicated, and comparisons
across experiments suggested that the difficulty of the concurrent
process has little effect on the perseveration of motor preparation.

The effects of smoking on ERPs in simple and demanding tasks
Aaron B. Ilan1 and John Polich2
The Scripps Research Institute

After abstaining from nicotine, cigarette smokers performed auditory
and visual oddball tasks as well as Stroop color naming and memory
scanning tasks.  After smoking two cigarettes, subjects were run
through these paradigms again.  Behavioral and electrophysiological
measures before and after smoking were contrasted to evaluate the
cognitive effects of smoking.  In the auditory oddball task, smoking
had a negligible effect on both behavioral and electrophysiological
measures.  In the visual oddball task, P300 amplitude tended to
decrease after smoking.  Both Stroop facilitation and interference were
observed on RTs in the color naming task. P300 latency, however, was
shortest in the incongruent word condition, suggesting that the locus
of the Stroop effect is relatively late in the information processing
sequence.  Additionally, P300 amplitudes to both neutral and
incongruent color words were diminished after subjects smoked.  In the
memory scanning task, smoking had little effect on ERPs when a small
set size was used, but amplitudes of all components tended to decrease
after smoking when a large set size was used.  The scanning rate on RT
was somewhat slower after smoking.  In general, the effects of smoking
on ERPs depended on the sensory modality and attentional demands of the
task.  Smoking had no effect in a simple auditory oddball paradigm, but
component amplitudes were generally diminished after smoking in the
three visual tasks: visual oddball, Stroop color naming, and memory

Single-trial P300 analysis: Evidence for a graded context updating
Tiffany A. Ito and John T. Cacioppo
The Ohio State University

The amplitude of the P300 in the auditory oddball paradigm is larger
to rare than to frequent stimuli.  This finding has led to proposals
that relate the P300 to context updating processes.  By ensemble
averaging across trials, however, two possible processes by which
context updating is achieved are indistinguishable.  In one, frequent
and rare stimuli trigger updating operations but these operations are
more extensive or intense for unexpected than expected stimuli;
accordingly, the P300 amplitude is smaller for frequent than rare
stimuli.  It is possible, however, that context updating unfolds in an
all-or-none fashion but that frequent stimuli are simply less likely to
trigger updating.  To distinguish between these two hypotheses, a
single-trial method for analyzing ERPs was developed and applied to the
midline EEG data obtained in a standard oddball paradigm (N = 11).
Ensemble averaged ERP data replicated prior research showing a P300
latency of 350 ms, maximal amplitude over centroparietal regions, and
larger amplitude to rare than frequent stimuli.  A single-trial
magnitude analysis including all trials (comparable to ensemble
averages but with peak alignment) yielded the same results, confirming
the comparability of the methods.  Where the two models of context
updating make differing predictions is in the amplitude analysis, in
which trials without a detectable P300 are excluded.  Results revealed
fewer detectable P300s elicited by frequent than rare stimuli but the
P300 amplitude was nevertheless greater for rare than frequent
stimuli.  These results rule out the notion that context updating
processes, once triggered, unfold in an all-or-none fashion.

The effect of sleep deficit, stimulus degradation, and knowledge of
results on response force in simple and choice reaction task
Piotr Jaskowski and Dariusz Wlodarczyk
Medical University of Luebeck, Germany; Medical Academy of Poznan, Polen

Some recent findings support the idea that response force (RF)
measured during reaction time experiments might reflect changes of the
level of activation.  Indeed, some variables which are assumed to
affect subjects' energetical resources, like arousal caused by
occasional delivering of electrical shocks or high stimulus
intensities, evoke more forceful responses.  We performed an experiment
in which the effect of sleep deprivation (SD), knowledge of results
(KR) and stimulus quality (SQ) on response force is studied in simple
and choice reaction tasks.  As expected both simple (SRT) and choice
(CRT) reaction times increase with sleep deficit. Knowledge of results
shortens SRT and CRT while they are impaired by degradation.  The
letter was suprisingly found also in case of SRT, although much less
than for CRT.  Based on Sanders' (1983) conclusion that sleep
deprivation affects both arousal and activation we expected detrimental
effect of sleep on force amplitude. On the other hand, KR was expected
to enlarge force by its compensatory effect on arousal. No effect of
sleep deprivation on response force was found. KR increased response
force independently of sleep deprivation. Surprisingly, force was found
to be larger for degraded stimuli in the choice reaction task, but not
in simple one, and this effect was independent of sleep deprivation.
This paper was supported for a grant from State Commitee for Scientific
Research (KBN)]

EEG coherence: Effects of sex, hemisphere, and pubertal timing
Jochen Kaiser and John H. Gruzelier
Charing Cross and Westminster Medical School

The selective elimination of a substantial amount of cortical synapses
constitutes the final step of brain maturation which coincides with
puberty. It has been suggested that the number of pruned connections
depends on the rate of maturation at puberty, i.e. pruning might be
prolonged in late maturers resulting in low synaptic density. Extremes
in the normal variation of synaptic density have been implicated in the
aetiology of schizophrenia. EEG coherence, a measure of
cortico-cortical connectivity, was used as a non-invasive tool for the
investigation of interindividual differences in synaptic density.  We
expected higher short-distance coherence with lower synaptic density
because local differentiation should be reduced. In N = 38 healthy
adults we recorded EEG from 12 bipolar electrodes during phases of rest
and photic stimulation at various flicker frequencies. Interhemispheric
coherence did not vary as a function of age at puberty or sex. In
contrast, both short- and long-distance intrahemispheric coherence was
higher for late than for early maturers. This effect was more
pronounced during flicker stimulation, for the male subsample and in
the left hemisphere. In general, males tended to have higher coherence
than females and coherence was higher in the left hemisphere than the
right. These findings are interpreted as reflecting underlying
differences in maturational rate which is slower in males than females
and in the left than the right hemisphere.

The effect of sociality on event related facial activity in the context
of a video game
Arvid Kappas and Anna Pecchinenda
Laval University

The present study investigated the influence of sociality on facial
EMG at three sites (Zygomaticus Major, Corrugator Supercilii,
Masseter), skin conductance, heart rate, and self-report in the context
of a pacman-type video-game. Sociality was manipulated by having
subjects play several game episodes either alone, or with the belief
that either a friend or a stranger was observing them via a camera
system. Sixty female subjects were randomly assigned to the three
groups. The manipulation of social context significantly influenced
facial activity to specific game events (being bitten by a monster,
biting a monster, reaching a power pill), as well as electrodermal
activity, and self report. Heart rate was only influenced by parameters
of the video-game and by repetition. Contrary to previous studies using
mostly funny films as stimuli, the results indicated that in response
to specific events (negative or positive), subjects who thought they
were observed by a friend showed less overall activity than those in
the two other conditions. This finding was observed for all three
muscle sites and for all three event types. Inspection of facial
activation over time suggests, however, that groups differed in
specificity of facial activity, with subjects observed by a friend
showing an inhibition of activation in the absence of significant
events and clearly demarcated reactions in their presence. Thus, it
seems that subjects were more clearly communicating their responses to
ongoing events when they thought they were observed by a friend.

Neurophysiological mapping of primary motor cortex:
Microstimulation-triggered averaging of EMG activity in awake
performing rhesus macaques.
Jennifer Hill Karrer1, Brian J. McKiernan2, and Paul D. Cheney2.
1 University of Kansas, 2 University of Kansas Medical Center

The motor homunculus has served as the fundamental map of the primary
motor cortex (MI) since the 1930's.  However, using repetitive ICMS
(intracranial microstimulation with high frequency trains) to evoke
movements, others have reported significant areas of overlap in the
motor homunculus.  The goal of this study was to apply a more sensitive
output mapping method, microstimulation-triggered averaging of EMG
activity, to identify MI zones representing forelimb muscles.  Two
rhesus macaques were trained to perform two motor tasks: 1) a timed
reaching and prehension task, and 2) a targeted push/pull task.
Averages of rectified EMG activity were computed from individual
microstimuli (10-20 uA) delivered to MI in penetrations placed a 0.5 mm
spacing.  EMG activity was recorded from a total of 22 forelimb muscles
(5 shoulder, 6 elbow, 4 wrist, 5 forearm digit, 2 intrinsics).  Both
excitatory and inhibitory effects were mapped. 1728 stimulation sites
were employed to compile the MI output maps.  Consistent features were
observed in both macaques.  A primary distal output zone formed a
central core that was surrounded medially, rostrally, and laterally by
a horseshoe shaped proximal muscle representation.  The lateral arm of
the proximal muscle representation showed extensive overlap with the
primary distal muscle representation.  This study not only provides a
new model of how MI is arranged, but also some indication of the
functioning of MI during complex, multi-joint movements.

Smooth pursuit eye tracking heritability in monozygotic and dizygotic
Joanna Katsanis, William G. Iacono, Mark Harris, and Micah Hammer
University of Minnesota

The present study examined the heritability of smooth pursuit eye
tracking  by assessing the smooth pursuit eye movements of 30
monozygotic (MZ) and 30 dizygotic (DZ) healthy twin pairs. Subjects
were 17 and 18 year old males drawn from an epidemiological
investigation (the Minnesota Twin Family Study). Infrared recording was
used to monitor horizontal smooth pursuit tracking while subjects
observed a luminous spot that moved with a sinusoidal motion at a
frequency of .4 Hz for 18 cycles. The electrooculogram was used  to
monitor vertical eye movements with sensors placed on the outer canthus
and above one eye. A comprehensive assessment of eye tracking was
carried out using both global (gain, RMS error, total amount of smooth
pursuit time) and specific measures (frequency, amplitude and velocity
of different types of saccades). Overall, the within-pair similarity
for the various ocular motor measures was greater for the MZ than the
DZ twins.  These findings suggest that eye tracking performance is
under partial genetic control and support the notion that measures of
smooth pursuit eye movements have the potential to identify genetic
risk for certain types of psychopathology in late adolescence.

Use of affect-toned odours to modulate acoustic startle reflex
Hossein Kaviani, Glenn D. Wilson, Stuart A. Checkley, and Jeffrey A. 
Institute of Psychiatry, University of London

Acoustic startle reflexes, as indexed by eyeblink activity, were
investigated in 24 Ss (9 men and 15 women) under conditions of pleasant
odour, unpleasant odour and no-odour, variably ordered. Affective
ratings of these odour categories confirmed that emotional state was
effectively manipulated by the smell stimuli, using MANOVA, (P < .001)
and paired t-test (for pleasant \ unpleasant differences from neutral),
(P < .001). Raw amplitude, logarithmic transform, percentage of mean
amplitude during neutral conditions, and startle magnitude (probability
x amplitude) were statistically calculated. The results were not
essentially different. For this reason, only the raw amplitudes are
presented. Startle reflexes were modulated in the expected direction -
unpleasant odours enhancing startle amplitudes and pleasant odours
reducing them relative to the no-odour condition, confirmed by  MANOVA,
(P < .001) and paired t-test (for pleasant \ unpleasant differences
from neutral), (P < .01). The role of odour hedonics in startle
modulation was further confirmed by the finding of a correlation
between rated unpleasantness of the odour and the eyeblink amplitude
while being exposed to it (  =  .71, P = .003). Sex differences were
not statistically significant across rating scores, (P = .89) as well
as startle amplitude, (P = .61). These findings support the theory of
Lang and associates that the startle reflex is an objective measure of
emotional state.

Modulation of the acoustic startle reflex by emotionally-toned
Hossein Kaviani, Jeffrey A. Gray, Stuart A. Checkley, Veena Kumari, 
Philip J. Corr and Glenn D. 
Institute of Psychiatry, University of London

It has been widely reported that the eyeblink component of the
acoustic startle reflex can be modulated by emotionally-toned slide
stimuli; pleasant slides reduce eyeblink amplitudes whereas unpleasant
slides enhance them. However, the effect is weak and the procedure
labourious. We describe the development of a video-film stimulus
procedure which is briefer and produces more robust affective
Two parallel film-sets were selected on the basis of affect
ratings, each containing equal numbers of pleasant, neutral and
unpleasant film clips  (n = 9). 22 Ss (11 men & 11women )were tested.
There were main effects of Valence across affective ratings and
startle amplitude in both film-sets (p < .001), confirmed by
MANOVA with repeated measures.  No sex effect was observed (p >.05), on
any of the dependent measures. EMG activity of orbicular oculi
correlated negatively with ratings of affect, Spearman rho (5) = -.89 ,
P < .05 (film-set 1)  & rho (5) = -0.94, P < .01 (film-set 2). Film
clips were evaluated with respect to their reliability (test-retest and
parallel form) in modulating acoustic startle reflexes to white
Startle reflex modulation effects using these film stimuli were
sufficiently powerful and consistent to indicate that the
procedure is promising as a measure of emotional states with
application to personality differences, psychotropic drug effects and
treatment outcome studies.

Dissociation of ERP topographies for verbal and nonverbal auditory
oddball tasks using principal components analysis
Jurgen Kayser, Craig Tenke, Jennifer Watson, and Gerard Bruder
New York State Psychiatric Institute

ERP asymmetries were examined in twelve normal right-handed subjects
during oddball tasks using binaurally presented verbal (consonant-vowel
syllables) and nonverbal (complex tones) stimuli (20% targets).
Response hand was counterbalanced across subjects. ERPs were recorded
from 30 scalp electrodes. A principal components analysis with varimax
rotation performed on ERP waveforms revealed five factors (92.2%
variance) identified as N1 (peak latency 100 ms), P2/N2 (200 ms), early
P3 (300 ms), late P3 (430 ms), and slow wave (500-1000 ms). Factor
P2/N2 loaded on P2 for nontargets and on N2 for targets. Repeated
measures ANOVAs of factor scores showed enhanced N2, early P3, and late
P3 for targets compared to nontargets, but not N1 or slow wave. For
targets, significant task (verbal/nonverbal) x hemisphere interactions
were found for N2, early P3, and late P3. In the nonverbal task, N2 was
maximal over the right hemisphere, particularly over lateral-temporal
regions. For the verbal task, N2 was maximal over the left hemisphere,
particularly over lateral-parietal regions. Early P3 was present only
in the nonverbal task, while late P3 was present in both oddball tasks.
Early and late P3 components both inverted frontally and were largest
at Pz. Early P3 was more pronounced over the right hemisphere (greatest
asymmetry at medial-central sites), while late P3 was more pronounced
over the left hemisphere, particularly for the verbal task (greatest
asymmetry at parietal-temporal sites). These distinct, asymmetric ERP
topographies presumably reflect differential involvement of cortical
structures in identification of phonemes (left parieto-temporal) and
complex tones (right anterior-temporal).

Effects of text connectivity on colour-cued selective attention
event-related potentials (ERPs)
Marion Kellenbach and Patricia Michie
1Groningen University, 2University of Western Australia

Recent evidence has suggested ERPs associated with colour-cued
selective attention may be modulated by language variables.  In
contrast to the selection negativity previously reported in colour-cued
selection studies, a long duration midlatency positivity associated
with attention (selection positivity) was recently reported in a
colour-cued selective attention paradigm utilising connected text
stimuli (Nobre and McCarthy, 1987, Society for Neuroscience Abstracts,
13, pp. 852).  Three experiments investigated this polarity reversal of
colour selective attention effects on ERPs by systematically varying
features of the Nobre paradigm, including the connectivity of the text,
type of target stimuli, type of response, inclusion of function words
and word repetitions, and stimulus onset asynchrony (SOA).  The
selection positivity was reliably elicited, and shown to be comprised
of several overlapping subcomponents distinguishable on the basis of
latency, scalp distribution, and the independent influences of text
connectivity and SOA variables.  In addition, an unprecedented colour-
selection enhancement of the exogenous P1 component, previously thought
to form part of the ERP signature unique to spatial selection, was also
observed.  Elicitation of this early effect was shown to be sensitive
to task demands as defined by response type.  The implications of these
findings for the temporal and neuroanatomical organisation of cognitive
processes are considered, and the data are interpreted within the
context of a parallel, interactive attentional network under the
control of an executive attentional system.

Reactivity and resiliency in older adults
Arlene R. King, Marcia Taborga, and Robert W. Levenson
University of California, Berkeley

Significant relations between life stress and disease have been
reported frequently. However, effect sizes are typically small and
individual differences abound. We used these individual differences to
operationalize the construct of "resiliency" as the likelihood of
remaining in good health given high levels of life stress. We
hypothesized that people who show such resiliency will also be less
physiologically reactive to more acute stressors.
	To increase variability in health outcomes, we studied an older
sample than is typical for this research. Stressful life events (over
the past year) and illness symptoms were assessed by questionnaire.
Resiliency was operationalized as the residual scores when illness data
were regressed on life stress data, thus indicating whether the person
had more illness (low resiliency) or less illness (high resiliency)
than would be expected based on the amount of life stress experienced.
	Physiological (cardiovascular, electrodermal, somatic)
	responses of 58 older married couples (average age:
husbands=64, wives=62) were studied in the laboratory during a
15-minute discussion of a mutually-distressing relationship issue.
Reactivity scores were calculated by subtracting the average
physiological levels during a five-minute silent pre-interaction period
from the average levels during the 15-minute discussion.
	Results indicated that resilient individuals were less
cardiovascularly reactive to the stressful interaction (smaller heart
rate increases, less shortening of pulse transmission times to the ear
and finger). We conclude that lower physiological reactivity to acute
interpersonal stress may help explain why some older people are more
resilient to life stress-related diseases.

Dimensional complexity of the EEG in schizophrenics under cognitive
challenge: Differences to normal subjects.
Peter Kirsch, Christoph Besthorn, Jochen Rindfleisch, and Robert Olbrich
Central Institute of Mental Health

The calculation of the fractal dimension is a rather new and rapidly
growing approach of analyzing the human EEG basing on chaos theory or
the theory of non-linear systems.  It was demonstrated that the
complexity of the EEG signal differs in dependence of the psychological
state of the organism.  While the complexity of the signals decreases
with the increasing depth of sleep, an increase of the dimension was
found under task compared to resting conditions.  Analysis of the EEG
in schizophrenics showed an increased dimensional complexity under
resting conditions.   These results were interpreted as reflecting the
looseness of thoughts often found in these patients.  The aim of the
present study was to investigate whether or not schizophrenic patients
show different responses to cognitive challenge compared to normal
control subjects.  It was expected that the complexity of the signal
under cognitive challenge is lower in schizophrenics than in normal
control subjects reflecting the impaired information processing
abilities of the patients.  87 schizophrenic and 30 matched control
subjects performed two different types of the continuous performance
task.  The EEG was analyzed by calculation the Grassberger-Procaccia
correlation dimension.  The results revealed no differences between
schizophrenics and controls under resting condition.  In contrast,
during the first minute under task conditions the control subjects
showed a significant decrease of the dimension while no changes were
found for the schizophrenic group.  These results occurred for both
types of the cognitive task.  The results are interpreted as reflecting
the ability of normal subjects to adapt their information processing
system to the cognitive challenge. In contrast, schizophrenic subjects
do not show any adaptation to the task.

Impaired information processing and autonomic conditioning in
Peter Kirsch
Central Institute of Mental Health

A reduced acquisition of autonomic conditioned responses as well as
electrodermal nonresponding are often reported deficits of
schizophrenic patients. In the framework of cognitive approaches, these
deficits were interpreted as reflecting a reduced amount of controlled
information processing capacity in schizophrenics. It could be shown
that schizophrenics did not differ from normal subjects in their
responses to aversive unconditioned stimuli (UCS). In contrast, they
did not show a sufficient discrimination between a paired conditioned
stimulus  (CS+) and an unpaired conditioned stimulus  (CS-). The aim of
the present study was to investigate the relation between impaired
controlled information processing and autonomic conditioning in
schizophrenics. 15 schizophrenic patients and 15 normal control
subjects were investigated in a non aversive classical conditioning
procedure. A letter reproduction task served as UCS while the CS was a
screen providing information about the task. Electrodermal responses
(EDR) , heart rates and reaction times (RT) were used as dependent
measures. It was expected that, with respect to the physiological as
well as the behavioral measures, schizophrenics showed a poorer ability
to differ between different stages of CS information content than
normals.  The results confirmed the expected differences. No
differences of the EDR to the different CS were found for the
schizophrenic group. In contrast, the normal subjects showed clear
evidence for differential conditioning. The same differnces between the
groups occured for the RT measures.  We interpret the results as
indicators of the reduced amount of controlled information processing
acquired by the schizophrenic patients during conditioning.

The effects of emotional disclosure and traumatic life event history on
blood pressure and heart rate in college-aged females.
David J. Klein and John T. Cacioppo
The Ohio State University

Emotional disclosure, the verbal expression of facts and emotions
related to a traumatic event,has been found to have beneficial effects
on health outcomes and affect. It has been hypothesized that the
effects of disclosure result from the confrontation an upsetting event,
which decreases negative physiological effects of trauma-related
inhibition and rumination.  Ninety-one female participants were
selected from over 400 based on answers to a questionnaire about their
traumatic event history.  Two groups of subjects were formed from those
who either reported having experienced or reported having never
experienced a traumatic event (e.g., assault).  Subjects then wrote in
groups of 5 or less for 30 minutes about one of three randomly assigned
topics: (1) their most traumatic experience (real disclosure
condition), (2) a traumatic event (described by a short paragraph) that
they never experienced according (imaginary disclosure condition), or
(3) the physical layout of the campus (no-trauma control). A month
later subjects were run individually at which time their cardiovascular
activity was recorded at one-minute intervals while they: (a) sat
quietly (baseline) and (b) talked anonymously into a tape recorder for
6 minutes about the facts and feelings surrounding the most traumatic
experience of their life (disclosure period). Results indicated that
those with a history of trauma had higher blood pressure and heart rate
across the trauma disclosure conditions.  In addition, blood pressure
and heart rate were elevated during the disclosure period relative to
baseline for all those with and without a history of trauma and across
trauma-disclosure conditions.

Gender differences in susceptibility to vection-induced motion sickness
and gastric myoelectric activity
Alexandrea H. Klose, Lisa M. Willoughby, and Senqi Hu
Humboldt State University

The purpose of the present study was to investigate gender differences
in susceptibility to vection-induced motion sickness.  Twelve men and
14 women sat in an optokinetic drum for 8 minute baseline and 16 minute
drum rotation periods.  During the drum rotation period, subjects'
symptoms of motion sickness (SMS) were measured.  Subjects'
electrogastrograms (EGGs) and the ratios of spectral intensity at 4-9
cycles per minute (cpm) between drum rotation and baseline periods,
gastric tachyarrhythmia, were obtained.  Repeated measures of analysis
of variance on scores of SMS showed significant differences between men
and women, F(1, 24) = 3.94, p < 0.05.  Women reported significantly
more severe symptoms of motion sickness than men while viewing the
rotating drum.  The mean scores of SMS were 5.45 for men and 6.55 for
women.  One-way analysis of variance on ratios of spectral intensity at
EGG 4-9 cpm activity between drum rotation and baseline periods
revealed a significant difference between men and women, F(1, 24) =
4.68, p < 0.05.  Women developed more severe abnormal gastric
tachyarrhythmia than men.  The mean ratios of spectral intensity at EGG
4-9 cpm were 2.07 for men and 3.54 for women.  These results indicated
that women are more susceptible to motion sickness than men.

Can the temporal locations of heartbeat sensations be measured using a
simultaneity paradigm?
Kelley A. Knapp and Jasper Brener
SUNY at Stony Brook

In the method of constant stimuli (MCS) subjects judge whether tones
presented at various time intervals after the R-wave are simultaneous
with their heartbeat sensations.  The median interval (MDI) chosen
under these conditions is assumed to identify when the heartbeat
sensation occurs relative to the onset of ventricular contraction.  It
may be inferred that heartbeat sensations detected <100ms after the
R-wave are transduced through receptors in or close to the heart, while
sensations detected later are sensed by more peripheral receptors.
Thus, the MDI may help to identify broadly which receptors transduce
heartbeat sensations.  The validity of the MDI as a measure of the
relative temporal locations of sensations was tested using vibrations
and tones.  On the first of three sessions, 35 subjects adjusted (ADJ)
the time interval between vibrations and tones until the two stimuli
were simultaneous.  Subjects then completed two tasks in which they
judged the simultaneity of vibrations and tones presented at 6
different interstimulus intervals.  In MCS1 but not in MCS2, one of
these vibration-tone intervals was equal to the median of the ADJ
task.  If the MDI is a valid measure of when the vibration is sensed
relative to the tone then all MDIs should be equal.  However MDIs were
found to be significantly different: ADJ (20 ms), MCS1(14 ms), and MCS2
(4ms), indicating a degree of uncertainty in measuring the relative
temporal locations of sensations using the simultaneity paradigm.  This
research was supported by NIH Grant #HL42366 to JB.

Event-related potentials during dual task performance: Tracking and
auditory discrimination
Timothy F. Knebel
NASA Langley Research Center

This study examined the relationships between tracking difficulty,
tracking error, counting error, and ERP amplitude during auditory
discrimination.  Twenty-four dextral volunteers performed a tracking
task across three levels of difficulty while simultaneously counting or
ignoring tones of two different pitches.  EEG was recorded and averaged
at frontal, central, and parietal electrode sites.  Peak amplitude
measures were obtained for components:  N1, P2, N2, and P3.
	The amplitude of P3 was significantly diminished in the
difficult tracking level compared to the easy and medium
tracking levels.  P3 amplitude was larger for counted stimuli
especially at the central and parietal regions.  N2 amplitude was
greater for counted stimuli at the frontal region.  Tracking error,
measured as root-mean-square error (RMSE), increased significantly from
the easy tracking level to the most difficult.  N2 and P3 amplitudes
were significantly and negatively correlated with RMSE.  Counting error
rate was negatively correlated with P3 amplitude.  Results are
interpreted consistent with theories of selective attention and
resource allocation.

Cardiovascular load perception: Results from studies using the method
of reproduction vs. the tracking method
Volker Kollenbaum1, Bernhard Dahme2, and Guenther Kirchner2
1University of Kiel, 2University of Hamburg

The interoception of heart rate, systolic blood pressure and
myocardial metabolism have been studied in two separate experiments.
In both experiments subjects were confronted with three levels of each
of these three cardiovascular parameters during bicycle ergometer
testing.  The intensity levels were 25, 50, and 75 % of the individual
range of aerobic load.  In the first experiment the subjects had to
reproduce these same values of the parameter in question by regulating
the physical load of the ergometer, in the second experiments subject
had to estimate the level of the parameter in question and its decay
after load had stopped by a tracking device. In the first experiment
complete data were gathered from 14 subjects (7 f, 7 m), in the second
one from 18 subjects (8 f, 10 m).  Reproduction or tracking accuracy
were taken as interoceptive indices.  Heart rate and blood pressure
were measured beat to beat by a Finapres device, the myocardial
metabolism was determined by the Robinson-Index.  By both methods
subjects underestimated heart rate.  In regard to blood pressure both
methods show divergent results.  There is no substantial bias in the
perception of the Robinson-Index.  We found a high internal consistency
in heart rate reproduction, but no monotone consistency at all in blood
pressure reproduction.  The physiological processes underlying
interoception and reproduction of cardiovascular load have to be still
explored.  The clinical implications of the bias in heart rate
perception during cardiac load will be outlined.

A method for two-dimensional self-regulation of slow cortical
potentials: Toward non-motoric communication
B. Kotchoubey, H. Schleichert, W. Lutzenberger, and N. Birbaumer
University of Tuebingen

Healthy subjects were presented with a sequence of two alternating
tones with 2-s ISIs. Subjects learned to produce rhythmic oscillations
of their slow cortical potentials (SCP) by generating directed SCP
changes in certain intervals between tones, while a moving object on a
screen provided them with continuous EEG feedback. Furthermore,they had
to produce two different EEG signals: a) positive or negative SCP shift
atvertex, and b) SCP asymmetry between the right and the left central
area. Thus, four conditions were to distinguish: (1) larger Cz
negativity in the second half-cycle than in the first half-cycle (the
on-screen object moving upwards), (2) larger Cz negativity in the first
than in the second half-cycle(the object moving downwards), (3) C3 over
C4 dominance in the second ascompared with the first half-cycle
(rightwards), and (4) C4 over C3 dominance in the second as compared
with the first half-cycle (leftwards). In the first session, 14
subjects differentiated significantly between the conditions (1) and
(2), but not between (3) and (4). Later, five subjects participated in
6 to 15 training sessions; four of them demonstrated highly-significant
differentiation between the right and the left conditions as well. The
differentiation was achieved within less than 300 ms after the
discriminative signal, i.e.much faster than in studies employing
traditional SCP biofeedback technique. The data suggest a hope of the
possibility to control a cursor on a two-dimensional screen by means of
SCP, which can be applied in severe motor disorders where no other
means of communication exist.

Cardiac reactivity to food intake by dieting status and gender
Jean Kristeller, Jeff McKee, and Thomas Johnson
Indiana State University

Previous work by this investigator identified a tachycardic response
(8-10 bpm) during food intake (peanut M&Ms), modified  by psychological
variables (anxiety and dieting concern). A tachycardic response to
sucking has been demonstrated in infants, but this tachycardic response
during food intake has otherwise not been well documented. This study
extends the investigation to a variety of foods and further examines
the effect of dieting concern (Restrained Eating) and gender. Heart
rate was measured on a continuous basis (using the BIOPAC system) in 16
female and 9 male undergraduates.  After a resting baseline, subjects
ate, on signal, four foods (pretzels, chips, grapes, peanut M&Ms),
varying by sweet and fat level.  Average heart rate was calculated over
each one minute eating trial. Heart rate increased significantly
(p<.0001) from baseline for all foods, ranging from 6.4 bpm (pretzels)
to 7.9 bpm (grapes), and rapidly returned to baseline between eating
trials. There were no significant food type or gender effects. There
was a significant interaction (F=6.44, p<.025) between level of
restraint and type of food; low restrained participants showed high
heart rate responses to all foods (about 9 bpm), while high restrained
participants responded the least to non-sweet foods (about 4.5 bpm) and
at a moderate level to sweet foods (about 7 bpm). These data support
the reliability of a tachycardic response during food intake, one which
may be affected by psychological factors. Investigating this effect may
have implications for understanding conditioned responses to foods
characteristic of both normal and disordered eating patterns.

Effects of drinking milk and water on gastric activity and
vection-induced symptoms of motion sickness
John J. Lagomarsino, Kelly R. McVicker, Donald A. McCarty, and Senqi Hu
Humboldt State University

The purpose of the present study was to investigate
the effects of drinking milk or water on electrogastrograms
(EGGs) and the susceptibility to vection induced motion sickness.
A vertically striped optokinetic drum was used to provoke
symptoms of motion sickness in the study.  Twenty-seven subjects
were divided into three groups: drank 300 ml milk, drank 300 ml
water, or drank nothing.  EGG activity and subjective symptoms of
motion sickness (SSMS) were measured throughout the experiment.
The results indicated that the subjects who drank 300 ml milk
reported significantly less symptoms of motion sickness than
those who drank 300 ml of water or who drank nothing, F(2, 23) =
5.93, p < 0.001.  The mean scores of SSMS were 3.40 for the milk
drinking group, 5.36 for the water drinking group, and 5.98 for
the no drinking group.  EGG recordings indicated that the
subjects in the no drinking group showed significantly higher
abnormal 4-9 cycles per minute (cpm) EGG activity during drum
rotation period than those in the milk and water drinking groups,
F(2, 23) = 4.91, P < 0.01. The mean ratios of EGG 4-9 cpm spectral
power estimates between drum rotation and baseline periods were
1.01 for the milk drinking group, 2.02 for the water drinking
group, and 19.20 for the no drinking group.  In conclusion, milk
drinking before optokinetic drum rotation reduced vection-induced
symptoms of motion sickness and associated abnormal gastric
myoelectric activity.

A comparison of event-related potential and skin conductance measures
of classical conditioning.
William J. Lammers
University of Central Arkansas

This study assessed the correspondence between central (event-related
potential) and peripheral (skin conductance) measures in a traditional
skin conductance classical conditioning paradigm.  Ten subjects
listened to two 8s tones (CS+ and CS-) randomly presented during two
phases of conditioning.  During habituation, no UCS was presented.
During acquisition, the CS+ was always followed by white noise (1s; 100
dB).  Subjects were instructed to perform a counting task to the tones
and to also determine which tone preceded the loud noise.
The first interval SCR (1-4s post CS onset) was larger to the CS+
than to the CS- during acquisition but it appears that the difference
may have already existed during habituation.  The second interval SCR
(4-7s post CS onset) showed the clearest evidence of differential
conditioning, with a larger SCR to the CS+ than to the CS- during
acquisition (with no initial difference during habituation).  In
contrast, the contingent negative variation (CNV) prior to CS
termination at the Fz scalp site increased from habituation to
acquisition for both the CS+ and CS-, with no difference between the
CS+ and CS-.  Correlation coefficients showed significant relationships
between the first, second, and third interval SCRs.  However, no
consistent patterns of correlations were observed between SCR measures
and CNV measures.
Often, very similar terms have been used to characterize the
mental processes reflected by the second interval SCR and the CNV
(e.g., expectancy; attention toward an impending stimulus).  The
current findings with a classical conditioning paradigm suggest that
these measures may in fact represent different brain processes.

Caffeine raises blood pressure and heart rate on the job
James D. Lane, Barbara Phillips-Bute, and Carl F. Pieper
Duke University Medical Center

Caffeine is known to raise blood pressure at rest and during stress in
the laboratory.  Caffeine's effects on ambulatory blood pressure (BP)
and heart rate (HR) during normal workday activities were measured in
21 healthy normotensive volunteers, all habitual coffee drinkers.  High
(500 mg = 5 cups of coffee) and low (100 mg = 1 cup of coffee) daily
doses of caffeine were administered on different days in
counterbalanced order.  Half of each dose was given in the morning and
the remainder at midday.  Ambulatory BP and HR were recorded every 15
min with Accutracker monitors for 6-8 hours until the end of the
workday.  Subjects completed concurrent diary entries for location,
posture, and perceived activity and stress.  Mood and symptoms were
assessed by questionnaire at morning, midday, and evening.  Nineteen
subjects (11 M, 8 F) provided usable ambulatory data.  Caffeine dose
affected both BP and HR, even when posture, activity, and stress were
statistically controlled.  The high dose raised BP by 3-4 mmHg and HR
by 3 bpm.  There was only minimal evidence of caffeine withdrawal in
the low dose condition.These results, together with those of earlier
studies, suggest that the consumption of caffeinated beverages raises
BP and HR during everyday activities.  Moreover, the cardiovascular
effects of caffeine are not accounted for by dose-related differences
in posture, activity, and stress.  Although the effects are moderate in
size, the widespread consumption of caffeinated beverages means that
caffeine's effects may make significant contributions to coronary
disease risk in the adult population.

Inside picture processing:  Emotional modulation of  ERPs from the
cortical surface
Peter J. Lang, Robin Gilmore, Bruce N. Cuthbert, Margaret M. Bradley, 
Steven Roper, Harald Schupp & Jean Cibula
University of Florida

Scalp EEG electrodes are remote sensors separated by space and
barriers of flesh and bone from the activity of neural cells.  Thus,
signal amplitude is both greatly attenuated and spatially smeared,
diminishing our ability to locate the electrical dipoles that drive
evoked potentials.  This problem is particularly acute in studies of
emotion, as the mediating structures are presumed to lie deep within
the brain.  In the present experiment, event related potentials were
recorded directly from the surface of the cortex during viewing of
pleasant, unpleasant, and neutral pictures (selected from the
International Affective Picture System).  Subjects were hospitalized
patients with implanted sub-dural electrode grids, being evaluated for
surgical treatment of epilepsy.  Preliminary results indicate that
evoked waveforms from electrodes on the parietal cortex are
characterized by a marked late positivity and that this potential shift
is generally sustained over the entire period of picture processing.
As found for slow waves recorded from scalp electrodes, more arousing
emotional stimuli prompt the largest responses.  These effects were
generally more pronounced for more posterior electrode sites, with good
differentiation between electrodes 1-2 cm apart.

Heart rate and behavioral measures of attention in 6-, 9-, and 12-month
old infants during object exploration
Jeffrey M. Lansink and John E. Richards
University of South Carolina

Behavioral and physiological indices are used to identify infant
attention phases which differ with respect to cognitive processing.
These indices differentiate an engaged phase of attention from a
relatively inactive phase.  Studies employing behavioral measures
(focused and casual attention) and those using heart rate to index
attention (sustained attention and attention termination) have found
that infants are harder to distract during active than inactive
attention phases.  The present study used a multi-operational
definition of infant attention using behavioral and heart rate
Subjects for this experiment included 14 6-month olds, 16 9-month
olds, and 14 12-month olds.  A toy was presented to the infant and
following a delay a peripheral distractor appeared.  There were four
delays: 1) sustained attention, a significant heart rate deceleration
of two seconds; 2) attention termination, two seconds after heart rate
had returned to the prestimulus level following a significant
deceleration; 3) focused attention, a two second behavioral judgement
of attention and 4) casual attention, a two second behavioral judgement
of inattention following two seconds of behavioral engagement.  Time to
localize the peripheral stimulus was the dependent variable.
There were longer distraction latencies during attentional
engagement than inattention for behavior and heart rate delays.  Heart
rate/behavior concordant measures with respect to attention engagement
or inattention showed the longest and shortest distraction latencies,
respectively.  There was no interaction of measurement method (heart
rate or behavioral) with attention type (engaged or inattention).
Heart rate and behavioral ratings combined provided the greatest amount
of discrimination of distraction latencies.

Prediction of resting blood pressure using self-reported cynical
hostility: Support for the 'Composite Hostility' (CHOST) scale
Mark R. Larson and Alan W. Langer
Syracuse University

The present study had two main purposes.  The first purpose was to
examine the relationship of resting blood pressure to self-reported
cynical hostility levels using the Cook-Medley Hostility Inventory
(Ho).  Meanwhile, because it has been suggested that some Ho scale
items may not truly be related to cynical hostility, the second purpose
of this study was to compare the overall Ho score to a Cook-Medley
subscale known as the composite hostility (CHOST) scale, which was
devised by Barefoot et al. (Psychosomatic Medicine, 1989, pp.46-57) as
a potentially more accurate measure of cynical hostility.  Subjects
were 64 young adult male volunteers from a university setting.  Two
resting blood pressures (BP's) were taken for each subject, spaced five
minutes apart.  As expected, the second of the two BP's proved most
stable, and these second measurements were therefore compared with the
Cook-Medley responses.  Results showed clear evidence of a relationship
between both systolic and diastolic blood pressures and cynical
hostility levels.  Moreover, as anticipated, correlations obtained
using the CHOST scale proved more significant than those using the
overall Ho scale, supporting the hypothesis that the CHOST scale may be
more useful for assessing cynical hostility (systolic BP, CHOST scale:
r=.267, p<.05; systolic BP, Ho scale:  r=.191, p=NS; diastolic BP,
CHOST scale:  r=.334, p<.01; diastolic BP, Ho scale:  r=.281, p<.05).

Relations between thalamic metabolic activity and alpha power
Christine L. Larson, Daren C. Jackson, Heather C. Abercrombie, R. Terry 
Ward, Stacey M. Schaefer, 
James E. Holden, Scott B. Perlman, and Richard J. Davidson
University of Wisconsin-Madison

Alpha electrocortical activity is a commonly used measure of regional
brain activation.  Alpha power reflects an awake, but resting state and
has been hypothesized to be inversely related to activation.  The
thalamus has been proposed to be an important site for regulating
cortical rhythmic activity.
To examine relations between thalamic activity and alpha power,
resting regional cerebral glucose metabolism (rCMRglu) and 28 channels
of EEG were measured simultaneously in 26 subjects (16 depressed, 10
control).  rCMR was obtained using fluorodeoxyglucose F18 positron
emission tomography (FDG-PET).  During the FDG uptake period (30
minutes), continuous EEG was recorded.  Artifact-free epochs of EEG
were referenced offline to the averaged ears reference.  Mean alpha
(8-13 Hz) power for the 30-minute recording period was computed.  The
PET data were transformed into Talairach space for statistical
comparison with the EEG.
Average alpha power across all electrode sites was computed.  This
metric was used with Statistical Parametric Mapping to map the
entire brain volume for correlations with metabolic rate (omnibus p<
.001).  Correlations were performed across all subjects since we did
not hypothesize any group differences in this relation.  A region
including the thalamus showed the most robust and consistent inverse
relation with the measure of global alpha power (r's from -.58 to
-.68), indicating that in humans there is an inverse relation between
thalamic activation and cortical alpha power. These data underscore the
role of the thalamus in cortical EEG and indicate that individuals who
have greater thalamic metabolic rate also have more global alpha

Electrophysiological measurement of parotid response to food in fasted
and nonfasted subjects
Daniel B. LeGoff, PhD, Christopher Davis, PhD, and Thomas Bauslaugh, MA
Simon Fraser University Burnaby, BC, Canada

Past research on appetite, eating behavior and physiological response
to food has indicated the potential importance of accurately measuring
changes in saliva flow rate (cotton swab absorption, dental ejector,
Lashley suction disk) are invasive, have relatively low sensitivity,
and are putatively reactive.  Recent research has established that flow
rate can be assessed using surface electrodes which are sensitve to the
myoepithelial and secretory cell changes in the parotid gland.  This
electrophysiological technique (electrosaliogram, ESG) was used to
distinguish the salivary responses of fasted (n=10) and nonfasted
(n=10) subjects to food and nonfood odours.  It was found that the ESG
reliably distinguished food from nonfood response levels, as well as
fasted from nonfasted subjects.  Overall, ESG response to food minus
nonfood baseline was significantly correlated with time since last
meal.  It is concluded that the ESG provides a non-invasive and
accurate measure of salivary response to food stimuli, and is sensitive
to degree of food deprivation.

Psychopathy and startle modulation during affective picture processing:
A replication and extension
Gary K. Levenston1,Christopher J. Patrick1, Margaret M. Bradley2, and 
Peter J. Lang2
1The Florida State University, 2University of Florida

Startle modulation was assessed during viewing of pleasant, neutral,
and unpleasant pictures in 18 psychopathic and 18 non- psychopathic
prisoners classified using Hare's Psychopathy Checklist.  Pleasant
slides depicted thrilling activities and erotica; unpleasant slides
depicted personal threat, victimization, and mutilations.  Blink reflex
magnitude to acoustic probes (50 ms, 105-dB) was recorded at 300, 800,
1800, 3000, or 4500 ms following slide onset.
A Group X Valence X Probe Time interaction was found, F(8, 27) =
2.95, p<.05.  Nonpsychopaths showed evidence of an attentional prepulse
effect at 300 ms (i.e., both affectiveude and amplitude analysis of ERP signals.

Dimensional complexity and nonlinearity of sleep EEG recorded from
insomniacs and normal controls
Derek Loewy1 and Walter Pritchard2
1University of Ottawa, 2R.J. Reynolds Tobacco Co.

The techniques of nonlinear dynamics were used to compare the sleep
EEG of chronic insomniacs and normal controls.  Six subjects diagnosed
with chronic primary insomnia and six normal controls slept for one
night in a sleep laboratory.  EEG was recorded from C3 referenced to
the right mastoid.  Data were recorded with 12 bits precision, sampling
rate 240 Hz, and 0.3-60 Hz on-line filters.  Ten second epochs of EEG
were analyzed.  These were obtained during wakefulness with eyes open
(fixating) and with eyes closed.  EEG data were also recorded during
the following sleep stages: Stage 1 (first epoch after sleep onset),
Stage 2 with "spindles", early Stage 2 without spindles (after first
spindle or "K-Complex"), late Stage 2 (last 1/3 of the night), slow
wave sleep, REM sleep with rapid eye movements ("phasic REM"), and REM
sleep without rapid eye movements ("tonic REM").  EEG "dimensional
complexity" (DCx) was estimated using the Takens-Ellner Method and
singular value decomposition.  The embedding lag was determined using
the method developed by Rosenstein et al. (1994).  A measure of
nonlinearity was also obtained by computing the difference in DCx
estimates for these original data and DCx estimates for surrogate data
sets generated using the Gaussian-distributed amplitude-adjustment
procedure.  No significant differences in DCx were found between
insomniacs and normals for any wake/sleep stage.  DCx was lowest for
Stage 2 with spindles and highest for Stage 1.  Nonlinearity was found
to be significant for Stage 2 with spindles, phasic REM, and awake with
eyes open.

Effects of clonidine on event-related potential indices of auditory and
visual information processing
Christopher T. Lovelace1, Connie C. Duncan2, and Walter H. Kaye3
1National Institute of Mental Health, 2Uniformed Services University of 
the Health Sciences, 
3University of Pittsburgh

Clonidine, an alpha-2 adrenoceptor agonist, inhibits activity in the
ascending noradrenergic projections originating in the locus coeruleus
and terminating in the cortex; this system has been implicated in
arousal and attention.  In separate testing sessions, eight healthy
adult women were given 0.5, 1.0, or 2.0 micrograms/kg intravenous
clonidine or a saline placebo in randomized order.  Event-related
potentials (ERPs) were recorded while subjects performed choice
reaction time tasks.  Two highly discriminable stimuli (auditory or
visual, in separate trial blocks) were presented with relative
probabilities of .10 and .90.  The high dose increased reaction time
and decreased accuracy of responses in the auditory task.  There was a
similar, though nonsignificant, effect on accuracy in the visual task.
The high dose also markedly reduced auditory P300, whereas no
significant changes in performance or P300 were observed following the
low or medium dose.  Clonidine did not affect the amplitude of visual
P300 or slow wave in either modality.  The medium and high doses
increased the latency of visual N100.  Mismatch negativity (MMN) was
quantified in the auditory .10 - .90 difference ERPs.  The low dose
reduced the amplitude of MMN, but the medium and high doses did not.
Clonidine did not affect MMN latency.  The differential effect of
clonidine on auditory and visual P300 reinforces the view of separate
auditory and visual attentional systems, with greater resilience in the
visual system.

The influence of hunger on the evaluative process
D.I. Lozano and S.L. Crites
University of Texas at El Paso

A study of the effect of hunger on the evaluative process was
conducted.  Event-related brain potentials and behavioral data were
recorded to assess the effects of hunger on 1) the categorization of
stimuli, as indexed by ERPs; and 2) the behavioral responses to
stimuli.  Participants abstained from food for at least 15 hours prior
to the time of the experiment.  The experiment required participants to
categorize trait adjectives, foods, and animals as positive or
negative.  Participants were exposed to two blocks of 300 words, one in
which the context words (trait adjectives) were positive and one in
which the context words were negative.  Both positive and negative
target words (foods and animals) were embedded in the positive and
negative context blocks.  Participants were given a meal following
completion of the categorization task and were required to complete the
same categorization task following the meal.  The ERP data show that
the amplitude of the P300 is larger following a meal, consistent with
previous research (Geisler & Polich, 1992, Psychophysiology, 76-85).
In contrast to previous findings, the amplitude of the P300 is larger
when positive items are embedded in a negative context than when
negative items are embedded in a positive context.

Event-related potentials and serial position effects with 12-item lists
D.I. Lozano, S. Moreno, J.V. Devine, and S.L. Crites
University of Texas at El Paso

Event-related brain potentials and behavioral data were collected to
probe stimuli in a serial probe recognition task.  Participants were
presented with lists of twelve pictures followed by either a matching
or a non-matching probe.  Serial position effects were examined by
selecting the matching probes from list items 1-3, 5-7, or 10-12.  The
probe matched an item on the list on 50% of the trials.  The behavioral
data show that responses were more accurate when probes were presented
at the end of the list than when probes were presented at either the
middle or beginning of the list.  In addition,responses to the probes
at the end of the list were faster than the responses to either probes
at the beginning or at the middle of the list.  As has been found
previously, the ERP waveforms evoked by matching and non-matching
probes began to differ around 200 ms following probe presentation.
Furthermore, there were differences in the ERPs according to the
position of the probes.  The amplitude of a  positive component
occurring around 300 ms following probe presentation was largest for
probes located at the end of the list.  In addition, the amplitude of a
negative component occurring around 360 ms following probe presentation
was larger for probes that were located at the end of the list than for
those located at either the beginning or the middle of the list.

Cortisol reactivity, self-esteem, and depression
Kristen A. Luscher, Angela Scarpa, Christine N. Christensen, and Kadee 
J. Smalley
Eastern Washington University

Previous literature has found a relationship between low self-esteem
and depression.  The mechanism underlying this relationship, however,
has not been fully understood.  One possible mechanism is cortisol
reactivity which has also been associated with depression.  This
preliminary study investigates cortisol reactivity as a factor which
mediates or moderates the relationship between self-esteem and
Fifty-three participants (20 male), ages 19-58 (mean 27), were
exposed to two laboratory stressors (a set of uncontrollable stressors
and an anagram task).  Self-report self-esteem and mood ratings were
obtained prior to the stressors.  Of the 53 participants, saliva
samples were taken before and after the stressors from 15 (6 male)
participants, ages 21-45 (mean 27), for the determination of cortisol
levels.  Cortisol reactivity was measured by subtracting baseline
cortisol levels from post-stressor task levels.
Results supported a relationship between self-esteem and
depression (r = .57, p = .025).  Results did not support a
relationship between self-esteem and cortisol reactivity, indicating
cortisol reactivity is not acting as a mediating variable.  In order to
test whether cortisol reactivity functions as a moderating variable, a
two-way ANOVA was performed.  Results indicated significant main
effects for self-esteem (p =
.006) and cortisol reactivity (p = .007).  Results also indicated a
significant interaction (p = .012), whereby depression is highest in
those exhibiting a combination of low self-esteem and high cortisol
In conclusion, a psychobiological explanation for depression may
exist, such that individuals with lower self-esteem and high cortisol
reactivity are at the greatest risk for depressive symptomatology.

Coping-related drinking in high anxiety sensitive individuals
Alan MacDonald and Sherry Stewart
Dalhousie University

The present study evaluated Sher's (1987) stress- response dampening
(SRD) theory to explain the increased sensitivity of high anxiety
sensitive (AS) subjects to alcohol- induced SRD effects compared to low
AS controls. High and low AS subjects were divided into alcohol and
placebo groups and participated in a hyperventilation challenge known
to induce physiological symptoms feared by high AS subjects (Donnell &
McNally, 1989). Reactivity was measured using affective, cognitive and
somatic self-report scales and measures of heart rate (HR), blood
volume pulse (BVP) and skin conductance level (SCL).  Preliminary
analyses suggest that high AS placebo subjects showed greater
reactivity to the challenge on affective and cognitive measures than
low AS placebo subjects. Furthermore, both high and low AS alcohol
subjects showed less reactivity on somatic self- report measures
compared to placebo subjects. However, on affective and cognitive
measures, the strongest reductions in reactivity were seen in high AS
subjects, suggesting that alcohol may be reinforcing for high AS
subjects by eliminating their tendency to become fearful and
catastrophize about aversive experiences. Based on previous
psychophysiological findings (Stewart & Pihl, 1994) and preliminary
analyses, high AS placebo subjects were expected to show greater SCL
reactivity in response to the hyperventilation challenge than low AS
placebo subjects. In addition, all subjects receiving alcohol were
expected to show less SCL reactivity compared to placebo subjects, with
the strongest alcohol-induced SCL reductions expected for high AS
subjects. The results are discussed in terms of the theoretical links
between AS and psychophysiological vs. subjective-emotional responses
to stress.

Is a frontal positive slow wave in the ERP specific for emotion-focused
Stefanie Maier1, Oliver Diedrich2, Gabriele Becker1, Ewald Naumann1, and 
Dieter Bartussek1
1University of Trier, 2University of Tuebingen

Emotional slides elicited a frontal positive slow wave in the ERP with
larger amplitude for an emotion-focussed processing group than for a
structural processing group. This effect was interpreted as a
consequence of the activation of specific neuronal structures when
subjects attend to the emotional content of the stimuli. The aim of the
present study was to show whether the emotion-focussed processing of
neutral slides elicits a frontal positive slow wave as well.
Therefore, ERPs were recorded from subjects watching either emotionally
neutral or negative slides (presented for 6-7.5 sec). Four experimental
groups were examined. Each subject performed three phases with 23
slides each: In phase 1 all subjects rated neutral slides along the
cognitive dimensions "shape" and "size", in phase 3 all subjects rated
negative slides along the emotional dimensions "arousal" and
"valence".  In phase 2 the different groups rated either neutral or
negative slides along either the cognitive or the emotional dimensions.
EEG was recorded from F3, C3, P3, Fz, Cz, Pz, F4 C4, P4 from 300 ms
before until 4500 ms after stimulus onset.  The results show that the
frontal positive slow wave is elicited during emotion-focusssed
processing of neutral slides, but also during cognitive processing of
negative and neutral slides. Thus the frontal positive slow wave seems
not specific for emotion-focussed processing of emotional slides. A
reanalysis of the experiment shows that other features than the
emotional content of the slides have an influence on the ERPs as well.
Supported by Deutsche Forschungsgemeinschaft.

A comparison of the Schandry and Method of Constant Stimuli procedures
for assessing heartbeat perception
Jennifer Mailloux and Jasper Brener
State University of New York at Stony Brook

Although Schandry's heartbeat counting task is widely used to assess
the ability to perceive cardiac sensations, performance on this task
may be determined by processes other than sensitivity to cardiac
sensations.   Subjects may perform well on Schandry's task if they
accurately estimate their heart rate.  However, the Method of Constant
Stimuli (MCS) requires that subjects discriminate between stimuli that
are either coincident or delayed with respect to their own heartbeat
sensations and performance on this task is not influenced by the
accuracy of subjects' beliefs about heart rate.  This study compared
performance on these tasks.  In one session, beliefs about the effects
of posture and exercise on heart rate were assessed before subjects
were required to count their heartbeats while their actual heart rates
were manipulated by changes in posture and exercise.  In an attempt to
influence heart rate beliefs, half the subjects (N=40) received
heartbeat-contingent feedback between two presentations of the
heartbeat counting task, whereas the other half received feedback at a
rate 20 percent slower than their actual heart rates.  On another
session (MCS), subjects judged the simultaneity of heartbeats and tones
presented at various intervals after the R-wave.  Subjects classified
as heartbeat detectors on the MCS procedure were also good at counting
their heartbeats on the Schandry task.  Other subjects performed poorly
on MCS but well on Schandry's task.  These subjects' may use beliefs
about heart rate to enhance performance on Schandry's task.  Supported
by NIH Grant #HL42366 awarded to Jasper Brener.

Blind separation of event-related brain response components
Scott Makeig1,2, Tzyy-Ping Jung1,3, Anthony J. Bell3, Dara Ghahremani3, 
and Terrence J. 
1Naval Health Research Center, San Diego CA; 2University of California 
San Diego; 3The Salk 
Institute for Biological Studies

The problem of objectively decomposing event-related brain responses
into neurophysiologically meaningful components is a major difficulty
in the evoked response field.  Traditional methods of identifying and
measuring response subcomponents based on measuring the amplitudes and
latencies of peak excursions in the waveforms at individual scalp sites
fail when subcomponents overlap substantially, while current source
localization procedures based on fitting single or multiple dipole
models give ambiguous results when source geometry is unknown or
complex.  The Independent Component Analysis (ICA) algorithm of Bell
and Sejnowski (1995) is an artificial neural network which maximizes
the overall entropy of a set of non-linearly transformed input vectors
using stochastic gradient ascent, without regard to the physical
locations or configuration of the source generators.  Trained on one or
more multichannel electric or magnetic evoked responses, the algorithm
converges on spatial filters which separate the input data into
independent time courses and distinct scalp topographies arising in
multiple, spatially-stationary 'effective brain sources.  Response
decompositions produced by the ICA algorithm can be used to measure the
effects of experimental manipulations on individual response
components, even when these are overlapping in time or space.
Typically, response components identified by the algorithm are
recaptured in repeated analyses, regardless of changes in initial
weights, sensor montage, and data length.
I will explain the theory and practice of ICA decomposition and
its differences from PCA, demonstrate results of EEG simulations,
and present applications to EEG and MEG data analysis.

Independent component analaysis of event-related potentials during a
selective attention task.
S. Makeig1,2, L. Anllo-Vento2, P. Jung1,3, A.J. Bell3, T. J. Sejnowski3, 
and S. A. Hillyard2.
1Naval Health Research Center; 2University of California San Diego; 
3Salk Institute for Biological 
Studies, La Jolla, and Howard Hughes Medical Institute.

ERP studies of visual-spatial attention indicate that cortical 
processing of stimuli appearing in the attended location is augmented as 
early as 80 ms after stimulus onset. However, separation of the multiple 
brain processes contributing to the surface-recorded components of ERP 
waveforms has proven difficult. Recently, an `infomax' algorithm for the 
blind separation of linearly mixed inputs has been devised (Bell and 
Sejnowski, 1995) and applied to EEG and ERP analysis (Makeig et al., 
1996). The neural generators of sources are not specified by the 
algorithm and may be either physically compact or distributed. Results 
of applying this Independent Component Analysis (ICA) algorithm to 
single-subject and group-mean ERPs recorded during a visual selective 
attention experiment (Anllo-Vento and Hillyard, 1996) suggest that ERP 
waveforms represent a sum of overlapping, discrete and time-limited 
brain processing events whose amplitudes are modulated by selective 
attention without affecting their time course. The source components
identified by the algorithm appear to index independent stages of visual
information processing. Spatial attention operates on early source 
components in a manner similar to a sensory gain-control mechanism, 
while later components appear to reflect further processing of stimulus 
features and feature conjunctions.

Phase of menstrual cycle modulates eye-blink startle potentiation and
Rachel Manber, Keith W. Burton, John J. Allen, and Alfred W. Kaszniak
University of Arizona

	 Emotional modulation of the eye-blink startle reflex was
hypothesized to be sensitive to the emotional changes that often
accompany the premenstrual phase of the menstrual cycle.  Seventeen
regularly menstruating healthy young women (ages 18 to 24) were studied
in both midfollicular and late luteal phases.  In each session the
magnitude of the eye-blink startle response to positive, neutral, and
negative color slides was measured and the PANAS was administered.
A significant valence effect replicates previous findings when
data from both phases were pooled (p<.05).  The main valence
effect, however, was significant only during the luteal  phase
(p<.05).  There were trends for greater potentiation (the difference
between response to negative and positive slides, p=.055) and greater
startle magnitude for neutral slides (p<.09) in the late-luteal
compared to the follicular phase.   The correlation between background
mood (positive minus negative subscales of the PANAS ) and startle
potentiation was significant  (r=.55) during the follicular (but not
luteal) phase.  Thus, during the more emotionally quiescent period more
dysthymic mood was associated with a greater attenuation of the startle
potentiation  and magnitude (r=.48 to .56).
Though not conclusive, these results suggest differential
emotional processing in the two phases of the menstrual cycle.  We
did not, however, find a significant phase difference in activity of
the corrugator and zygomatic muscles.  It appears that the startle
response may provide a more sensitive index of alterations in emotional
responding during phases of the menstrual cycle than the reactivity of
the corrugator and zygomatic muscles to the same stimuli.

Electrophysiological (ERP) correlates of encoding and retrieval in
episodic memory
Jennifer A. Mangels, Terrance W. Picton, and Fergus I. M. Craik
Rotman Research Institute

To assess the time course of neural activity during encoding and
retrieval in episodic memory, we recorded event-related potentials
(ERPs) from 46 electrodes as subjects studied and recognized words.
During encoding, subjects studied words under full or divided
attention. The divided-attention task required subjects to make motor
responses to tones presented in either a predictable pattern ("easy")
or a random sequence ("difficult"). During retrieval, subjects
classified items depending on whether recognition was accompanied by
conscious recollection of the study-episode ("remember") or whether
recognition was based only on familiarity of the stimulus ("know"). In
all encoding and retrieval conditions, the N180 observed over
occipito-temporal sites was larger over the left hemisphere and may
reflect processes associated with the automatic identification of
verbal stimuli.  From 400 to 2400 ms, a sustained positive wave emerged
maximally over prefrontal sites. This frontal positivity was
lateralized to the left hemisphere at encoding and to the right
hemisphere at retrieval. In addition, frontal activation was modulated
by attention at encoding and by successful recognition at retrieval.
Specifically, frontal positivity was evident in the full attention and
"easy" divided-attention encoding conditions, but was not apparent in
the "difficult" divided-attention condition. Frontal positivity was
associated with correctly recognized words, but not with new words. A
centro-parietal positive wave (P650) was also observed at retrieval
that was larger for "remember" responses than for "know" or "new"
responses. These findings support the lateralized role of the
prefrontal cortex in attentionally controlled aspects of episodic
encoding and retrieval.

Cardiovascular and anger responses to stress during pre- and
midmenstrual phases
Amber A. Marcia, Kelly L. Thompson, Lauren J. Miller, and Senqi Hu
Humboldt State University

The purpose of the present study was to investigate cardiovascular and
anger responses to stress during pre- and midmenstrual phases of the
menstrual cycle.  Fifty women between the ages of 18 and 45, who had
regular 25 to 32 day menstrual cycles, participated in the study.
Subjects initially completed a questionnaire assessing levels of state
and trait anger (State-Trait Anger Scale; Speilberger, Jacobs, Russel,
and Crane, 1983) and a Likert-type scale rating anger, frustration,
andirritability.  Subjects were then asked to perform a "Serial-7"
mental arithmetic task for 10 minutes.  We examined the differences
between women in the midmenstrual phase and women in the premenstrual
phase on state and trait anger, Likert ratings of anger, irritation,
and frustration, and heart rate before and after the laboratory-induced
stressor task.  Results revealed significant differences between the
two phases on baseline ratings of frustration, F(1,38) = 5.32, p < .05,
and on post-stressor Likert ratings of anger, F(1,38) = 6.09, p < .05.
In addition, significant differences were revealed between the two
phases on baseline ratings of total anger as measured on the
State-Trait Anger Scale, F(1,38) = 5.96, p < .05.  Finally, a one-way
analysis of variance also revealed significant differences between the
two phases on the increase in heart rate that occurred from baseline to
post-experimental recordings, F(1,38) = 5.82, p < .05.  These findings
indicate that women are more reactive to stress during the premenstrual
phase of their menstrual cycle than they are during the midmenstrual
phase of the cycle.

Priming pictures and words: An investigation of the N400 and the LPC.
Frances Martin
University of Tasmania

	This study compared ERPs produced by the processing of pictures
and words in a cross modal recognition memory paradigm.  In the
acquisition phase, subjects were presented with a series of sentences,
on word at a time, concluded by two stimuli.  The stimuli (either a
word or a picture) were identical, related to the target, or
unrelated.  The target (word or picture) was either congruous with the
sentence stem or incongruous.  Subjects (n=54) were divided into four
groups according to the stimuli they viewed in each phase (word-word,
picture-picture, word-picture, picture-word).  The memory phase
involved the presentation of stimuli which were either seen in the
acquisition phase (old) or unseen (new).  Results indicated firstly
that the ERP waveforms to word and picture stimuli differed in both the
acquisition and memory phases, pictures showing a striking bipolar
scalp distribution (frontal negativity and parietal positivity), while
words revealed a more equipotential distribution across the scalp.
Throughout acquisition, both incongruous pictures and words elicited
enhanced N400s, congruous pictures revealing enhanced LPC amplitudes as
compared to words.  The effect of the prime was evident only for
incongruous targets when preceded by an identical prime, attenuating
N400 amplitude.  Enhanced LPC amplitudes and reduced N400 amplitudes
were evident to seen stimuli throughout the memory phase as compared to
unseen, especially when pictures were viewed in acquisition.

Psychophysiological characteristics of narcissism during active and
passive coping
Christina M. McCann1, Sarah Reiff2, Sidney R. Ornduff2, and Robert M. 
1University of North Texas, 2State University of New York at Stony Brook

Psychophysiological characteristics of narcissism were evaluated
during active and passive coping tasks.  Measures of state anxiety,
skin conductance response (SCR), heart period (HP), and preejection
period (PEP) were obtained from 36 undergraduate men with scores in the
upper (n=18) and lower (n=18) terciles on the Narcissistic Personality
Inventory (NPI).  Five 100-dB(A), 1-s tones were presented during two
5-min tasks, with a 27-s countdown phase preceding each tone.  Subjects
could avoid the tone during one task (active coping), but not during
the other (passive coping).  Tasks were presented in counterbalanced
order, and were preceded by 5-min baseline rest periods.  Analyses
focused on psychophysiological reactivity (changes from pretask
baseline) during the countdown phases.
As expected, decreases in PEP and increases in SCR frequency were
greater during active as compared to passive coping (p<.001),
particularly when the active coping task was presented first.  There
was a significant multivariate effect of narcissism on cardiovascular
reactivity during the tasks (p<.005), with significantly greater
decreases in PEP and increases in HP in high as compared to low NPI
subjects.  Increases in SCR frequency tended to be smaller overall in
high NPI subjects as well, particularly when the active coping task was
presented first.  High NPI subjects also showed greater habituation of
SCR reactivity over trials (p<.02), and reported lower anxiety across
experimental periods (p<.06), as compared to low NPI subjects.  These
results provide preliminary support for theoretical links between
narcissism and psychopathy, and suggest that narcissism may be
characterized by deficiencies in behavioral inhibition.

Sex differences in physiological reactivity to the acoustic startle
Loren McCarter and Robert W. Levenson
University of California, Berkeley

Cardiovascular disease is the leading cause of death among both men
and women, yet between the ages of 20 and 50, men are twice as likely
as women to die from this illness. Sex differences in cardiovascular
reactivity to stressful events seem to mimic these sex differences in
cardiovascular disease incidence, with men showing greater
cardiovascular reactivity than women to many stressors.  To the extent
that heightened cardiovascular reactivity is of etiological
significance in the development of cardiovascular disease, it is
tempting to speculate that greater cardiovascular reactivity in men
than in women may help account for observed sex differences in
cardiovascular disease mortality rates. One important and
under-examined issue is whether sex differences in physiological
reactivity are specific to the cardiovascular system or whether they
are also found in other systems such as electrodermal activity or
We measured cardiovascular, electrodermal, and respiratory
reactivity continuously while exposing 100 men (mean age=20.49) and 133
women (mean age=20.37) to a simple acoustic startle stimulus (100 ms,
110 dB burst of white noise). Men showed the familiar response of
greater cardiovascular reactivity to this stressor than women (in
systolic blood pressure and in peripheral vasoconstriction). In
contrast, women showed greater electrodermal (skin conductance level)
and respiratory (respiratory depth) reactivity than men. These findings
suggest that greater reactivity to these kinds of simple stressors in
men, as compared to women, may be most pronounced in the physiological
measures that are of greatest purported significance in the etiology of
cardiovascular disease.

Physiological responses to the emotional qualities of music
Kimberly A. McCoy1, Robert W. Levenson1, and Carol L. Krumhansl2
1University of California, Berkeley, 2Cornell University

Music is used to stimulate, sooth, and adjust our emotional states
in widely-ranging contexts, including recreational, religious,
military, commercial, social, and work settings.  We examined the
emotional effects of listening to six three-minute excerpts of
instrumental classical music chosen to target sadness, fear, and
happiness. Self-reported emotion ratings were obtained from independent
samples of Cornell and Berkeley undergraduates. Physiological measures
(cardiovascular, somatic, respiratory, electrodermal) were also
obtained from the Berkeley sample.
Self-report data from the two samples were highly similar:
discriminant functions derived from Cornell participants' ratings
correctly classified the targeted affect for Berkeley participants at
far greater than chance levels. Although the targeted emotion was
always strongest, all six excerpts produced rather complex blends of
emotions. Thus, we pooled data across excerpts to examine relations
between physiology and the targeted emotions.
In the Berkeley sample, happiness ratings were related to slower
heart rate, longer pulse transit times, and lower systolic blood
pressure. Sadness was related to greater skin conductance and higher
systolic blood pressure. Fear was not associated with any physiological
measure, but rated "anxiety" was associated with higher systolic blood
pressure. This pattern of relationships, which is consistent with a
simple arousal model (i.e., positive affect associated with relaxation,
negative affect with activation), could not be accounted for by
differences in excerpt intensity or familiarity.
We conclude that carefully selected classical musical excerpts can
produce reliable patterns of self-reported emotional experience that is
accompanied by coherent patterns of physiological change.

Fetal heart rate and transplacental stress reactivity: A new paradigm
for psychophysiology
James A. McCubbin, Erma J. Lawson, Jeffrey J. Sherman, and Jane A. 
University of Kentucky College of Medicine

Maternal systemic and transplacental stress reactivity during
pregnancy is understudied, but is a potentially important area for
psychophysiological investigation.  Although stress during pregnancy
has been linked to birth outcome, there is little appreciation for the
potentially pathogenic role of exaggerated maternal neuroendocrine and
circulatory reactivity.  A series of recent studies in our laboratory
suggest that transplacental effects of stress reactivity may have
clinically relevant impact on fetal development and birth outcome.  A
prospective study of blood pressure responses to a standardized
arithmetic stressor in 40 primigravidae found significant correlations
between diastolic blood pressure responses during pregnancy and
subsequent birthweight and gestational age (McCubbin, et al., American
Journal of Obstetrics and Gynecology, in press).  These data suggest
that exaggerated maternal reactivity has transplacental consequences
that may negatively affect the developing fetus.  To assess more
directly the transplacental impact of maternal stress, simultaneous
measurement of fetal heart rate in 20 primigravidae found significant
increases in fetal heart rate, concomitant with maternal responding.
These studies, taken together, suggest that maternal systemic responses
cross the placenta and may negatively affect fetal development.
Psychophysiological studies of neuroendocrine and circulatory stress
reactivity during pregnancy can address a number of important issues:
1) early identification of mothers at risk for low birthweight and/or
premature delivery,  2) assessment of transplacental fetal
responsivity,  and 3)  exploration of new intervention strategies to
minimize the systemic and transplacental pathogenicity of exaggerated
maternal reactivity.  Supported by NIH HL35195, HL32738, and GCRC

The effect of clonidine and naltrexone on blood pressure responses to
James A. McCubbin, John F. Wilson, Jeffrey J. Sherman, Jane A. Norton, 
and George Colclough
University of Kentucky College of Medicine

Endogenous opioids appear to be involved in antihypertensive
pharmacotherapies that inhibit sympathetic nervous system tone, but
thus far this issue has not been adequately clarified.  Although
several investigators report that the opioid blocker naloxone
antagonizes the blood pressure-lowering effects of clonidine, others
have reported no effect.  Moreover, clonidine is used to ease
withdrawal from opiate addiction and has been shown to release
beta-endorphin from brainstem slices and from pituitary tissue in the
rat.  A study of hypertensive patients reports significantly lower
plasma levels of beta-endorphin in the patient group, and treatment
with clonidine normalizes both blood pressure and beta-endorphin
levels.  Based on these findings, it is reasonable to hypothesize that
the blood pressure-lowering effect of clonidine may depend, at least in
part, on an opioid-releasing mechanism.  The present study examined the
effects of clonidine and the opioid receptor antagonist naltrexone on
blood pressure responses to stress in young adults with mildly elevated
blood pressure.  Results suggest that clonidine reduces baseline blood
pressure levels, but has no effect on blood pressure responses to
stress.  Opioid blockade partially antagonizes the effect of clonidine
on diastolic blood pressure levels at rest, but does not appear to
interact with clonidine on blood pressure reactivity.  These results
indicate that the blood pressure lowering effects of clonidine are
mediated via both opioid and nonopioid mechanisms.  Supported by NIH
HL35195, HL32738, and GCRC RR2602

Genetic influences on the spontaneous EEG: An examination of
15-year-old and 17-year-old twins
Kathryn A. McGuire, Joanna Katsanis, and William G. Iacono
University of Minnesota

This study investigated the presence of genetic influences on the
spontaneous EEG (electroencephalogram) by examining the twin similarity
of 15- and 17-year-old monozygotic (MZ) and dizygotic (DZ) twin pairs.
A total of 50 (MZ=33, DZ=17) 15-year-old and 69 (MZ=45, DZ=24)
17-year-old twin pairs were assessed.  Spontaneous EEG activity was
recorded at Cz referenced to linked ears, O1-T5, and O2-T6 while
subjects rested with eyes closed.  Measures of both relative and
absolute power were derived.  Spectral powers were calculated for the
delta, theta, alpha, and beta frequency bands.  The intraclass
correlations in all bands across all sites were significant for the MZ
twins, indicating moderate to high degrees of similarity on these
measures.  Overall, the DZ intraclass correlations were significantly
lower than those of the MZ twins.  This pattern of findings suggests a
significant genetic influence on the spontaneous EEG of middle
adolescent children.  Examination of developmental trends revealed that
twin similarity in EEG tended to increase from age 15 to 17 as assessed
by relative and power spectra.  Further investigation of developmental
trends for relative power spectra revealed a significant decrease in
theta activity and a significant increase in beta activity from age 15
to 17.  Analysis of absolute power spectra revealed a significant
decrease in total power in virtually all frequency bands across all
sites from age 15 to 17.  These findings suggest the presence of
developmental changes in EEG during adolescence.

Kid's cortical ERPs: Emotion and attention in picture processing
Mark H. McManis, Margaret M. Bradley, Bruce N. Cuthbert, Harald Schupp, 
and Peter J. Lang
University of Florida

Data from our laboratory has demonstrated that children, like adults,
are physiologically reactive to affective pictures: Heart rate, skin
conductance, facial muscle activity, and startle reflexes vary with the
affective content of the pictures. Recent studies have shown that ERPs
are also related to the affective picture content in adults, with
larger positivity for arousing, relative to neutral, pictures. The
present study extends work with the cortical evoked response to child
Children (7-10 years old) and adults each viewed a series of
pictures  varying in affective content. An extra effort was made to
include pictures that young children find very unpleasant; parents of
the children were fully informed of the content of the pictures. Heart
rate, skin conductance, and  corrugator region EMG were recorded during
a 6 s picture viewing period. An acoustic startle probe was delivered
during the presentation of each picture. EEG was recorded from F3, Fz,
F4, Cz,  and Pz, referenced to linked ears. All cortical sites were
corrected off-line for eye movement.
Heart rate, skin conductance and corrugator EMG responses were
similar to the patterns obtained previously in adults and children. As
found in cortical studies of adults, children showed a fronto-central
negativity to picture onset peaking  at 400-500 ms, and marked
positivity at Pz. Children also showed the pronounced P3 to task-
irrelevant startle probe. The results encourage further use of the
picture paradigm in research with children.

Normal and disabled readers discriminate alliterating from
non-alliterating spoken words: An event-related brain potentials study
W. Brian McPherson
University of Arkansas for Medical Sciences, Little Rock, Arkansas

	Sixteen adolescent disabled readers and 16 age-matched normal
	readers judged whether spoken words did or did not alliterate.
Single syllable, real word digitized speech stimuli were used.
Measures included event-related brain potentials (ERPs) recorded from 5
pairs of bilateral electrodes, reaction times, and response accuracy.
Reading disabled subjects were split based on the Bradley Oddity test,
a test requiring auditory phonological discriminations.  The 6 disabled
readers who responded correctly to all 24 trials on the test were
called Phonetics.  The remaining 10 disabled readers, who had, on
average, 4 incorrect responses, were referred to as the Dysphonetics.
	The normal reading controls had both higher accuracy and faster
	responses than either reading disabled group.  The Phonetics
were marginally faster than the Dysphonetics, but the groups did not
differ in accuracy.  The normal readers' ERPs revealed an N400 priming
effect for alliterating targets.  The alliterating targets were
significantly less negative than the non-alliterating targets from 250
to 450 msec post target onset.  This effect was widely distributed
bilaterally across the scalp and reached its peak over parietal and
temporal-central-parietal regions.  The Phonetic group did show some
similar N400 priming, but only over temporal-central-parietal sites.
The Dysphonetic group did not present any priming for this epoch,
however, they did show near significant priming over the left parietal
site between 400 and 500 msec post target onset.

Increased fMRI activation of anterior cingulate to infrequent compared
to frequent stimuli in an auditory oddball task
Vinod Menon, Kelvin O. Lim, Judith M. Ford, and Adolf Pfefferbaum
Stanford University and VAPA Health Care System, Palo Alto

Event-related potentials (ERPs) to an auditory oddball have been
investigated extensively.  Recent brain imaging studies have indicated
activation of the anterior cingulate during the oddball task when
compared to a resting baseline.  In the present study, functional
magnetic resonance imaging (fMRI) was used to investigate activation of
the anterior cingulate by directly comparing the activation elicited by
the target infrequent tone to that elicted by the non-target frequent
tone.  Four right handed subjects participated in the study.  The
stimuli were eighty 1000Hz and twenty 2000Hz tones presented randomly
with an interstimulus interval of 8s.  Subjects pressed a response ball
to infrequent tones.  Axial functional images were acquired every 4s at
12 locations with a slice thickness of 7mm.  Images were normalized to
stereotaxic coordinates and analyzed using Statistical Parametric
Mapping.  Foci of activation were detected using a reference function,
consisting of +1 for the infrequent tone and -1 for the frequent tone,
convolved with the haemodynamic response.  Results reported here are
for Z>2.3 (p<0.001,uncorrected).  Increased activation of the anterior
cingulate (Brodmann's Areas 24 and 32) to infrequent tones was detected
in all four subjects.  There was greater fMRI activation in the
anterior cingulate to the infrequent than the frequent tone.  The
results suggest that the activation of the anterior cingulate is not
merely related to enhanced generalized attention during the task, and
may have implications for understanding the neural substrates of the
P300 component of the ERP.

EEG characteristics in adolescent males at risk for developing
Amy K. Mertz, Kathryn A. McGuire, Joanna Katsanis, and William G. Iacono
University of Minnesota

It has been well documented that male offspring of alcoholic fathers
are at a higher risk than the general population for developing
alcoholism.  In addition, several studies have shown that alcoholics
exhibit unusual EEG characteristics, such as decreased alpha activity
or increased beta activity.  However, it is unclear whether these EEG
characteristics are caused by chronic alcohol ingestion, or whether
they pre-date the onset of alcoholism.  In order to examine this
possibility, we recorded the resting EEG at Cz referenced to linked
ears, O1-T5, and O2-T6 in three samples of 14- and 15-year-old male
offspring of a) fathers with alcoholism (high-risk; n=37), b) fathers
with alcohol abuse (moderate-risk; n=16), and c) fathers without a
diagnosis of alcoholism (low-risk; n=21).  Preliminary analyses of the
data indicated that the high-risk group had significantly lower levels
of alpha activity in the relative and absolute power spectra across all
recording sites relative to those of the low-risk group.  Significantly
higher beta activity was also seen  in the relative power spectra of
the high- and moderate-risk groups at O2-T6 and of the moderate-risk
group at O1-T5 compared with the low-risk group.  These preliminary
findings suggest that certain components of the EEG can function as
possible biological markers for the development of alcoholism.

Sequential changes in the auditory evoked potentials during target
H. J. Michalewski, A. Starr, T. Aguinaldo, and M. Roe
University of California, Irvine

Stimulus sequence effects were examined in auditory detection tasks
(oddball paradigm) that required (1) a button press to targets and (2)
a mental count of the targets.  Sequence effects were evaluated by
sorting and averaging single trials to nontargets relative to the
position in the sequence following a target.  Normal subjects (n = 13)
listened to a series of 300 tones consisting of 240 nontargets (low
tone) and 60 targets (high tone).  Targets occurred randomly in the
sequence with the restriction that the succession of two targets was
not allowed.  Tones (250 ms) were presented every 2 sec.  Long time
constant (16 sec) recordings were made from midline and lateral scalp
electrode placements.  Sweep duration was 1.44 sec and included a 0.76
sec prestimulus period.  Nontarget potentials varied in amplitude as a
function of position following the target.  A slow prestimulus negative
potential (readiness potential, RP), the N100, and a late slow wave
were smaller to nontargets immediately following the targets than to
nontargets immediately before the targets.  The amplitudes of these
components recovered as a linear function of the number of nontargets
in the sequence.  P200 amplitude was larger to nontargets following the
target than to nontargets immediately preceding the targets.
Sequential changes (RP, N100) were either absent or reduced in the
count condition; P200 amplitude changed for both press and count
conditions. The results describe the presence of brain activity
subserving response preparation and stimulus processing that are
systematically related to the position of the stimulus in a sequence.

Conditioned gamma band coherences between visual and somatosensory
brain areas after differential classical conditioning of painful stimuli
Wolfgang Miltner1, Matthias Arnold1, Herbert Witte1, and Christoph 
1Friedrich-Schiller University of Jena, Germany, 2Eberhard-Karls-
University of Tuebingen, Germany

In a recent ERP study (Waschulewski-Floru=DF, H., Miltner, W.,
Brody, S., & Braun, C. (1994). Classical conditioning of pain responses.
International Journal of  Neuroscience, 78, 21-32) we demonstrated that
during the last 250 ms of presentation of a visual CS+ reinforced by a
succeeding painful stimulus (UCS) there was a significant increase of
electrical negativity over the receptive field of the stimulated finger 
at SI. This negativity did not develop during the last 250 ms of CS-
presentation. Results indicate that occipital brain areas processing the
CS+ communicate with primary somatosensory brain areas processing the 
UCS.coherence analysis. 16 subjects participated in the experiment. The 
UCS consisted of intracutaneous electrical stimuli. Red and green light 
served as CS+ and CS-. UCS was applied during the last 10 ms of CS+ 
EEG data were recorded from 31 electrodes, corrected for artifacts and 
then submitted to a coherence procedure (calculated for different 
frequency bands) using pairs of electrodes on occipital, primary and 
secondary somatosensory areas of both hemispheres. Significant 
differences of coherence between CS+ and Cs- conditions were onserved 
only within the gamma band with stronger coherences between electrodes 
of visual and somatosensory areas during CS+ condition than during CS- 
condition. Results indicate that during conditioning brain areas 
involved in the information processing of the visual CS+ became rapidly 
synchronyzed within the gamma band (i.e, around 40 Hz).

Cortical reorganization in phantom limb pain changes after regional
anesthesia of the amputated limb
P. Montoya, N. Birbaumer, W. Lutzenberger, H. Flor, W. Grodd, K. Unertl, 
and W. Larbig
University of Tuebingen

In an earlier paper (Flor et al. 1995), we have demonstrated extensive
reorganization of the primary somatosensory projection areas in phantom
limb pain patients. The magnitude of reorganization correlated 0.95
with the subjective phantom limb pain. Since phantom limb pain persists
usually for years, often for life-time, the plasticity of the
reorganization itself constitutes a critical question for its
therapeutic modification. To test the stability of cortical
reorganization, nine unilateral upper limb amputees were tested before
and after plexus brachialis blockade of the stump. SEPs elicited by
pneumatic stimulation of the 1st and the 5th finger as well as of the
lip (bilateral) were recorded from 60 electrodes. The functional
somatotopic organization of the somatosensory cortex representing the
stimulated regions was obtained using an equivalent current dipole
model and anatomical information provided by individual MRI scans.
Significant group differences were found in the session before
anaesthesia, indicating that in patients with phantom limb pain the
cortical representation of the lip ipsilateral to the amputation side
was significantly more shifted towards the finger region than in
pain-free patients. In addition, a significant reduction in the
magnitude of cortical reorganization was observed in patients without
pain during anaesthesia compared to patients with enduring pain. These
results strongly support the plastic (Hebbian) nature of cortical
reorganization and provide further support for the hypothesis that
plastic changes in amputees are strongly associated with the experience
of phantom limb pain.
Supported by the German Research Society (DFG).

Dissociative experiences, tracking task workload, and EEG
Richard A. Moraga1 and William J. Ray1
1Penn State University

In this study we assess individual differences in a simulated flight
performance task.  Thirty individuals scoring high (n=15) or low (n=15)
on the Dissociation Experiences Scale (DES) performed a series of
flight tracking tasks Conditions varied according to flight mode (i.e.,
automatic vs. manual) and task workload (i.e., additional monitoring
tasks).  We describe their 17 site EEG using linked ears as reference;
256 HZ sampling rate; .03-35 HZ bandpass; 2 EOG channels for artifact
correction.  The EEG data were Fourier analyzed into theta, alpha, beta
and gamma frequency bands and log transformed for ANOVASs (BMDP2V).
EEG task engagement indices, similar to those previously reported (Pope
and Bogart; 1993), using ratios of beta to theta and alpha were also
examined. Main effects involving EEG bands were found for tracking mode
in all bands but theta (e.g., F(1,28)=5.52,p=.03, for gamma activity);
and in all bands for task workload condition (e.g., F(3,84) =26.17,
p<.001, for theta).  Significant interaction effects were also found
for tracking mode and dissociation and for tracking mode, workload
condition, and dissociation.  Analysis of an EEG engagement index
yielded a significant main effect for tracking mode and significant
interaction effects for tracking mode and workload,mode and site, and
workload and site.  These results and site specific EEG findings will
be discussed using performance data, topographic mapping, and current
theories of attention.  (Supported in part by NASA-Langley)

The olfactory P300 in young and older adults
Charlie D. Morgan1, Mark W. Geisler2, James W. Covington3, Dennard W. 
Ellison3, John Polich4, 
and Claire Murphy2,3
1San Diego State University/University of California, San Diego Joint 
Doctoral Program in Clinical 
Psychology, 2UCSD Medical Center, 3San Diego State University, 4The 
Scripps Research Institute.

Olfactory event-related potentials  (OERPs) were recorded monopolarly
at the Fz, Cz and Pz electrode sites in 16 young adults (8M/8F;
M=23.22.4 yrs) and 16 older adults (8M/8F; M=69.75.7 yrs) using a 45
sec inter-stimulus interval and amyl acetate as odorant in a
single-stimulus paradigm.  A cognitive magnitude estimation task was
employed to produce a P3 component.  Results demonstrated that young
adults had significantly larger amplitudes and shorter latencies than
older adults for N1, P2, N2, and P3 components. In addition, older
males produced the smallest P2 and P3 amplitudes.  Latencies, however,
were not significantly affected by gender for any of the components.
Morphology and timing of the OERP components resembled that found in
the visual and auditory systems with the maximal P3 amplitude occurring
at the Pz electrode site for young adults and a more equal amplitude
distribution across electrode sites for older adults.  Further, P3
latency followed the N1-P2-N2 complex by 100-200 ms, as found in the
other modalities.  The effects of age on OERP amplitude and latency
were largest for the P3 compared to the earlier components.  These
age-related changes may be due to degenerative changes in the
peripheral olfactory system associated with aging and cognitive slowing
of the central nervous system that has been shown to affect ERPs in the
auditory and visual modalities.  Results suggest that the OERP may be
useful in assessing age-related decline in olfactory functioning in
both clinical and research settings.  Supported by NIH grant# DC02064
(CM) and training grant# DC00032 (MWG).

Fate of four Sokolovian deductions in the electrodermal and vasomotor
components of the orienting reaction: The picture remains confused
James W. Morrison1, John J. Furedy2, and Pierre Flor-Henry1
1Alberta Hospital Edmonton, 2University of Toronto

Sokolovian orienting reaction (OR) theory introduced in the sixties
has a number of testable deductions for its two most important
components, the electrodermal SCR and the peripheral vasomotor response
(VMR). Of these deductions, there are the following clearly derivable
four: habituation (decrease over repeated trials), reinstatement
(increase to change following repetition), super reinstatement
(response to change exceeding that to the first of a series of repeated
trials), and dishabituation (response to repeated stimulus immediately
following change trial exceeding that to the trial immediately
preceding the change trial). These deductions were tested in the
cross-modal preparation (e.g., Furedy, 1968, J.  Exp. Psych., 70-78)
where, following 15 repeated (tone or light) trials, a change trial
(light or tone) occurred on the 16th trial, and the repeated trial
(tone or light) on the 17th trial.  Stimulus duration was 300 ms, and
the intertrial interval varied around 45 s (mean). Sex and laterality
(bilateral recording in completely right-handed subjects) were also
varied between- and within-subjects, respectively. In the SCR (based on
31 and 25 males and females, respectively) all except the
dishabituation effect emerged reliably, p < 0.01, and neither sex nor
laterality yielded significant main or interactive effects. In the VMR
(18 and 13 males and females, respectively--data loss due to recording
difficulties), there was no habituation or dishabituation, but
significant reinstatement, p < 0.01, and super reinstatement, p < 0.05,
effects were found. Aside from the SCR/VMR differences (reported
elsewhere in the literature), the results also suggest (considering the
stimulus-omission preparation, which generally yields dishabituation
but not reinstatement effects) the need to distinguish
stimulus-specific and state-related novelty effects.

It's shocking! Conditioning affective categories
Brad Moulder, Margaret M. Bradley, Bruce N. Cuthbert, and Peter J. Lang
University of Florida

A stimulus paired with aversive electric shock produces an augmented
startle response, compared to a stimulus not specifically associated
with shock. In this experiment, we investigated whether this type of
hedonic learning is stimulus specific.  To test this, a variety of
pictures of the same affective valence -- either pleasant or
unpleasant-- were presented as to-be-conditioned stimuli (i.e. paired
with electric shock) to determine whether startle responses are
potentiated for members of the same valence category.
	Pleasant and unpleasant pictures were chosen from the
International Affective Picture System, and each was presented for a 6
second viewing interval.  To assess sensitization, startle reflexes
were measured in the viewing period both before and after shock
exposure.  In a subsequent acquisition phase, different pictures from a
specific valence category (pleasant for some participants, unpleasant
for others) were paired with electric shock for 9 trials (CS+);
pictures of the other category were never paired with shock(CS-).
Startle probes were presented during acquisition as well as during a
subsequent extinction phase. In extinction, new pictures of each
valence, as well as pictures previously encountered during acquisition,
were presented and probed.
	Replicating earlier studies, startle reflex magnitude increased
following exposure to shock. Pictures paired with shock led to
potentiated startle responses, compared to those that were not. In the
extinction phase, reflex magnitudes for pictures never seen before, but
of the same valence as the CS+, were similar in size to those elicited
for previously shocked pictures. Taken together, these data suggest
that hedonic learning shows generalization by valence category.

Individual differences and startle response modulation
Kevin B. Muse, Almut I. Weike, and Alfons O. Hamm
University of Greifswald, Germany

Eysenck's theory of extraversion and Zuckerman's theory of sensation
seeking suggest that extraverts and sensation seekers are motivated to
engage in a wide range of behaviors to compensate for arousal
deficits.  Most psychophysiological studies, however, have not examined
the motivational dispositions of these subjects.  To examine affective
processes in extraverts and sensation seekers, we have tested startle
modulation in a large sample (N= 90; extraverts and sensation seekers
are defined by median splits on the Eysenck Personality Inventory and
the Sensation Seeking Scale) using the picture perception paradigm.
During viewing of negatively valenced slides, the startle eyeblink
response is potentiated, while during viewing of positively valenced
slides, the response is inhibited.  Recent findings show that this
modulation effect occurs only when stimulus materials are sufficiently
arousing.  If extraverts are underaroused, they should show less
startle response modulation than introverted subjects.  To assess
orienting, arousal, and interest elicited by slide material, we
measured electrodermal activity, heart rate, blood pressure, and
subjective emotion and interest ratings (emotion and interest ratings
were performed separately during a second round of slide viewing).  The
data from this study show the expected interaction between valence and
arousal characteristics of slide stimuli for the startle response.
Preliminary analyses also reveal that the startle responses of low
sensation seekers are about twice as large as those of high sensation
seekers, showing that high sensation seekers are less responsive to
these stimuli than low sensation seekers.

Gastric myoelectrical reactivity to emotional stimuli
Eric R. Muth, Robert M. Stern, Julian F. Thayer, and Kenneth L. Koch
The Pennsylvania State University

The purpose of this study was to investigate gastric myoelectrical
reactivity to emotional stimuli with known autonomic nervous system
(ANS) response patterns. Electrogastrograms (EGGs) and
inter-beat-intervals (IBIs) were recorded during baseline (BL), shock
avoidance (SA), and cold face stress (CF), each lasting 4 min.  During
BL, subjects sat quietly.  During SA, subjects played a reaction time
game with the threat of shock for slow reaction times.  SA produces
sympathetic nervous system (SNS) activation and parasympathetic nervous
system (PNS) withdrawal.  During CF, subjects sat quietly with a bag of
cool water (6-9 degrees Celsius) on their forehead.  CF produces PNS
activation and SNS withdrawal through a dive reflex-like response.
After each task subjects reported their experience of various emotions,
ranging from serenity to excitement.  IBIs decreased significantly
during SA compared to BL (t[23]=1.79, p<.05) and increased
significantly during CF compared to BL (t[23]=4.03, p<.05), supporting
the efficacy of the tasks in inducing ANS changes.  Interest,
excitement and activation were all significantly increased and serenity
and relaxation were significantly decreased during SA compared to BL
(p<.05). The CF task did not induce any emotional changes compared to
BL. The percentage of tachyarrhythmic activity (abnormal gastric
activity) in the EGG was significantly higher during SA compared to BL
(t[19]= 2.69), p<.05) and there was marginally less 3 cpm (normal
gastric activity during SA than during BL (t[19]=1.82, p<.10).  The CF
task did not significantly effect EGG activity compared to BL. The SA
task was effective in inducing emotional and gastric myoelectrical
changes characteristic of SNS arousal and PNS withdrawal.  In
conclusion, the EGG can be used to study gastric myoelectrical
reactivity to emotional stimuli.

Assessing the relationship of frontal activation asymmetry and
affective style: Methodological considerations
Ewald Naumann, Dirk Hagemann, Christian Fleischer, Oliver Diedrich, and 
Dieter Bartussek
University of Trier

Several studies demonstrated that intensity of emotional response to
affectively evocative stimuli of specific valence can be predicted by
resting EEG asymmetry in anterior regions of the brain.  Left frontal
activation was associated with reports of more intense positive affect
in response to stimuli of positive valence.  Right frontal activation
was associated with more intense reports of negative affect in response
to stimuli of negative valence.  While reported results suggest a
remarkable consistency of phenomenon across separate investigations,
each study differs in means of emploied methods to gain results.  The
aim of the present study is to replicate these results and to
investigate their independency of emploied procedures.  The design of
the present study was kept close to that of the example studies. EEG of
37 subjects was recorded from electrode locations F3, F4, T3, T4, C3,
C4, P3, P4, A1, A2 during eight 1-min baselines resting periods with
subjects keeping eyes open and closed. All placements were referenced
to Cz. After recording of EEG, subjects looked at 30 slides selected
out of the International Affective Picture System. Subjects rated their
emotional reaction to each slide immediately after seeing it using 8
emotion-rating scales. A Fast Fourier Transform was performed on the
EEG data to obtain measures of cortical activation asymmetry. All
different procedures described in the example studies were performed to
the same set of raw data.  After all asymmetry metrics and affective
indices were computed, correlations between these measures were
calculated to evaluate invarianve of methods and to assess the
psychophysiological relationships. Results indicate that there is no
invariance of emploied methods: Different procedures of assessing the
psychophysiological relationships in question led to different
outcomes.  Supported by Deutsche Forschungsgemeinschaft

The dissociation between skin conductance and secondary task reaction
time in a visual discrimination task
David L. Neumann, Ottmar V. Lipp, and David A. T. Siddle
The University of Queensland

A dissociation between two putative measures of processing resources,
skin conductance responding and secondary task reaction time (RT), has
been observed in auditory discrimination tasks. The present research
examined this dissociation effect in a visual discrimination task.
Subjects were presented with circles and ellipses and instructed to
count the number of longer-than-usual presentations of one shape
(task-relevant) and to ignore presentations of the other shape
(task-irrelevant). Concurrent with this task, subjects made a speeded
motor response to an auditory probe. Skin conductance responses were
larger during task-relevant shapes than during task-irrelevant shapes.
Probe RT tended to be slower during task-irrelevant shapes than during
task-relevant shapes at probe positions of 50 and 150 ms whereas the
difference was reversed at a 2000 ms probe position. Only the
difference at the 150 ms probe position was reliable. The results
demonstrate that the dissociation effect generalizes to a visual
discrimination task. The time course of secondary task RT to probes
presented following shape onset was investigated in a second
experiment. The nature of the probe positions were varied between four
groups (Group A: 10, 100, 2000 ms; Group B: 30, 150, 3000 ms; Group C:
50, 200, 3000 ms; Group D: 100, 300, 4000 ms). Probe RT during task-
relevant and task-irrelevant shapes showed a parallel downward trend
when probes were presented within 300 ms following shape onset. The
similarity of this time course to the psychological refractory period
(PRP) effect is discussed.

N400 effects related to incongruities in mental calculation problems
Michael Niedeggen, Kerstin Jost, and Frank Roesler
Philipps Universitaet

Retrieval of basic arithmetic facts stored in memory networks
specialized for numerical representation can be strongly affected by
interference. Using simple multiplication tasks, we examined the
influence of different types of incorrect answers on both, reaction
times and event related potentials (ERPs). Following the successive
visual presentation of two digits (a,b), subjects had to verify an
offered solution, which could be either the correct product (a*b) or a
related answer (e.g. a+b, or (a+1)*b, or a*(b+1) ).  In order to
evaluate the time course of spreading activation we varied the stimulus
onset asynchrony (SOA: 200 or 500ms).  Processing of correct and
related false answers could be clearly differentiated on the basis of
ERPs: Similar to semantically incongruous words during sentences,
presentation of false results in simple computational problems elicit a
N400-like wave with a maximum at centro-parietal electrode sites.
Additionally, subtypes of incorrect answers differ with regard to this
component: Latency and duration was shorter for trials easy to reject
(e.g.  deviant sum of the preceding digits) then for those requiring
longer evaluation time (e.g. multiples of the first or second digit).
Elongation of the SOA lead to an enhancement of the N400 effect as well
as a shortening of its onset.  These data suggest that the N400 effect
associated with the verification of simple arithmetic facts parallels
that observed during lexical decision. This supports the idea that
arithmetic facts are stored in networks for which the same functional
principles apply as for semantic networks.

N400 elicited by words and pictures
Arti Nigam, James E. Hoffman, and Robert F. Simons
University of Delaware

	The N400 ERP component is observed during processing of
linguistic stimuli that are semantically unexpected.  In an earlier
study we observed N400 effects with both pictures and words, with the
same ERP scalp distribution for both kinds of stimuli (Nigam, Hoffman
and Simons, 1992).  However, some more recent studies such as that by
Ganis et al. (1996) report substantially different ERP scalp
distributions for pictures and words.
	At issue is whether the N400 ERP component reflects a process
	in a conceptual system that is involved in the processing of
both verbal and visual information.  Thus, it is important to clarify
to what extent the 'word N400' and 'picture N400' are similar, and to
examine causes of observed differences in scalp distribution across
studies.  The present study uses the basic methodology of the above
studies, with the goal of replicating the results of one or the other.
Since the greatest difference between word and picture ERPs has been
noted at the frontal-midline site, this site was included in a larger
set of scalp recordings. The possibility of significant procedural
differences was also addressed.  Word and picture ERPs were compared
within-subjects, using a new, counterbalanced set of stimuli.
	The word N400 was somewhat larger overall than the picture
	N400; however, both showed the same scalp distribution.
Further, the scalp distribution in both conditions resembled that of
the pictures condition reported by Ganis et al.  Thus, the difference
in observed ERP scalp distribution across studies is isolated to the
words condition.  Possible explanations are being explored in a
separate study, the data of which will also be presented.

Some touching situations: The relationship between context and gender
in physiological responses to human touch
Wendy J. Nilsen and Scott R. Vrana
Purdue University

Researchers have found a reliable decrease in heart rate when an
individual is touched by another human being, and this decrease has
been described as reflexive in nature.  However, this conclusion may
not be justified, because all of these studies failed to fully examine
the context in which the touch occurred.  The aim of the present study
was to investigate if changes in the context in which human touch
occurs effect physiological responses to being touched.  Context for
the ten-second touch was manipulated in three ways: professional touch,
in which subjects were touched while having their pulse taken; social
touch, in which the experimenter comforted the subject by touching them
on the wrist; and a no-touch control in which the subjects were told
their pulse was being taken automatically, without being touched.
Social context was also manipulated by employing both same-sex and
opposite-sex touch experimenters.  Heart rate, blood pressure, and skin
conductance were measured in 69 male and 69 female undergraduate
subjects.  The results indicate that there is a complex relationship
between context and gender in physiological responses to human touch.
In the professional touch and no-touch control condition subjects'
heart rate decreased, although the magnitude of the change was mediated
by both the gender of the experimenter and subject.  In the social
condition heart rate increased for most subjects, although again there
was an interaction between the sex of both the toucher and the person
being touched. These data prove problematic for the hypothesis that the
response to human touch is a reflex.  Instead, they suggest that
physiological responses to touch do not occur in an interpersonal
vacuum and that both context and gender are important factors in
determining physiological responses to touch.

N400 as a measure of incongruence within the sentence and across
sentence boundaries.
M. Niznikiewicz, B. F. O'Donnell, P.G. Nestor, S. Karrigan, and R. W. 
Harvard Medical School, Brockton VAMC.

N400, an ERP measure sensitive to semantic context, has typically been
studied using word pairs which differed in semantic relatedness, or
words which violated the preceding context of a sentence.  We report
the relationship of N400 to varying levels of context violation within
and across sentences, and working memory demands.  These preliminary
data were recorded from five normal subjects who read two-sentence
paragraphs from a computer terminal.  A target word in the second
sentence was either congruent (paragraph congruent:PC) or incongruent
(paragraph incongruent:PI) with information provided in the preceding
sentence, which demanded working memory demands, or was nonsensical
with respect to information provided within the same sentence.  The
three conditions were presented in the high and low memory load. ERPs
were recorded from 28 electrode sites over 924 ms epoch to each target
word.  As expected, in the SI condition (traditional incongruency
paradigm), N400 amplitude was sensitive to nonsensical target words,
and the latency did not differ as a function of memory load (402 ms in
low vs 406 ms in high memory load).  N400 latency was sensitive to
memory load in both PC and PI condition, which required holding the
contents of a paragraph in memory to make a semantic judgement.  The
N400 latency in the load memory load, PC condition was 478 ms, and in
the PI condition, it was 490 ms.  In the high memory load the N400
latency was 524 ms in the PC, and 466 ms in the PI condition.

Effects of nicotine on stress-related changes in baroreceptor activity
Matthew Orenstein and Larry D. Jamner
University of California, Irvine

Pharmacological and psychological stressors have been associated with
reductions in baroreceptor activity.  However, few studies have
investigated the combinatory effects of these stressors on baroreceptor
activity.  This study examined the moderating effects of nicotine on
baroreceptor activity during resting and psychologically stressful
conditions as a function of gender.  Nine men (5 nonsmokers, 4 smokers)
and 11 women (6 nonsmokers, 5 smokers) were exposed to a Paced Auditory
Serial Addition Task (PASAT) while continuous measures of systolic
blood pressure, diastolic blood pressure, and cardiac
inter-beat-intervals were obtained under nicotine (delivered
transdermally) and placebo conditions.  Nicotine and placebo sessions
were counterbalanced within-subjects and conducted in double blind
fashion.  Baroreceptor activity was assessed in terms of sensitivity
(ms/mmHg) and adjustments (percent of cardiac cycles related to
sensitivity), using the sequence method (Watkins et al., 1995).
	Overall, the PASAT significantly reduced baroreceptor
sensitivity and adjustments under both placebo and nicotine
conditions (ps <0.007).  Compared to placebo, nicotine was observed to
marginally reduce resting baroreceptor sensitivity (22.4 ms/mmHg vs.
18.4 ms/mmHg; p=0.083), but not adjustments.  In contrast, during math
stress, nicotine was observed to significantly reduce both sensitivity
(12.3 ms/mmHg vs. 16.9 ms/mmHg; p=.038) and adjustments (5.6% vs. 7.5%;
p=.003).  The effect of nicotine on baroreceptor sensitivity was found
to be a function of gender.  During the math task, compared to placebo,
nicotine was found to reduce the baroreceptor sensitivity of men
(21.1ms/mmHg vs. 12.0 ms/mmHg; p=0.035), but not women (13.6 ms/mmHg
vs. 12.6 ms/mmHg; ns).  These effects were found to be independent of
smoking status.

Interaction of emotion with magnitude, habituation and relationships of
P300 and startle responses in children
Edward M. Ornitz
University of California at Los Angeles 

Large P300 components of the ERP accompany startle in response to loud
auditory stimuli.  Startle was measured both as orbicularis oculi EMG
and vertical EOG, and P300 was recorded at Pz in 34 normal 7-to-11 year
olds in response to 40 104 dB noise bursts presented every 23 s.  Both
the startle response and the P300 habituated toward asymptotic levels
after the first 28 trials, suggesting that both startle and the
subsequent cognitive evaluation of the startling stimulus, reflected in
the P300, are modulated by a common neurophysiological mechanism
extrinsic to the direct startle pathway.  A modest significant
correlation between the P300 and the vertical EOG peak latencies for
the initial trials suggests that cognitive evaluation of the startling
stimulus may also include evaluation of the reflex response to that
stimulus.  Analyses of the within-subject associations between startle
and P300 showed that amplitudes and habituation varied independently
within the individual subject, suggesting that the P300 is not a
component of the startle response.  Rather, it reflects an evaluation
of the startling stimulus, decreasing in amplitude as the surprising
value of the startling stimulus decreases with habituation.  Ratings of
arousal and affective valence available from 21 of these subjects
revealed significant correlations between arousal and both magnitude
and habituation of P300 but not of startle, reflecting the influence of
one dimension of emotion on the cognitive evaluation of startle, but
not on startle per se.

Ignored yet processed: An examination of a negative ERP modulation
elicited by repeated unattended words
Leun J. Otten, Michael G. H. Coles, and Emanuel Donchin
University of Illinois at Urbana-Champaign

We previously reported studies in which to-be-attended and
to-be-ignored words were presented concurrently under demanding viewing
conditions.  Subjects viewed simultaneously presented letter strings,
some of which were repeated on the subsequent trial.  They made lexical
decisions about the string designated as relevant by its color and a
spatial cue, and gave a response only when the string was a nonword.
When an attended word was repeated, it evoked an event-related
potential (ERP) that was more positive-going than the ERP elicited by
the word's first presentation.  By contrast, repeating an unattended
word evoked a more negative-going ERP.
These phenomena were further investigated in two experiments.  In
Experiment 1, words were repeated in the same or a different font.  The
negative ERP modulation was observed when words were repeated in a
different (but not the same) font.  In Experiment 2, words were always
repeated in the same font while different responses were given to words
and nonwords.  For those subjects who showed the negative ERP
modulation, responses were faster but less accurate when the unattended
string was repeated.
These results suggest that to-be-ignored words can be processed to
at least the level of their orthographic identities.  This processing
may influence subsequent performance by altering the way in which both
to-be-attended and to-be-ignored words are selected or suppressed.

ERP waves of ADHD- and normal control children in an AX-continuous
performance test
C.C.E. Overtoom, C. Kemner, H. van Engeland and M.N. Verbaten
Utrecht University, The Netherlands

The CPT-AX developed by Halperin et al. (1988) aims to differentiate
inattention and impulsivity by the type of errors that were made.  1)
Impulsivity: a response to a letter other than 'X' following an 'A'.
2) Inattention: misses and responses to an 'X', not preceded by an
'A'.  In the present study, we tried to replicate earlier findings that
ADHD- children showed deficits in impulsivity and inattention, and we
investigated if those impairments were reflected in the P3 amplitude.
Sixteen ADHD-children and sixteen control children performed the
Halperin version of the CPT-AX. The children had to press a button upon
the appearance of the letter X after the A (10%). ERPs were recorded at
Oz, Pz, Cz, and Fz.
ADHD-children performed worse compared to control children, but
only in the inattention score. The P3 amplitude in response to the
target, as well as the P3 to the warning stimulus, were smaller in the
ADHD-children compared to the controls. The response to letters not
being X's but preceded by an A (A-not-X), evoked a larger
fronto-central N2 in the ADHD group compared to the control group.
The smaller parietal target P3 of the ADHD-children and the absence
of a deficit in impulsivity may indicate that inattention is higher
in the ADHD-group. The larger frontal N2 of the ADHD-children to the
A-not-X letters, has not been reported before and may reflect a
disinhibitory effect involved in these children's greater inattention.

Cardiovascular dynamics of social and asocial stressors:
Generalizability of reactivity across stressors
Rebecca Palacios-Esquivel and Joe Tomaka
University of Texas at El Paso

Cardiovascular reactivity (CVR) to stress is implicated in the
etiology of cardiovascular pathology.  Most research, however, has
focused on CVR to asocial laboratory stressors (e.g., mental
arithmetic, cold pressor) which are popular because they can be
administered in a standardized manner.  These studies frequently assume
that reactivity to asocial laboratory stressors generalizes to real
life social stress.  Despite the importance of the social/asocial
dimension, few studies have directly compared physiological responses
between asocial and social stressors. The few studies that have
compared social and asocial stressors have found only marginal
associations, but such comparisons are indirect, and/or the studies
suffer from a number of limitations.  The purpose of this study was to
examine more closely  the relationship between cardiovascular reactions
to asocial and social stressors.  Subjective, behavioral, and
physiological responses were compared across the performance of two
active coping tasks (one social and the other asocial).  Physiological
responses consisted of heart rate, preejection period, cardiac output,
systolic blood pressure, diastolic blood pressure and total peripheral
resistance.  In general, high correspondence between asocial and social
tasks was found for HR (r = .54), PEP (r = .57), CO(r = .53), and DBP
(r = .36), but not for SBP, MAP, and TPR (all NS). Canonical analyses
improved prediction within variables (canonical r's ranging from .40 to
.88) and across variables (canonical r = .87).  We conclude that there
is high correspondence between reactivity to asocial and social
stressors and that the choice of statistical procedure can affect
interpretation of data addressing task correspondence.

Epinephrine, arousal, and emotion: A modified replication and extension
of Schachter and Singer's classic experiment
Stephen Palmer, Elizabeth Mezzacappa, Edward S. Katkin, Stefan Wiens, 
Christopher Saunders, 
Richard Hilliard, Robert Vincent, and Rollin Gallagher
State University of New York at Stony Brook

Schachter and Singer (1962, Psychological Review, 379-399) reported
that epinephrine induced arousal interacts with social context to
affect emotional experience and expression.  Marshall and Zimbardo
(1979, Journal of Personality and Social Psychology, 970-988) have
argued, however, that epinephrine itself induces negative emotion and
that environmental cues are relatively unimportant.  This experiment is
a modified replication of Schachter and Singer's original experiment,
and addresses some of the criticisms that were directed at their study.
Subjects were 21 male undergraduates who were given a titrated
injection of either epinephrine or a placebo (saline) and misinformed
about what symptoms to expect.  Subjects then viewed 6 film clips
selected to induce anger, fear, or amusement (Gross & Levenson, 1995,
Cognition and Emotion, 87-108) and rated their own emotional response
to each clip on 14 different emotion scales.  Results indicated that
epinephrine Ss reported significantly more fear, t = 2.11, p < .05, and
more surprise, t = 2.26, p < .05 than did placebo subjects during the
fear films.  Epinephrine Ss also reported significantly more surprise
during the anger films, t = 2.12, p < .05.  However, epinephrine Ss
also reported significantly more tension during one of the two amusing
films, t = 2.06, p = .05, and more surprise during the other, t = 4.00,
p < .01.  Environmental cues appeared to play some role in determining
Ss' emotions, but they interacted with the negative affective bias
produced by the epinephrine so that only fear, surprise, and tension
were affected significantly.

To each his own: Autonomic reactions to phobic stimuli
Daniela Palomba, Michela Sarlo, Alessandro Angrilli, and Luciano 
University of Padova

Previous research in our laboratory showed marked differences in
autonomic response patterns of normal subjects viewing a surgery video
compared to other unpleasant arousing material.  The present study
explored differences in autonomic response patterns induced by prepared
stimuli (blood, spiders), threatening stimuli, and a neutral control
stimulus, in blood phobics and spider phobics.  Twenty-four subjects
were selected according to their scores on the Fear Survey Schedule,
Spider Questionnaire, and Mutilation Questionnaire.  Subjects viewed
four short video sequences (132 sec each) depicting a threat situation,
a heart surgery, a spider documentary and city landscapes,
respectively.  Heart rate (HR), skin conductance level (SCL), and
spontaneous eye blink rate were measured.  Blood phobics produced HR
increases to both the surgery and the threat video sequences, while
spider phobics increased their HR only in front of the phobia-related
stimulus.  Moreover, in blood phobics HR showed a biphasic response,
with an early increase followed by a decrease in the second part of the
film sequence.  SCL was higher to all unpleasant material in both
groups. Moreover the highest SCL response was observed in blood phobics
in front of the feared stimulus.  The data support the hypothesis of
differential autonomic response patterns between spider phobics and
blood phobics

Differential EEG activation while viewing aesthetically preferred and
nonpreferred natural environments.
Russ Parsons, Louis G. Tassinary, and Daniel Bontempo
Texas A&M University

Previously we reported the inability of facial electromyography and
startle blink responding to discriminate among aesthetically distinct
outdoor environments in a simple rating task (Parsons, Tassinary &
Bontempo, 1995). Brain researchers have drawn important distinctions
between left and right hemispheric EEG activation, and between anterior
and posterior sections of the hemispheres. Positive and negative
affective processing is typically associated with left and right
frontal regions, respectively, while the left and right posterior
portions are respectively associated with verbal and spatial tasks. In
this study, ten sites (F3, F4, F7, F8, T3, T4, P3, P4, O1 and O2) of
EEG activity were recorded while eighty research participants viewed
and rated 40 color slides of aesthetically preferred and nonpreferred
natural areas. The sociality of the rating instructions and
interactions with the experimenter during the ratings were manipulated
between subjects to determine the affect of social context on responses
to nonsocial natural environments. Alpha power was examined for the
first 6 seconds of the 8 second slide presentations, with all sites
showing significantly greater activation (alpha suppression; p< .04 for
all sites) for highly preferred scenes. Examination of hemispheric
asymmetry ratio scores (R-L/R+L; Davidson, 1988) for selected areas
indicated a trend towards (frontal, p= .07) and significantly greater
(parietal, p= .04; occiptal, p= .03) left hemisphere activation
associated with highly preferred scenes. Scene X Gender interactions
for several sites suggested that males tended to show greater
activation for the highly preferred scenes than females, while the
social context manipulations were generally not effective.

Event-related potential measures of auditory priming and recognition as
a function of lag: Age and task differences
J.V. Patterson, C. Cotman, and C. Sandman
University of California, Irvine

Event-related potential (ERP) measures of auditory priming and
recognition were compared in young and older controls.  Three lists of
words matched for word frequency and length were prepared for 3
conditions (priming, recognition, target detection).  Each list was
presented in the auditory modality and contained the following:  (1)
words repeated immediately (0 lag); (2) words repeated after 5
intervening words (5 lag); (3) words not repeated (fillers); and (4)
infrequent target words not repeated.  In the recognition condition,
subjects pressed one of two reaction time (RT) buttons to indicate
whether each word was a first ("new") or second ("old") presentation of
a word in the list.  In the priming condition, subjects made syllable
judgments of each word.  In a target detection condition, subjects
pressed one RT button to infrequent target words.  For both young and
older subjects, RTs were longer to new than to old words in both the
priming and recognition conditions; this difference was diminished at 5
lag, especially for the older group.  ERP differences between new and
old words were observed in the region of N400 and the late positive
component (300-600 msec).  Differences in the ERP between new and old
words were larger for 0 than 5 lag, and for recognition compared to
priming.  ERPs to old words differed with task.  ERP latency was
prolonged as a function of lag.  Age differences were observed for ERP
latency, amplitude, and topography.  These results provide evidence of
ERP differences between priming and recognition, and as a function of
age, task, and lag.

The P50 during wakefulness and sleep: The effects of stimuli parameters
and filter bandpass.
Andrea Perrino and Kenneth B. Campbell
University of Ottawa

A positive wave at around 50 ms (P50) is among the first components of
the auditory evoked potential that is affected by sleep.  The
direction of this effect is, however, equivocal.  Some researchers have
shown a decrease while others have shown an increase in amplitude
during sleep.  Different stimulus parameters and EEG filter settings
can account for these inconsistencies.  Eight young adults (21-25 yrs
old) were presented with "standard" 80 dB SPL,1000 Hz tone pips.
Subjects were tested during Wakefulness, Stage 2, Slow Wave and REM
sleep.  Stimulus duration (4 ms, 16 ms, 32 ms), rise time (2 ms and 8
ms) and interstimulus interval (550 ms, 1100 ms and 2200 ms) were
manipulated.  The high pass digital filter setting had a pronounced
effect. When it was set at 10 Hz, the P50 wave was larger when the
subjects were awake compared to when they were asleep.  This filter
setting removed a slow overlapping negative wave during the waking
state.  The removal of the overlapping negativity caused P50 to appear
artifactually large.  The sleeping ERP was overlapped by a slow
positive wave.  The 10 Hz filter setting removed this slow wave causing
P50 to appear artifactually small.  However, when the high pass filter
was set at 0.1 Hz the slow wave emerged.  Its summating effects caused
P50 to be small when the subject was awake and large when the subject
was asleep. Although these effects were consistent across all stimulus
parameters, differences were not always significant.

Effects of hunger on ERPs to the identification of tachistoscopically
presented food-related and food-unrelated words
R. Pietrowsky, W. Plihal, and J. Born
University of Bamberg

Event-related brain potentials (ERPs) were recorded during the
tachistoscopical presentation of word stimuli to hungry and satiated
subjects.  Stimuli consisted out of food-related words (food names) and
food-unrelated words (names of neutral objects and erotic words).
Presentation of each stimulus was repeated until the subject signalled
identification or was finished after a maximum of 10 repetitions. ERPs
were averaged separately for hunger and satiety, word categories,
number of stimulus repetition, and electrode site. Hungry subjects
needed less stimulus repetitions until identification of food names and
erotic words as compared to satiated subjects. ERPs were characterized
by P2, P3, and slow wave (SW) components. In general, P3 and SW were
increased in satiated subjects as compared to hungry subjects. In the
prior-to-last presentation P3 to food-related words was significantly
increased in hungry subjects as compared to satiated subjects.  Also,
in hungry subjects SW to food-unrelated words was significantly
increased compared to SW to food-related words during the last few
presentations prior to identification. This result is discussed in
terms of the notion that SW indicates inhibition within a lexical
store: At hunger, food-related words are expected.  Identification of
food-unrelated words thus requires the inhibition of the irrelevant
activation of food-related items within a lexical store.

The effects of acute and chronic psychological stress on autonomic,
neuroendocrine, and immune responses: A comparison of caregivers of
spouses with Alzheimer's disease and matched controls
Kirsten M. Poehlmann1, Mary H. Burleson1, John T. Cacioppo1, William B. 
Malarkey1, Bert N. 
Uchino2, Janice K. Kiecolt Glaser1, and Ronald Glaser1
1Ohio State University, 2University of Utah

Although psychological stress has been shown to affect various aspects
of physiological functioning and health, the effects have not been
uniform.  We examined the effects of acute and chronic stress on
cardiovascular, neuroendocrine, and immune function.  We studied 27
women caring for a spouse with a progressive dementia (high chronic
stress) and 37 controls category matched for age and family income (low
chronic stress).  All subjects performed a 12 minute laboratory
stressor (6 minute mental arithmetic and 6 minute public speech).
Measures were taken before (low acute stress) and immediately following
(high acute stress) exposure to the laboratory stressors.  Cellular
immune function was assessed by measuring natural killer (NK) cell
cytotoxicity and the blastogenic response to Concanavalin A (Con A) and
phytohemagglutinin (PHA) of peripheral blood leukocytes.  Chronic
stress was associated with elevated ACTH, enhanced cardiac sympathetic
activation (as indexed by PEP), elevated SBP, diminished blastogenic
responses to mitogens, and decreased NK cell cytotoxicity.  Acute
stress was also associated with elevated ACTH, enhanced sympathetic
activation (as indexed by PEP, EPI, HR and HP), elevated SBP, and
diminished blastogenic responses to mitogens but was associated with
increased NK cell cytotoxicity.  Subsequent analyses revealed that the
differential effect of acute stress on NK cell cytotoxicity was
attributable to an increase in the number of NK cells in peripheral
blood.  These data provide important evidence for the simultaneous
effects of acute and chronic stressors.

Fractal dimension of short EEG time series: Surrogate or real?
Hubert Preissl, Werner Lutzenberger Friedemann Pulvermueller, and Niels 
University of Tuebingen

Fractal dimension has been proposed as a useful measure for the
characterisation of electrophysiological time series. One of the
problems of this approach, is the difficulty to record time series long
enough to determine the fractal dimension of the underlying system.
Using time series of different length we demonstrated that there is a
monotoneous relation between fractal dimension and the number of
data-points. Furthermore an extrapolation scheme was applied in order
to obtain the fractal dimension of the system. This procedure shows
clearly that it is not possible to obtain stationary EEG time series
which are long enough to calculate a good estimation for the farctal
dimension of the system. In addition a monotoneous relation between
time series length and fractal dimension appears with surrogate data
generated from the original EEG time series. We conclude that it is
feasible to use fractal dimension as a tool to characterize the
complexity for short EEG time series, but it is not possible to decide
with EEG data whether the time-series is the output of a chaotic system
or not.  Supported by the German Research Society (DFG PU 97/2 and SFB

Short-term memory and EEG oscillatory subcycles.
William J. Ray1, Richard Moraga1, Christ Molnar1, and Vilfredo 
1Penn State University, 2University of Rome

Recent models of short term memory suggest a model involving "40
hertz" cortical subcycles with lower frequency (5 to 12 hertz)
oscillation. One extension of this model would suggest that individual
memory span differences would be reflected in the EEG, particularly the
ratio of gamma to theta activity. In order to test this model, we
examined 16 individuals who varied in terms of short-term memory on a
digit probe task of 8 elements. In this study EEG was recorded using a
Neuroscan system with 17 EEG sites referenced to linked ears during a
baseline condition and 25 digit probe trails. Overall, on both baseline
trails and test trials, EEG differences were seen in relation to memory
span in the more posterior areas and temporal areas of the cortex. In
terms of baseline activity, there were significant correlations between
EEG beta (20-30hz.) activity in the occipital (r=.52 (O1) & .59 (O2))
and the number of items remembered on the test trials. A similar
pattern was found for gamma (35-45hz) activity.  However, using the
ratio of high to low frequency activity did not improve the
correlational relationship with items remembered.  Analysis of the EEG
during the task itself suggests a complex relationship between memory
span, site, and EEG frequency. During the memory task, Ss with higher
memory scores showed significantly greater high frequency EEG activity
in the posterior areas of the cortex as compared with lower scorers who
showed this activity in the temporal areas.  (Supported in part by
NASA-Langley & NATO)

The ensemble averaged impedance cardiogram: A comparison of scoring
Sarah Reiff, Stefan Wiens, Tamera Schneider, Elizabeth S. Mezzacappa, 
and Robert M. Kelsey
State University of New York at Stony Brook

The validity and utility of two methods of scoring the ensemble
averaged impedance cardiogram were evaluated.  Impedance cardiographic
and electrocardiographic (EKG) recordings from 40 undergraduate men and
women were scored by four raters using a conventional method, involving
ensemble averaging after extensive beat-by-beat editing of waveforms,
and a streamlined method, involving ensemble averaging without
beat-by-beat editing.  Signals were ensemble averaged over consecutive
20-s intervals of the last minute ofediting.  Signals were ensemble
averaged over consecutive 20-s intervals of the last minute of a
pretask baseline period and the first minute of a vocal mental
arithmetic task period. Analyses of averages for each minute focused on
time-based measures calculated from the EKG R-wave (e.g., R-B interval,
R-X interval) and amplitude-based measures derived from the dZ/dt
waveform.  Signals from four subjects (10%) were judged unscorable, so
these subjects were dropped from further analyses.  For the remaining
36 subjects, interrater reliability was satisfactory, ranging from r =
0.82 to r = 1.00 for time and amplitude measures.  Similarly,
correlations between the two scoring methods ranged from r = 0.94 to r
= 1.00 for these measures.  Although paired t-tests comparing the two
methods revealed statistically signific measures.  Although paired
t-tests comparing the two methods revealed statistically significant
differences for a few measures, the differences were negligible (e.g.,
less than 1 ms difference for time intervals).  Scoring was
significantly faster with the streamlined method (less than 1 min per
minute of data) as compared to the conventional method (more than 2 min
per minute of data; p<.001).  The results support the validity and
utility of the streamlined method, and attest to the robustness of the
ensemble averaging procedure.

Attentional inertia in 14-, 20-, and 26-week-old infants
John E. Richards and Theresa L. Gibson
University of South Carolina

Fixations to extended stimuli (> 30 min) by 3 to 5 year old children
and college age students have been explained with a model of
"attentional inertia" (Anderson et al., 1987;  Burns & Anderson, 1993;
Choi & Anderson, 1991).  This model posits that attention to a visual
stimulus increases over the course of a fixation.  The current study
examined fixation durations and HR changes in young infants (3 to 6
months) to determine if the attentional inertia model was applicable to
Infants at 14, 20 and 26 weeks of age (N = 5 per age) were
presented a Sesame Street movie ("Follow That Bird") in session 1,
and computer-generated patterned stimuli accompanied by abstract sounds
in session 2.  The 20 min sessions were videotaped and fixation
direction was judged off-line.  HR was recorded during the entire
There was a decreasing conditional probability that a look would
end as fixations became longer.  This conditional probability
distribution was consistent with the occurrence of "attentional
inertia" in the fixations.  The HR level decreased over the duration of
a fixation and returned to prestimulus level immediately prior to the
fixation offset.  The HR change during infants' fixations to extended
visual stimuli suggest an increasing engagement of attention over the
course of a fixation.  The pattern of fixations and associated HR
changes provide independent evidence of increasing attentional
engagement with increases in look duration.

Validation of the ambulatory measurement of stroke volume by impedance
Harriette Riese, Eco de Geus, and Lorenz van Doornen
Vrije Universiteit, Amsterdam

A detailed insight into the cardiovascular effects of stress in
real-life situations requires the measurement of other variables in
addition to blood pressure and heart rate. Ambulatory impedance
cardiography, as performed by the AMS (Ambulatory Monitoring System)
allows the measurement stroke volume (SV) and cardiac output. At the
present stage of development, however, validation of this technique,
which only recently became available, is needed. The present study had
three goals: 1) to assess the test-retest stability of
impedance-derived SV during laboratory-stress 2) To compare SV levels
during a lab session with ambulatory measurements in comparable field
situations with respect to posture and activity. 3) To compare
ambulatory impedance-derived SV's with SV's as measured with the
Portapres continuous finger blood pressure measurement device.
Thirteen subjects wore both devices during 24 hours. In the morning and
in the afternoon they were exposed to a 5 min serial subtraction task
followed by a 5 min resting period. Test-retest reliability of
impedance-derived SV levels during the resting periods was .63 and
during serial subtraction .86.  The comparison of laboratory levels
with comparable periods during the rest of the registration period
showed low correlations overall. For example, lying quietly in the lab
correlated .29 with sleeping. Silent reading in the lab correlated .11
with reading or watching TV at home. The intra-subject correlations
between SV's as measured with the two devices will be presented, and
limitations of ambulatory SV measurements will be discussed.

Problems with small sample sizes in psychophysiological research.
Todd C. Riniolo and Stephen W. Porges
University of Maryland, College Park

	  Psychophysiological findings, which rely on small samples,
are often insensitive in detecting differences with a small or
medium effect size (ES), and may only achieve statistical
"significance" when an ES estimate is positively biased.  Thus,
only ES estimates from small samples that do not accurately index
the population ES achieve significance.  Without knowledge of the
entire sampling distribution, biased estimates of the population
ES may be common and spurious findings reported.
	  A simulation was conducted consisting of 10,000 computer
generated replications creating sampling distributions of ES
estimates from paired t-tests calculated with 12, 25, 50, and 120
degrees of freedom (df). Samples were drawn from a  normal
population with a small ES (.2) difference (Cohen, 1988).
with the Central Limit Theorem, the mean ES estimate from each
sampling distribution accurately indexed the
.2 ES (independent of df), with a reduction (.31, .21,
.14, .09) in the standard deviation occurring as the df
increased. Using only significant results, the mean of the ES
estimate was inflated (.73, .51, .37, .26) respectively for 12,
25, 50, and 120 df.
	  Since psychophysiological research often relies on small
samples: 1) ES estimates may be positively biased; 2) single
studies may provide unreliable ES estimates; and 3) true
psychophysiological "laws" may not be detected and unreliable
relations may be reported. Implications from these data suggest
that: 1) psychophysiological data registries be made available via
the internet to provide a full range of results for meta-analyses;
and 2) studies use larger sample sizes.

A paced breathing procedure for the inter- and intraindividual
estimation of cardiac vagal tone
Thomas Ritz, Claus Wagner, and Bernhard Dahme
University of Hamburg, Germany

A method for the control of respiratory modulations of respiratory
sinus arrhythmia (RSA) was tested, together with an interindividual
index of cardiac vagal tone. In the first study 22 asthmatics and 22
healthy subjects performed static muscle contractions of the forearm
and forehead. In the second study 12 healthy subjects were exposed to
forehead temperature stimulations with either cold or warm water bags.
Both experiments were designed to influence vagal activity. Total
respiratory resistance (TRR), respiration, and heart period (HP) were
recorded. RSA was determined by the peak-valley method.  Three baseline
measurements of 3 min duration with paced breathing at 8, 13, and 18
cpm were performed. Separate intraindividual regressions against
respiration period were calculated for RSA and RSA normalized for tidal
volume (RSA/VT).  Both RSA and RSA/VT revealed highly positive
correlations with respiration period over the paced breathing
frequencies. The slopes of these regressions were tested as
interindividual indices of vagal tone. For asthmatic and healthy
subjects, the slopes correlated positively with baseline HP, RSA, and
an interindividual index of vagal tone proposed by Grossman and Kollai
(1993). In addition, for asthmatic subjects the slopes were positively
correlated with baseline TRR and the Anger/Irritability subscale of the
Asthma Symptom List. To correct for intraindividual modulations of RSA
by respiration, deviations from the predictions of RSA under paced
breathing were calculated. Corrected and/or normalized indices of RSA
proved to be more sensitive than simple RSA for the detection of
experimental effects.

Cortisol as an opportunistic potentiator of blood pressure responses to
laboratory stress
Mark P. Roy 1, Clemens Kirschbaum 2, and Andrew Steptoe 3
1 Penn State University, USA; 2 University of Trier, Germany; 3 St. 
George's Hospital Medical 
School, University of London, UK.

Large cardiovascular stress responses  per se are unlikely to be
pathogenic, however a sustained response may directly produce damage,
or increase the window of opportunity for other pathogens.  In vitro
and animal studies suggest that cortisol, a stress responsive hormone
in humans, has vasoactive properties,  and is a prerequisite in a rat
hypertension model.  Hypothesis: Concurrent large cardiovascular and
cortisol reactivity will be associated with more sustained blood
pressure (BP) increases,  than isolated large cardiovascular reactors.
90 male probationary firefighters (Age:19-31) took part in a laboratory
session.  After 1 hour of habituation, the session consisted of seven
sampling trials (10 minute): baseline; mental arithmetic; inter-task
recovery; speech task; and three recovery trials.  Within trial BP & HR
were monitored beat to beat,  and saliva was sampled during each trial
for cortisol determination.  For each measure (SBP, DBP, HR and
Cortisol),  large trial main effects were found (p<0.0001).  Group by
trial interaction effects (SBP, DBP and HR p<0.0001;  Cortisol,
p<0.001), suggested that cardiovascular reactivity was independent of
baseline levels, whilst cortisol reactivity was baseline dependent.
Chi2 analyses showed no association between being a cardiovascular
reactor and a cortisol reactor.  The impact of coincidental
cardiovascular and cortisol reactivity was analyzed by grouping on
reactivity, the groups being: concurrent reactors; cardiovascular
reactors; cortisol reactors; and non-reactors.  Significant interaction
effects (SBP & DBP,  p<0.0001), were associated with the largest and
most sustained BP responses being in the concurrent reactor group.
Results are discussed in terms of underlying mechanisms, and the
significance for reactivity in hypertensive pathophysiology.

How the cookie crumbles: A lab study study of affective distress,
dietary restraint, and cardiovascular activation
Thomas Rutledge and Wolfgang Linden
The University of British Columbia

	We adopted an affective distress-induction procedure from the
cardiovascular psychophysiology paradigm to test the relationship
between stress and eating behavior in a female college population.
Participants were assigned to one of three conditions, an
experimental-inhibition (EI), experimental-expression (EE), or control
group (C), in order to manipulate the opportunity to express negative
affect, and to provide a means of assessing the impact of differential
stress recovery upon eating behavior.  All groups responded to a 12-min
lab challenge that involved three active coping tasks and groups EI and
EE were additionally repeatedly interrupted in an attempt to augment
distress.   Expression of distress during the 12-min recovery period
was either encouraged (EE) or discouraged (EI).  Measures of dietary
restraint and positive and negative affect were included to complement
blood pressure and heart rate indices.
	The analysis of food consumption data showed that physiological
	stress was the strongest predictor of post-stress eating
behavior showing less eating with greater concomitant arousal.
Self-reported affect failed to correlate with either food consumption
or physiological stress.  Dietary restraint also proved to be a
reliable predictor of food consumption, even after controlling for
physiological stress.  Finally, an examination of physiological
recovery data revealed that those subjects who were permitted to
express their affect showed a more rapid heart rate recovery relative
to those prevented from diffusing their emotions.  Variations in eating
behavior were not associated with differential recovery however.

Hemisphere differences and aware/ unaware processes in classical
Sara Saban, Kjell Morten Stormark, Dag Hammerborg, and Kenneth Hugdahl
University of Bergen

The present experiment investigated effects of brain asymmetry in
classical conditioning, with aware/ unaware modes of extinction.
Subjects were conditioned to one of two consonant-vowel syllables as
the CS+, which was paired with an aversive noise. The other syllable
(CS-) was never paired with the unconditioned noise.  The syllables
were presented binaurally during acquisition. The conditioned
acquisition phase thus followed a standard conditioned discrimination
procedure. During the extinction, the two CS syllables were presented
in one ear only, either left or right, with one of four new syllables
simultaneously presented in the other ear on each trial. Thus, the
extinction phase used a dichotic listening procedure, which allows for
probing one cerebral hemisphere at a time with the CS+ and the CS-. In
addition, half the subjects were attending the right ear stimuli (left
hemisphere), while the other half were attending the left ear stimuli
(right hemisphere). All physical parameters of the stimuli were the
same for both groups during all phases of the experiment, except for
the left versus right ear attention instructions during the extinction
phase.  Skin conductance responses (SCRs) were recorded, and the
results showed that conditioning occured in both =93groups=94 during
acquisition, indicated by a significant difference between responses to
the CS+ and CS-.  During the extinction phase, both groups showed
evidence of conditioning in the attended condition. However, there was
a significant laterality effect for the unattended condition, with
larger responses to the CS+ compared to the CS-, only for the left
hemisphere group.

Relations between foreperiod muscle tension and anticipatory heart
Dean Sabatinelli, Glen Griffin, and W. Keith Berg
University of Florida

In a forewarned reaction-time study, Haagh and Brunia
(Electroencephalography and  clinical Neurophysiology, 1985, 61, 30-39)
separated subjects into two groups according  to the presence or
absence of spontaneous soleus muscle tension during the foreperiod.
Group differences in measures of cortical anticipation were reported.
To examine the  generality of these findings, we similarly divided
subjects based on forearm EMG tension  during the foreperiod and
examined how this relates to a cardiovascular measure of
This relationship was examined in two separate choice reaction-
time studies.  Subjects were cued 6 s in advance by a visual warning
stimulus, and quickly responded to the visual Go stimulus by
compressing a device with their entire hand. In the first  study, EMG
data beginning 2s prior to the Go stimulus onset was utilized to
separate  subjects into two groups according to each subject's level of
pre-response EMG tension. The tense group showed 72% larger
anticipatory HR decelerations than the relaxed group. In the second
study, data collected over the entire foreperiod replicated the HR
findings of the first study, and showed that the EMG tension increases
began immediately after the onset of the warning stimulus. Subjects
were separable into increasing-tension and relaxed groups, with the
tense group showing 50% larger HR anticipatory  decelerations.
These data, together with those of Haagh and Brunia (1985),
suggest that  information on muscle tension available soon after
warning stimulus onset can predict changes in anticipatory components
of electrocortical and cardiovascular responding several seconds

Wait and see: Aversion and activation in anticipation and perception.
Dean Sabatinelli, Margaret M. Bradley, Bruce N. Cuthbert, and Peter J. 
University of Florida

     Anticipation and perception of unpleasant events have been found
to be associated with potentiated startle responses. This study
investigates the separate roles of aversiveness and activation in
mediating these effects.  Male and female subjects high in
self-reported snake fear first anticipated, and then perceived,
pictures of snakes. To examine the role of arousal, erotic and neutral
(household objects) pictures were also presented.  During a 6 s
anticipation period, a light cue signalled the category of an upcoming
picture, which was then presented for 6 s.  Startle blinks were
elicited by acoustic  probes in either the anticipation or perception
periods on each trial. If aversiveness mediates potentiated startle in
both anticipation and perception, subjects high in snake fear were
expected to respond with greater startle magnitude when processing
snake pictures, compared to low fear subjects.
     Results indicated that snake-fear subjects showed significantly
larger startle responses when viewing pictures of snakes compared to
low fear subjects, whereas the groups did not differ in their startle
responses when viewing erotic (or neutral) materials.  During
anticipation of snake pictures, however, group differences were absent.
Across all subjects, startle blinks during anticipation were reliably
larger for both erotic and aversive pictures, compared to neutral
stimuli. These data support the hypothesis that startle modulation in
perception is related to picture aversiveness. On the other hand, in
light-cued anticipation, acoustic startle responses are potentiated by
the arousing quality of the anticipated stimulus.

Startle responding to alcohol cues among alcoholics
Michael E. Saladin, David J. Drobes, Julian M. Libet, and Raymond F. 
Medical University of South Carolina

	Previous research has convincingly demonstrated that the eyeblink
component of the startle reflex elicited by a sudden, intense stimulus
is positively related to the aversiveness of foreground stimuli.  Based
on these findings, it might be expected that appetitive alcohol cues
should inhibit startle responses, particularly among alcoholics.  This
study investigated the effects of alcohol cues on startle responding
among alcoholics.
	Fifty alcoholics were exposed to two each of alcohol, water, and
non-cue trials, each lasting for three minutes.  Alcohol trials
consisted of exposure to the sight and smell of the subject's preferred
alcoholic beverage.  The order of alcohol and water trials was
counterblanced across subjects, and all of these trials followed the
non-cue trials.  During one trial of each type, six acoustic startle
probes were presented (50 ms, 115 dB) at random intervals.  The
resulting eyeblink was measured along the orbicularis oculi region
under the left eye.  The other trial of each type involved salivary
measurement (a traditional measure of alcohol cue reactivity), as well
as self-reported craving and affect.
Results indicated enhanced startle responding during the alcohol
trials, relative to water and non-cue conditions.  Subject ratings
suggested that alcohol trials were also associated with enhanced
alcohol cravings, and that this experience was highly aversive.  These
findings demonstrate that exposure to appetitive cues may elicit a
negative emotional response under certain conditions.  Potential
explanations for this pattern of response will be discussed, as well as
directions for further research.

Midline P3 amplitude interactions in schizophrenia and mania
Dean F. Salisbury, Iris A. Fischer, Martha E. Shenton, Andrea R. 
Sherwood, Paola Mazzoni, and 
Robert W. McCarley
Harvard Medical School/McLean Hospital

P3 was recorded from 29 schizophrenic, 17 manic, and 24 'normal'
control subjects, as they silently counted target tones (1.5 kHz, 97
dB, 15%) interspersed among standard tones (1 kHz, 97 dB) while
presented against 70 dB white noise. P3 amplitude was measured from
target waveforms as the mean voltage from 300 to 400 ms at Fz, Cz, and
Pz. Repeated measures ANOVA revealed significant midline P3 differences
for groups (F(2,67) =15.5, p <.001).  Post-hoc Scheffe tests revealed
that both psychotic groups were significantly smaller than controls at
each site. Although all groups showed the expected posterior
distribution of P3 (main effect of SITE: F(2,134) =49.0, p <.001),
midline distribution was significantly different between groups (GROUP
x SITE interaction: F(4,134) =2.4, p =.05). Subsequent analyses
revealed that the schizophrenic group differed from bipolars and
controls in midline P3 distribution (GROUP x SITE interactions: Sz vs.
BP F(2,88) =4.63, p =.01; Sz vs. CON F(2,102) =2.8, p =.07; BP vs.
CONF(2,78) =0.6, p =.54). P3 distribution in the schizophrenic group
showed relatively increased frontal voltage, with less difference from
controls at Fz than bipolars, but greater difference from controls at
Pz than bipolars. These midline P3 distribution effects suggest subtle
differences in P3 generator pathology in schizophrenia and bipolar
disorder, with relatively diffuse P3 reduction in mania, but more
preferential involvement of posterior P3 generators in schizophrenia.

Accessibility as a predictor of attitude functionality
Kristen Salomon1, John M. Enrst2, and Jim Blascovich3
1State University of New York at Buffalo, 2Ohio State University, 
3University of California, Santa 

Following Blascovich, Ernst, Tomaka, Kelsey, Salomon, and Fazio
(1993) the accessibility of object-evaluation associations as a
predictor of cardiovascular indices of attitude functionality during
demanding decision making was examined.  Forty-eight  participants
rehearsed attitudes towards one of two mutually exclusive sets of
abstract paintings.  One week later, participants performed a
decision-making task consisting of rapid pairwise judgments for either
the same (rehearsed) set or for the unrehearsed (novel) set of
paintings.  Facial EMG measures were recorded continuously during the
attitude rehearsal task as a manipulation check and cardiovascular
reactivity measures were recorded continuously throughout the
decision-making task.  For participants with consistent
object-evaluation associations and presumably greater attitude
accessibility, zygomaticus major EMG activity was related to rehearsed
attitude valence and intensity.  Within the rehearsed (i.e.,
attitude-relevant) painting condition, participants with consistent
object-evaluation associations exhibited more benign patterns of
cardiovascular reactivity (i.e., increased cardiac responses
accompanied by vasodilation; see Blascovich & Tomaka 1996) than
inconsistent participants (i.e., increased cardiac responses
accompanied by vasoconstriction) during the demanding pairwise judgment
task. These data suggest that attitude accessibility is a sufficient
cause of attitude functionality.

The effects of observers on cardiovascular reactivity during motivated
performance situations: Gender, liking, and evaluation
Kristen Salomon1 and Jim Blascovich2
1State University of New York at Buffalo, 2University of California, 
Santa Barbara

Liking and evaluation by a non-performing  observer were examined as
predictors of cardiovascular reactivity during  a motivated performance
situation.  Sixty-three participants (31 males and 32 females)
performed a mental arithmetic task while being observed by a same-sex
confederate.  To manipulate observer liking, participants were led  to
believe that the observer either liked or disliked them.  Observer
evaluation was manipulated by having the observer keep track of their
performance on a sheet of paper (evaluative) or simply sit and
"silently cheer on" the participant (non-evaluative). A three-way
interaction between gender, liking, and evaluation emerged.  Follow-up
tests revealed a significant interaction between liking and evaluation
for women, such that women in the like/non-evaluative condition
exhibited the most benign pattern of cardiovascular response (i.e.,
challenge), while women in the dislike/evaluative condition exhibited
the most harmful pattern (i.e., threat).  This interaction was not
significant for males (p < .12).  Furthermore, the means suggest that
males in the dislike/evaluative condition exhibited the most benign
pattern of cardiovascular reactivity (i.e., challenge).  These results
are noteworthy  in light of the existing research on evaluative social
support, most of which involve only female participants.

The effect of massage and touch on pain in the cold pressor test as a
function of high versus low need for touch.
Tracey L. Sampson, J. Alexander Dale, and Rod Clark
Allegheny College

	The effects of three conditions pulse palpitation, and  no
touch, were investigated during the Cold Pressor Test on 24
undergraduates(selected from 100) who had either a high (n=12)  or low
need (n=12) for touch as defined by a questionnaire designed by the
first author (after Hollender, 1970).  Gender distribution was
approximately the same in each group.  Heart rates were  monitored
throughout the experimental session via a Grass model 7 polygraph and
silver silver chloride electrodes in a standard V2 placement from an
adjacent experimental room. Pain ratings on a numerical  1-10 scale
were obtained at the end of each one minute pressor condition.  All
subjects experienced each condition of shoulder massage, pulse
palpitation and no touch(experimenter present). The order was
randomized differently for each subject and all possible orders were
used.   The mean of pain ratings for the massage condition and the
means of  pain ratings for both the pulse palpitation and  the no touch
condition were significantly different in that pain ratings were lowest
during the massage condition.  Also, a significant difference in heart
rate changes was found among those in

Modification of the electrically-elicited blink reflex by a
vibrotactile stimulus
Anita J. Sarno and Terry D. Blumenthal
Wake Forest University

Previous research has shown that the electrically-elicited eyeblink
reflex in humans can be modified by weak acoustic or visual pulses,
with eyeblink inhibition when the weak pulse precedes the electrical
stimulus, and eyeblink facilitation when the weak pulse follows the
electrical stimulus.  This study (N = 17) evaluated the impact of a
vibrotactile pulse on the electrically-elicited blink reflex, with the
vibration presented to the hand before (positive stimulus onset
asynchrony (SOA)), simultaneous with, or after (negative SOA) the
electrical stimulus. Relative to control blink responding, the
vibrotactile pulse inhibited the late R2 component of the blink reflex,
as shown by decreased response magnitude and probability (at 75 to 150
ms SOAs), decreased response duration (at 50 to 150 ms SOAs), and
increased response latency (at 100 to 150 ms SOAs). The facilitation of
R2 that is found with acoustic and visual pulses at negative SOAs was
not observed for any response measure at any SOA.  The magnitude of the
early R1 blink component was facilitated at 75 and 100 ms SOAs.
Previous research has shown that, when a weak pulse and a blink
eliciting stimulus are in the same sensory modality,  blink inhibition
can be observed but blink facilitation is generally absent. The present
findings suggest that, with respect to the blink reflex, an electrical
pulse to the forehead and a vibration to the hand are in the same
sensory modality. The overlap of the neural circuits subserving these
two stimulus inputs is supported by the facilitation of the early R1
blink component by the vibrotactile pulse.

Menstrual phase and effects on prepulse inhibition of startle
Umut Sarpel, Steven B. Schwarzkopf, and James R. Ison
University of Rochester

Gender differences in prepulse inhibition (PPI) of the acoustic
startle reflex (ASR) have been reported in humans, with females showing
decreased PPI (Swerdlow et al. Biol.Psych., 34:253-260, 1993).
Alterations in neurotransmission due to hormonal fluctuations were
implicated.  To explore these findings we 1) assessed gender
differences in existing data, and 2) are conducting a prospective study
of PPI targeting potential estrogen effects.
	In the ongoing study, data was gathered from 9 healthy females
with regular menstrual cycles.  None had psychiatric or neurological
disorders, hearing loss, or were using hormone-altering medication.
Subjects underwent ASR testing for PPI with prepulse intensities of 3,
6, & 12 dB above background.  Subjects were tested during menstruation
(low estrogen) and the follicular phase (high estrogen).
	Our preliminary gender study demonstrated a trend for lower PPI
in females compared to males.  Most striking was the significantly
greater variability of PPI in females compared to males (Levene test).
The current study indicates a significant reduction in 3 dB PPI during
the follicular phase (repeated-measures ANOVA, one-tailed).
Furthermore, variance was greater during the follicular phase for all
prepulse intensities (Levene).
	These preliminary findings are consistent with reported gender
effects for PPI, and speculation that it is associated with hormonal
fluctuation.  Increased variance and PPI reduction seem linked to the
estrogen peak during the follicular phase.  Estrogen has been proposed
to potentiate dopaminergic effects; furthermore, animal studies
indicate that increased mesolimbic dopamine leads to decreased PPI.
Therefore, we hypothesize that the estrogen surge of the follicular
phase may lead to enhanced dopaminergic tone, which in turn results in
the observed PPI changes.

Developmental and migraine-specific aspects of the
Gudrun Sartory and Bernhard Mueller
University of Wuppertal

Migraine in both adults and children is characterised by an elevated
Contingent Negative Variation (CNV). In order to shed light on the
functional significance of this finding,  the Bereitschaftspotential
(BP), thought to be partly responsible for the late CNV wave, was
investigated in children suffering from migraine (N=30), in healthy
age-matched controls (N=16), and in a group of healthy adults (N=20).
The children had a mean age of 11.7 years (SD=2.0) and the adults of 26
years (SD=4.3).
Subjects pressed a button 80 times with the right digit finger
while the motor-related potential was registered from 12
electrode sites (10-20 system) referenced to linked ears. EMG from the
right forearm and vertical EOG from the right eye were also recorded.
EEG was recorded with a time constant of 10 sec and a low band pass
filter of 35 Hz and sampled with 200 Hz from 1700 msec before to 200
msec after the button press.  Data were corrected for EOG artefacts,
averaged and base-line corrected with regard to the initial 200 msec.
Comparison of groups yielded differences in terms of topography
and amplitude. Adults showed more distinct lateralization with a more
pronounced Bereitschaftspotential at C3 than children whose BP was more
prominent in the mid-fronto-central region (Cz, Fz). Compared  to the
other two groups, migraineurs also showed an elevated BP at these
electrode positions. Healthy control children exhibited a positive
deflection 500 msec before the button press at Cz and C3 and up to the
button press at F3 and F4 which was not evident in the other two
groups.  The Bereitschaftspotential of healthy children thus appears to
be characterised by inhibitory potentials in the frontal area whereas
that of migraine children is more similar to that of adults in shape
while, however, being less lateralized and more elevated at mid-line

P300 and the scales of the Minnesota Multiphasic Personality Inventory
(MMPI) in normal subjects
Toshinori Sasaki1, Hiroki Yabe1, Fumio Saitoh2, Yasuharu Satoh1, Yutaka 
Fukushima3, and Sunao 
1Hirosaki University School of Medicine, 2Aomori Prefectural 
Tsukushigaoka Hospital, 3Aomori 
Central Hospital

Ogura et al (1994, Int J Pscychophysiol, 16 shown that students who
score high on the Minnesota Multiphasic Personality Inventory (MMPI)
schizophrenia scale (reflecting hebephrenia-like symptoms) have a
smaller P300 component of the event-related brain potential (ERP) than
students who score low.  This extreme-groups design is similar to
comparing schizophrenics to normal controls.  We asked whether P300
amplitude would reflect schizophrenia-like symptoms in subjects who
score in the normal range on the schizophrenia scale.
	We recorded ERPs from 18 students who tested within the normal
range on all scales of the MMPI and who had no family history of
psychiatric disorders.  No subject scored higher than 70 on the
schizophrenia scale.  ERPs were recorded during an auditory oddball
paradigm in which a 2000 Hz tone occurred on 20% of the trials and a
1000 Hz tone occurred on 80%.  Subjects pressed a response button to
the 2000 Hz tone.  Trials with eye blink artifacts were excluded.
Three subjects were excluded for poor task performance.  P300 amplitude
measured at Pz was regressed against MMPI scores.  Subjects with higher
scores on the schizophrenia scale had smaller P300s (r=-.58, p<.05).
P300 amplitude did not correlate with scores on any other scale.
	These data suggest that even within normal subjects, subjects
who have more schizophrenia-like symptoms have smaller P300s.

Glucose hypometabolism in the left anterior caudate nucleus of
depressives reverses following antidepressant therapy

Stacey M. Schaefer, Heather C. Abercrombie, Christine L. Larson, R. 
Terry Ward, Patrick A. Turski, 
Dean D. Krahn, Scott B. Perlman, James E. Holden, and Richard J. 
University of Wisconsin-Madison

Research implicates the basal ganglia in emotional dysfunction and
mood disorders such as depression.  Functional neuroanatomical studies
utilizing Positron Emission Tomography (PET) have reported abnormal
blood flow and metabolic rate in the caudate nucleus of depressives.
In order to better understand these observed differences, we examined
the regional cerebral metabolic rate for glucose (rCMRglu) in the
anterior caudate nucleus in 9 right-handed, unipolar depressed subjects
(6F) with 18F-fluorodeoxyglucose PET.  Subjects were tested before and
after an average of 17.25 weeks (SD=5.39 weeks) of successful
antidepressant treatment.   We identified the left and right anterior
caudate nucleus on each subject's Magnetic Resonance Image (MRI), which
was co-registered to his or her PET scan. rCMRglu was then extracted
from these regions on the co-registered PET scan.  Comparisons were
made with 10 normal controls (6F) tested at similar time intervals.
The control group consistently demonstrated a metabolic asymmetry
in the anterior caudate with greater rCMRglu in the left than the
right caudate on both testing occasions.  The depressives showed
relative hypometabolism in the left compared to the right anterior
caudate at the initial testing, which reversed following treatment to
significantly greater rCMRglu in the left.  Thus, the depressed group
did not demonstrate the pattern of asymmetry seen in controls until
after antidepressant therapy. The implications of these data for the
pathophysiology of depression and for differentiating between
state-dependent and independent components of the circuitry involved in
depression are discussed.

Autonomic indexing of long-term memory deficits in adult children of
Steven L. Schandler, Whitney V. Leach, Connie S. Thomas, Heather Platt, 
and Michael J. Cohen
Chapman University Veterans Affairs Medical Center, Long Beach

Visuospatial information processing appears as a reliable probe of
cognitive differences in adult children of alcoholics (ACA) and adult
children of nonalcoholics (NACA).  Compared to NACAs, the visuospatial
processing of ACAs is significantly disrupted.  The present study was
designed to isolate the location of disruption within the ACA's
information processing cycle.
Forty-four matched male subjects served as subjects.  Based on
interviews and questionnaires, 22 subjects were ACA from families in
which the father had a history of alcoholism. The remaining 22 subjects
were NACA.  During a single experimental session, each subject first
learned the spatial positions of eight low semantic content "nonsense
shapes" on an eight-position grid.  Immediately following learning, an
overlearning phase was initiated.  Finally, subjects received a high
interference task requiring them to learn the same shapes presented in
new spatial positions.
Learning performance, skin conductance and heart rate measures
were obtained.  In agreement with many previous studies, ACAs performed
significantly more poorly than NACAs on this form of visuospatial
learning task.  However, while NACAs had difficulty learning the
interference task, ACAs displayed little or no interference.  Using an
activation peaking template, NACAs showed significantly more activation
during the interference task, indicating the utilization of increased
resources to resolve the interference.  ACAs showed virtually
equivalent activation patterns on the learning and interference tasks.
From the data it appears that the ACA's visuospatial
processing disruptions reflect less complete and permanent information
storage.  Without stored, conflicting information, ACAs are less or
unaffected by subsequent interfering information.

Do you know what you did? Errors, confidence and the brain
Marten K. Scheffers and Michael G. H. Coles
University of Illinois at Urbana-Champaign

There is a component of the event-related brain potential, the
error-related negativity (ERN), that is associated with error-
detection in choice reaction-time tasks.  In the present study, we
examined the relationships among the amplitude of the component,
subjects' ratings of the accuracy of their responses, and patterns of
response activity.  Two groups of nine subjects performed a visual
choice reaction time task under either intact or degraded conditions.
Stimuli were the letters H and S, surrounded by compatible or
incompatible noise letters.  Subjects were required to respond with
left- or right-hand "squeezes" of zero-displacement dynamometers.
After each response, subjects had to judge the accuracy of that
response using a five-point scale, ranging from sure correct, unsure
correct, don't know, unsure incorrect, to sure incorrect.
Event-related brain potentials, electromyographic activity in the
responding arms, and response force were recorded on each trial.  In
the intact condition, subjects made few errors, were accurate in
judging their responses, tended to use only the extreme ends of the
confidence scale, and showed large ERNs when they erred.  In the
degraded condition, when subjects were correct, the ERN was largest for
"don't know" responses, and, when they were incorrect, the ERN was
largest for "unsure incorrect" responses.  Subjects' judgments about
their response accuracy depended on the relative levels of force
associated with correct and incorrect responses and on the relative
timing of correct and incorrect response activity.

An event related brain potential examination of using masked visual
cues to manipulate early attentional processing in elderly subjects
David M. Schnyer, John J. Allen, Alfred W. Kaszniak, and Chad J. 
University of Arizona

To examine preconscious attentional processing in elderly subjects,
65-85 year-olds made decisions about whether a target was above or
below a line.  A pattern indicating a vertical column of positions with
a center fixation line appeared; after disappearing a target appeared
in one of the locations.  On 66% of the trials, a prime was presented
for 60ms in one of these locations and then covered by the pattern;  in
half of the priming trials the prime marked the position in which the
target would subsequently appear and in half  it marked a position
opposite (on the other side of the fixation line) from the target.
Subjects reported no awareness of this priming manipulation. Reaction
time data revealed two groups of subjects, one where a consistent
prime/target correspondence produced facilitation and one where it
produced interference. ERPs revealed an early effect, such that, when
the target was preceded by the inconsistent prime, a distinct negative
peak at 60 ms appeared which was absent from the consistent priming and
no priming conditions. This peak was focused along the posterior
midline.  While this effect is earlier than those reported by other
studies of spatial attention, this study involved an exogenous cue and
the pattern of results is consistent with that revealed by conscious
manipulations of attention.  The two groups defined by facilitation or
interference reaction time effects, however, did not differ in ERP
effects.  The results indicate that elderly subjects are sensitive to
preconscious attentional manipulations in perceptual tasks.

A new auditory distraction task: Electrophysiological and behavioral
effects of task-irrelevant sound change
Erich Schroeger
University of Munich

Event-related potential (ERP) and behavioral effects of a new auditory
distraction task were determined. Subjects received tones that could be
of short (100 ms) or long (200 ms) duration equiprobably. They were
instructed to press a left button to long-duration tones (targets).
Tones could be of standard frequency or of low-probability (p=.1),
deviant frequency. The frequency difference between standard and
deviant was small (850 Hz), medium (200 Hz), or large (500 Hz). In
addition to these oddball blocks, control blocks were administered.
There, tones were of 10 different frequencies (involving those
frequencies which served as standard and deviant in the oddball
blocks). Each frequency was presented equiprobably, that is, there were
no standards or deviants. It was found that the frequency deviants
elicited MMN, N2b, and P3a components. Furthermore, these
task-irrelevant frequency deviants caused impoverished behavioral
performance to targets (increase in reaction time, increase in false
alarm rate, decrease in hit rate). Even with the small frequency
deviant distinct distraction effects were obtained: reaction times to
targets revealing a small frequency-change were 51 ms longer than those
to targets being of standard frequency. A comparison of the results
obtained in the oddball blocks with those obtained in the control
blocks showed that the electrophyiological and the behavioral
distraction effects were really due to impaired target processing
caused by deviants. It is argued that this new distraction task is
suited to yield distinct distraction effects with rather "small"
distractors. Pros and cons of this paradigm are discussed.

ERPs and blinks: Sex differences in response to erotic and violent
picture content
Harald T. Schupp, Bruce N. Cuthbert, Charles Hillman,Roy Raymann, 
Margaret M. Bradley, and Peter J. Lang
University of Florida

The research examined startle probe blinks, probe cortical ERPs, and
picture onset ERPs in 52 subjects (26 female) viewing 60 pictures (6 s.
exposure) from the International Affective Picture System. Stimuli
included 15 pleasant, sexually oriented pictures (erotic couples,
romantic courtship, attractive opposite sex nudes) and 10 affectively
positive control stimuli (sports scenes); 10 affectively neutral
pictures (non-expressive human faces, household objects); 15 unpleasant
pictures of aggression and violence (human violent acts, dead and/or
mutilated victims of violence, animals attacking) and 10 unpleasant
control stimuli (scenes of grief, disgusting objects).
	Both men and women showed similarly strong blink inhibition and
probe P3 reduction while viewing sexually evocative pictures. The sexes
differed, however, in their blink response to pictures of romantic
courtship (greater relative inhibition for women) and opposite sex
nudes (greater inhibition for men). Both sexes also showed similarly
strong reflex potentiation to violent content, particularly the scenes
of human aggression and animal attack. However, only pictures of human
aggression resulted in P3 reduction.
	For both sexes, onset ERP late positivity was largest to
	sexually evocative pictures. The sexes again differed, however,
in response to opposite sex nudes (less relative late positivity for
women). Pictures of human mutilation occasioned marked late positivity
-- more than for other aversive contents. Overall, probe blink and ERP
responses covaried significantly with subjects' valence and arousal
judgments, respectively; however, particularly in separate analyses by
sex, content specific responses were also clearly apparent.

Isolation rearing in rats: Long term neurochemical and behavioral
Steven B. Schwarzkopf1, Vaishali P. Bakshi2, David L. Braff2, Mark A. 
Geyer, and Kirsten M. 
1University of Rochester; 2University of California in San Diego

	 Long term behavioral changes occur when rodents are isolation
reared (IR) vs. socially reared (SR). We have found that prepulse 
inhibition (PPI) of the acoustic 
startle reflex (ASR) is reduced in IR animals.
Abnormalities of PPI are present in a number of psychiatric syndromes,
making this model potentially relevant to clinical phenomena. In this
study we examined behavioral as well as regional brain
catecholamine/indoleamine levels in IR animals.
	Twenty eight male Lister rats were used in this study.  After
weaning, animals underwent IR rearing (1 per cage, N=14) or SR
(2-4 per cage, N=14).  Twelve weeks later, animals were tested for PPI
and motility, and  sacrificed one week later.  Brains were rapidly
removed, frozen at -70C, and later dissected for isolation of frontal
cortex (FCtx), sensory/motor cortex (SMCtx), dorso-lateral striatum
(DStr), ventro-medial striatum (VStr), and hippocampus (Hipp). Each
region was assayed (HPLC) for dopamine (DA), serotonin (HT), and
norepinephrine (NE), and metabolites.
	Compared to SR rats, IR rats showed reduced PPI and exhibited
reduced spontaneous exploratory activity in a novel
environment. IR animals showed reduced DA (and metabolites) in VStr and
Hipp.  HT and its metabolite were increased subcortically (striatum)
but reduced in the Hipp.  NE was reduced in the Hipp and cortical
regions, with increases in the DStr.
	Results further support long term behavioral effects of IR in
rats and regional catecholamine and indoleamine alterations.  A
state of excess stress reactivity accompanied by depletion of some
neurochemical substrates and compensatory increases in others is

Associations between prepulse inhibition and facilitation of acoustic
startle, P50 ERP gating, and intensity dependence measures
Steven B. Schwarzkopf1, Gregory A. Light2, Steven M. Silverstein1, and 
Kirsten M. VanMeenen1
1University of Rochester; 2University of California in San Diego

Acoustic startle reflex (ASR) and cortical evoked response potential
(ERP) paradigms have been hypothesized to assess central inhibition
and/or facilitation.  These measures have also been speculated to have
relevance to psychopathologic states, with basic findings beginning to
clarify central mechanisms.  Relationships between these measures
remain unclear.   Prepulse inhibition (PPI), prepulse facilitation
(PPF), and habituation of ASR, P50 ERP suppression (P5SUP), and an ERP
measure of "intensity dependence" (ID) were examined in this study to
test for significant associations.
	28 normal subjects were tested on all three paradigms (ASR,
P50, ID).  Twenty subjects exhibited reliable ASR measures,
with 27 and 28 having reliable P50 and ID measures, respectively.
Despite gender differences in raw measures of the ASR, no gender
effects were noted for relative measures of inhibition, facilitation,
or ID.  As we have previously reported, PPI and P5SUP were not
positively associated, with a trend for P5SUP to be positively
correlated with ASR habituation.  No strong associations were present
between P5SUP and ID measures.  The strongest association (positive)
was between PPF and ID, both measures of central facilitation.
	Despite being equated at times in the literature, findings do
not support a clear association between PPI and P50
suppression, and provide continued, though modest, support for a
relationship between ASR habituation and P50 suppression.  If
replicated, results suggest that a form of central facilitation may be
assessed using the ASR with appropriate stimuli and ERP paradigms for
measuring ID. Findings will be discussed in terms of central mediation
of these measures.

Stimulus discriminability and dual task affect the N2 and P3
separately: An auditory oddball additive factors ERP study.
Sidney J. Segalowitz and Glenn Theal
Brock University

The N2 and P3 in the auditory oddball paradigm are hypothesized to be
associated with stimulus discriminability and stimulus evaluation.
Discriminability has been shown to affect the latency of the N2
component whether task-relevant or not.  Increased difficulty in
evaluation increases P3 latency.  One could draw from this that (1) the
N2 sensitivity to discriminability should be relatively automatic, i.e.
little affected by reducing attentional allocation to the stimulus; and
(2) the latency of the P3 after the N2 has occurred should not be
affected by stimulus discriminability but should be affected by a
secondary vigilance task that demands attentional resources.  We tested
this in an additive factors paradigm with 3 levels of stimulus
discriminability (easiest to hardest: discriminations between 800 vs
1500 Hz pitch, 100 vs 150 msec tone durations (at 400 Hz pitch), and
100 vs 125 msec durations) and 2 levels of Distractor task (with or
without a continuous verbal working memory task).  RT and error scores
demonstrated that the tasks behaved as expected (harder
discriminability and distraction significantly increased RT and error
scores).  N2 latency was increased by stimulus discriminability only (p
< .001) and P3 latency was affected by discriminability and distraction
(p < .01).  There was no interaction in either analysis, indicating
that automatic stimulus discriminability and the evaluation sensitive
to attentional resources are affected independently.

Comparing the habituation of electrodermal and electromyographic
responses to startling stimuli
Kimberle A. Seljos1, Anne M. Schell2, Michael E. Dawson1, and Veronica 
Y. Mejia1
1University of Southern California, 2Occidental College

Startle eyeblink magnitude, the skin conductance response (SCR), and
skin conductance level (SCL) were recorded while 87 participants
listened to 8 startling 104 dB white noises. Response magnitude and
percent (of first trial) magnitude scores for blink and SCR were
regressed against trial number to obtain regression coefficients
representing rate of change, or habituation, across trials (with
negative regression coefficients indicating habituation). Correlation
analysis revealed that for both blink magnitude and SCR, habituation
rates were significantly greater among participants with larger overall
response magnitudes (r = -.27 and r = -.53, respectively). For both
blink and SCR percent magnitude scores, however, participants with
larger overall response magnitudes had significantly lower habituation
rates ( r = .23 and .24); larger responders tended to maintain their
responsivity as a percentage of initial level to a greater extent.
Slower percent magnitude habituation rates for blink and SCR were also
associated with higher SCL, a measure of arousal (r = .24 and .23). Of
particular interest, the regression coefficients for percent blink
magnitude and percent SCR magnitude were positively correlated (r =
.40). Participants who habituated rapidly (in terms of percent change)
in one response system habituated rapidly in the other. Thus, when
stable physiological variables that affect overall response magnitude
(such as active sweat gland counts or characteristics of facial
musculature) are normalized by the use of percent change measures,
these two response systems respond in a similar manner to repetitive
stimulation by startling stimuli.

ERP measures of source and item memory in young and elderly subjects
Ava J. Senkfor and Cyma Van Petten
University of Arizona

Prior research has suggested that memory for items is dissociable from
memory for the context in which they were learned ("source" memory).
We recorded event-related potentials (ERPs) as healthy young and
elderly subjects made item and source recognition judgements on spoken
words.  In the item-study phase of the experiment, a list of concrete
nouns was  spoken in a male or female voice with a size judgement task
(large/small object).  In the subsequent item recognition test,
subjects decided only if words were old or new; half of the old words
were presented in the opposite voice as at study.  Correctly recognized
old words elicited more positive ERPs than new words, regardless of the
voice change manipulation, maximal from 400-800 msec poststimulus.  The
source-study phase was similar to item-study, but at recognition
subjects responded "new", "old/same voice", or "old/different voice".
An old/new difference similar to that of the item recognition test was
observed.  In addition however, correctly recognized old words elicited
a large late positivity than new words at prefrontal electrode sites,
regardless of the success or failure of judging the voice as same or
different. We interpret the prefrontal positivity as a reflection of
the attempt to recover voice information, while more posterior sites
reflect the success or failure of memory retrieval. All observed
effects were reduced in elderly as compared to young subjects.
However, the results also suggest that aging can dissociate different
ERP memory effects.  The amplitude of the old/new word recognition
effect and the prefrontal positivity were significantly correlated in
young subjects, but not in elderly subjects.

Cardiovascular reactivity and psychological stress-induced myocardial
ischemia in cardiac patients
David Sheffield, Ph.D., Brian M. Go, M.D., Paula L. Biles, Claudia G. 
Christy, R.N., Kathleen C. 
Light, Ph.D., and David S. Sheps, M.D.
University of North Carolina at Chapel Hill

In order to examine the relationship between cardiovascular
reactivity and psychological stress-induced myocardial ischemia, 47
patients with a history of coronary heart disease and documented
myocardial ischemia were studied by radionuclide ventriculography, at
rest and during a speech task.  Fourteen (30 %) patients developed
ischemia, defined as a reduction in ejection fraction >5 per cent or
new wall motion abnormality.  Patients who developed ischemia had
significantly larger diastolic blood pressure responses to the speech
task than non-ischemic patients (means (SEs) = 31.2 (4.0) mmHg and 19.8
(3.1) mmHg, respectively, p <.05) and had a tendency to have larger
systolic blood pressure reactivity (means (SEs) = 51.3 (6.6) mmHg and
39.7 (4.3) mmHg, respectively; p =.14).  There were no differences in
heart rate reactivity, or resting blood pressure and heart rate between
the two groups.  Further, differences in blood pressure reactivity
between ischemic and non-ischemic patients remained substantially
unaltered after adjusting for resting blood pressure.  There were no
differences in age, sex, and medication status between the two groups.
In addition, resting left ventricular ejection fraction did not differ
between the ischemic and non-ischemic patients.  Furthermore, coronary
angiography performed on 27 patients revealed no statistically
significant difference in the number of diseased vessels between
patients who developed ischemia and those who did not.  These data
suggest that blood pressure reactivity is related to psychological
stress-induced myocardial ischemia.

Effects of family history of hypertension and urbanization on blood
pressure in Zimbabwean medical students
Jeffrey J. Sherman1, James A. McCubbin1, and Jonathan Matenga2
1University of Kentucky College of Medicine, 2University of Zimbabwe

The increasing prevalence of essential hypertension is a growing
public health concern for Zimbabwe and other African countries. Two
important risk factors for hypertension are urbanization of the African
population and family history of hypertension. Positive family history
is associated with increased resting blood pressures and increased
reactivity to stress. However, the effects of family history and
urbanization on blood pressure is poorly understood. This study
explored the relationship between family history of hypertension and
urbanization on resting blood pressures and blood pressure responses to
a mental arithmetic stressor in a group of normotensive, black medical
students with (n=36) and without (n=34) a parental history of high
blood pressure. While all participants resided in Harare, an urban
setting, 19 had parents currently residing in a rural area. Blood
pressures were measured at rest and during a ten minute stress period.
Results indicate that those with a positive parental history of high
blood pressure had higher resting blood pressures, and greater systolic
blood pressure increases in response to stress, when compared to their
negative family history counterparts. Further, those with parents
residing in urban areas had higher resting systolic blood pressure than
those with parents residing in rural areas.  However, no reactivity
differences were apparent between the urban and rural groups. These
data suggest that while family history for hypertension influences both
resting blood pressures and reactivity, urbanization may influence only
resting blood pressures.  (Supported by NIH/Fogarty International
Center 5T37TW3041-02, NIH HL35195, and NIH HL32738)

Avoiding noise artifacts in airpuff startle stimuli
Brendan M. Shortley and W. Keith Berg
University of Florida

Airpuff stimuli typically used in studies of selective attention
modulation of startle nearly always  produces an accompanying sound, a
potentially serious confound.  Can subjects use this noise alone as the
discriminate stimulus in a study of selective attention and is there a
way to avoid this?
In three studies designed to examine this question subjects
received a 2 s warning stimulus followed randomly by either an airpuff
of 25 or 100 ms or no airpuff.  Subjects were instructed to
discriminate the duration of those airpuffs they judged to be present.
In the first two studies  (Ns=20 and 14, respectively) the airpuff
orifice was near the eye but directed away from the face so only
acoustic stimulation could occur.  In study 2 a 86-98 dBA masking sound
of 20 - 600 Hz was presented on circumaural earphones which attenuated
outside noise.  In study 3 (N=18) airpuff  was directed near the eye to
elicit tactile startle blink, and the airpuff noise abatement procedure
used in study 2 was again employed.  Judgments of the presence of
airpuff and its duration showed  that subjects easily detected and
discriminated an unaltered puff noise (study 1 - 88.6% accuracy),  but
with puff noise abated (study 2) responding did not differ from chance
(41.1% accuracy).   Judgments of duration of the tactile startle with
puff noise abated (study 3), were well above  chance (78.6%).
The results indicate that airpuff induced noise can be a serious
artifact in studies of selective attention, but can be eliminated.

Epoch length and accuracy of power spectral analysis
R.P. Sloan1, A. Choudhri1, M.M. Myers1, P.G. Grieve2, and H. Levin1
1Columbia University, 2Northrop Grumman Corporation

Fourier analysis of biological oscillations has yielded important
information about autonomic control of the heart and vasculature.
Nevertheless, there are virtually no standards for the relationship
between the length in time of an epoch of data to be analyzed and the
frequencies to be examined.  We addressed this matter by creating 8
series of 500 RR intervals, differing in constituent frequencies and
power, which were submitted to Fourier analysis of 6 different epoch
lengths:  30, 60, 90, 120, 150, and 300 sec.  LF (0.02-0.06 Hz), MF
(0.07-0.14 Hz), and HF (0.15-0.50 Hz) power was computed using standard
methods.  For each analysis, we computed percent error in each
frequency band.  As expected, analyses of 30 and 60 second epochs were
associated with greater methods.  For each analysis, we computed
percent error in each frequency band.  As expected, analyses of 30 and
60 second epochs were associated with greater than 30% and 10% error
respectively for LF and MF power.  Surprisingly, however,the percent
error from analyses of 90 second epochs for LF and MF power was not
substantially greater than that from analyses of 300 second epochs.  As
expected, for HF power, analyses of all epochs were associated with low
error.  These findings suggest that while longer epochs generally are
associated with reduced error, epochs as short as 90 sec provide
reasonably accurate estimates of LF and MF power and that for HF power,
epochs as short as 30 sec are adequate.

Movement related EEG potentials of isometric force production during
speed and accuracy tasks
Semyon Slobounov1 and William J. Ray2
Penn State University

We examine whether different speed and accuracy requirements in
discrete and repetitive finger movements influence the spatio-temporal
characteristics of the readiness potentials (RP) preceding these
movements.  Twelve right-handed healthy male and female subjects
produced a movement with either the dominant or non-dominant hand by
pressing the thumb and index finger against the load cell with
differential task requirements Both discrete and repetitive finger
force production movements were examined under conditions of speed and
accuracy. A NeuroScan system using an Electro-Cap Electrode helmet was
used to record continuous EEG at site Fp1, Fp2, Fz, Cz, C3, C4, Pz, P3
& P4 referred to linked mastoids. A multivariate 3-factor repeated
measures ANOVA using 2 task requirements (fast and accurate) by 2
movement patterns (discrete and repetitive) by 3 electrode locations
(C3, CZ and C4) was used.  Of particular interest was to determine
whether different speed and accuracy constraints in discrete and
repetitive isometric force influence the spatio- temporal
characteristics of the readiness potentials (RP) preceding these
movements. One finding for the RP amplitude showed enhancement at
higher rates of force development both for speed (r = .810, p < .01)
and accuracy task (r = .887, p < .01) during discrete and repetitive
tasks (r = .907, p < .01; r = .901, p < .01,respectively). Another
finding was that the amplitude of RP was consistently higher for fast
ratherthan accurate movements.  Moreover, the result from analysis of
RP onset times suggests that the RP begins earlier for the fast
movements than for the accurate movement.

The P300 is sensitive to evaluative and nonevaluative categorization
processes of very briefly presented pictorial stimuli
N. Kyle Smith1, Wendi L. Gardner1, Kenneth Hugdahl2, Gary G. Berntson1, 
and John T. Cacioppo1
1Ohio State University, 2University of Bergen

In prior ERP research, we have developed a variation of the oddball
paradigm that is sensitive to differences in evaluative (good/bad,
pleasant/unpleasant) judgmental processes.  Although we have
consistently found a larger late positive potential (presumably a P3b)
to evaluatively inconsistent than consistent stimuli, we have relied on
fairly long stimulus presentations (> 1000 ms) that were presented
foveally.  In pilot research, we developed a set of simple pictorial
stimuli for which subjects could accurately make evaluative
(positive/negative) and nonevaluative (animal/nonanimal) judgments when
the stimuli were presented for 180 ms in right or left hemifields.  We
subsequently used these stimuli to examine whether differences in
evaluative and nonevaluative categorization processes could be indexed
using ERPS when stimuli were presented just above sensory threshold
levels to nonfoveal as well as foveal regions.  Fourteen subjects were
exposed to sequences of five stimuli and performed a dichotomous
evaluative or nonevaluative categorization task. Categorization
inconsistency was varied by embedding either a negative picture in a
sequence of neutral pictures (evaluative task instructions) or a
picture of an animal in a sequence of nonanimal pictures (nonevaluative
task instructions).  These target stimuli were presented foveally or in
the left or right hemifield whereas all remaining pictures were
presented foveally.  Despite the use of stimulus presentations so brief
that the lateralized presentations appeared outside of the spotlight of
attention, results replicated prior research showing a larger P300 over
centroparietal regions to categorically inconsistent than consistent
stimuli.  Implications for investigating the neural substrates of
evaluative processing are discussed.

Modulation of the potentiated startle response: The role of temperament
Nancy Snidman and Jerome Kagan
Harvard University

The EMG startle response to an acoustic probe can be potentiated by
the threat of an aversive stimulus. The role of temperamental
differences in modulating this response was examined in children
categorized by behavioral reactivity during infancy and early
childhood.  One group was selected to be high in behavioral reactivity
at 4 months of age and behaviorally inhibited at 14 months and 4 years
of age, while another group did not show this consistent pattern.  As
part of a laboratory battery at 6 years of age, eyeblink startle
responses were elicited by an acoustic probe under threat and no threat
conditions.  A green light signaled trials during which an aversive
stimulus, an airpuff (60psi), delivered to the throat at the level of
the larynx, might be presented, and no light trials during which an
airpuff would never occur.  On each trial, a 95dB, 50-ms burst of white
noise was presented binaurally through headphones and EMG activity of
the m.orbicularis oculi was measured.  Preliminary data on the first 10
subjects revealed that the high reactive/inhibited children showed
greater blink magnitude to the acoustic probe during the threat
condition than the non-threat condition compared with other children.
The results suggest that subjects categorized as high reactive in
infancy and behaviorally inhibited in subsequent years may have a lower
threshold of arousal in limbic structures that could explain their
behavior in the laboratory and their potentiated startle response to a
threat condition.

Lateralized control of inotropic and chronotropic cardiac reactivity to
Kathleen Soderlund1, Beth Colaluca1, Stefan Wiens2, Sarah Reiff2, Tamera 
Schneider2, and Robert 
M. Kelsey2
1University of North Texas, 2State University of New York at Stony Brook

Sympathetic nervous system control of the heart is apparently
lateralized, with the left cardiac sympathetic nerve (CSN)
predominating over inotropic performance and the right  CSN
predominating over chronotropic performance.  The relationship of CSN
lateralization to central hemispheric lateralization (e.g.,
contralateral, ipsilateral, unilateral, bilateral) is less clear.  This
relationship was explored by requiring subjects to tap their left index
finger (right hemisphere activation) or right index finger (left
hemisphere activation) throughout a 1-min vocal mental arithmetic
task.  Preejection period (PEP), an inotropic index, and heart period
(HP), a chronotropic index, were recorded from 50 undergraduate women
who were right-handed according to the Edinburgh handedness
questionnaire.  All subjects completed an initial 10-min baseline
period, a 5-min mental arithmetic training task, a second 5-min
baseline period, and a 1-min mental arithmetic test task.  For the test
task, subjects were randomly assigned to either a left tap group
(n=17), a right tap group (n=16), or a no tap control group (n=17).
Left and right tap groups did not differ significantly in tapping
frequency (p>.75).  A multivariate analysis of PEP and HP reactivity
(changes from pretask baseline) during the test task, using initial PEP
and HP reactivity to the training task as covariates, revealed a
significant effect of tapping condition (p<.01).  PEP reactivity (left
CSN) was greatest in the right tap (left hemisphere) group, whereas HP
reactivity (right CSN) was greatest in the left tap (right hemisphere)
group.  These preliminary results suggest that central-sympathetic
neural control of the heart is primarily ipsilateral.

Stability of cardiovascular responses to forehead cold pressor
John J. Sollers III, Glenn S. Brassington, Bruce H. Friedman, Julian F. 
Thayer, and Lynn A. Rossy.
University of Missouri, Columbia and Washington University, St. Louis

Temporal and situational stability of cardiovascular responses to str
ess is assumed when implicating individual differences in
cardiovascular reactiv ity as a risk factor for the development of CVD.
However, the effects of forehea d cold pressor stimulation (FCPS) on
the cardiovascular system in a standing pos ture remains unexplored.
Hence, the purpose of this study was to test the tempor al and postural
stability of cardiovascular responses to FCPS across the sitting and
standing posture.
Participants were 11 female (aged 21-39, M=25.4, SD=5.4) and 11
male (aged 23-38, M=28.5, SD=4.6) graduate students in psychology.
Participants' heart rate (HR) systolic blood pressure (SBP) and
diastolic blood pressure (DBP) were measured each minute for 5 minutes
at baseline in the sitting and standing postures. Then participants'
HR, SBP and DBP were measured each minute in the sitting and standing
postures with exposure to FCPS (3min.). The temperature was maintained
at 5-10 degrees Celsius. Each participant was exposed to FCPS twice in
the sitting and standing posture and the order of testing was
counterbalanced.  Results indicated good test-retest reliability for
all cardiovascular measures w ith correlations being slightly larger in
the sitting as compared to the standin g posture for all measures
(sitting: .84< r > .95, standing: .70< r >.80). Resul ts indicated good
postural stability for all cardiovascular measures with correl ations
being slightly larger for SBP as compared to HR and DBP (HR: r= .72 and
.  72, SBP: r= .80 and .82, DBP= r=.71 and .76).  These data lend
support for the notion that cardiovascular response to FCPS is a stable
individual difference across postural contexts.

EEG maturation and the development of ocular artifacts
Riek Somsen and Bert van Beek
University of Amsterdam

The majority of EEG artifact treatment studies have paid little
attention to the specific problems of young children. In children the
amount of ocular artifact in the EEG decreases with age. Until age 6
often more than 75% of the EEG records is contaminated with eye and
head movements. The EEG frequency window which is most susceptible to
artifacts from 0 to 10 Hz also shows the greatest maturational decrease
between 0 to 10 years.  Finally, in children the physical distance
between electrodes is very small.  Both the EOG and EEG electrodes
record eye movements and cerebral activity.  Hence, these artifacts are
difficult to distinguish from cerebral EEG. We used a specifically
designed EEG processing program (eEG_ANalysis).  The children were
between 5 and 12 years (n=142).  The effects of different artifact
treatments in relation to age were studied. The conditions were: 1)
uncorrected EEG. 2) EOG-EEG transfer correction. 3) selection of EEG
epochs that were free of blinks and head muscle artifacts. 4) selection
of EOG epochs with minimal power between 1-6 Hz (minimal horizontal eye
movements). 5) 3 and 4 combined.
Correction reduced EEG power at all scalp locations and frequency
bands and removed more high frequency power with increasing age.  The
three selection conditions yielded much lesser and highly selective
reduction of spectral power at the frontal and central locations, only
for the Delta and Theta bands. Blink selection was relatively stronger
in the youngest, while removal of horizontal eye artifacts was stronger
in the oldest children.

The development of selective attention as indexed by heart rate
Riek Somsen, Maurits van der Molen, Bert van Beek, Monique Geers, Saar 
Langkamp, Nienke Stark.
University of Amsterdam

	Between and within channel auditory selective attention was
examined in two groups of children of 7 and 11 years old by presenting
tone pips of different pitch randomly to opposite ears. Some of the
pips had a slightly higher pitch. The children were instructed to count
the deviant pips at one ear and to ignore all pips at the other ear.
The amount of frequent pips which preceded a rare was varied. Series of
frequents which ended with a rare consisted of 4, 8, and 12 pips.
Cardiac interbeat intervals were sampled twice: one time for the
attended tone pips and one time for the non-attended tone pips. The
cardiac interbeat intervals responded differently to the attended tone
pips. Initially the heart was slowed when the rare was presented,
followed by increased heart rate when the subject was counting. When
adult subjects are waiting for the rare tone to occur, their heart rate
is slowing until the deviant stimulus is detected which is followed by
heart rate speeding. This expectancy response was absent in the youngest 
children and was present in the older children only for the series of 4 
and 8 pips. In the non-attended series no anticipatory deceleration 
occurred. The youngest children made relatively more counting errors. 
These results indicated that the developmental rates of the three 
selective attention functions was not the same: tone discrimination and 
channel separation were reasonably developed in both groups, while the 
expectancy response was absent in the youngest children, while it was 
present for relatively tone series in the 12 year olds.

Visuospatial attention effects on brainstem reflexes and cortical
event-related potentials
Douglas C. Sonnenberg, Kathy A. Low, and Steven A. Hackley
University of Missouri-Columbia

Previous research has shown that selective attention can enhance the
ability of a weak visual prestimulus to inhibit the blink reflex to a
subsequent startle stimulus. We studied this phenomenon in 20 young
adults, with concurrent recordings of visual evoked potentials, the
postauricular reflex, and behavioral indices of perceptual sensitivity
(d') and criterion (B).
Subjects judged the duration (300 or 500 ms) of the illumination of
one of two diodes, which were located 25 degrees to the left and right
of fixation.  A centrally located arrow either correctly or incorrectly
precued the location of that target: p(valid) = .80; p(invalid) = .20.
Event-related potentials were recorded at O1' and O2' sites on such
trials.  On other trials, an intense noise burst followed the target at
an asynchrony of 150 ms.  The ability of the target to produce prepulse
inhibition of the eye-blink and pinna-flexion components of acoustic
startle was measured.
Prepulse inhibition of startle-blink was reliably augmented on
valid as compared to invalid trials.  Similarly, the cortical P1 and N1
potentials evoked by the target/prepulse stimuli were enhanced by
spatially focused attention. (Postauricular reflexes and task
performance measures have not yet been analyzed.). Reflex modulation
and event-related potential data provide converging evidence that
spatially selective attention can influence low-level visual pathways.
Supported by National Institutes of Health grant MH47746 to SAH.

A simulation study of single-trial ERP latency estimation methods
Kevin M. Spencer and Emanuel Donchin
University of Illinois at Urbana-Champaign

One of the key assumptions underlying signal averaging is that each
event triggers a time locked response that has precisely the same
temporal pattern across trials. This assumption is violated if the
latency of ERP components vary from trial to trial. Such "latency
jitter" distorts the shape and amplitude of the ERP. Indeed, it is
improper to compare the amplitudes of ERPs computed over epochs with
varying degrees of latency jitter.
The effects of latency jitter may be counteracted by applying a
method that estimates the latency of the component in each trial.
With these estimates, the trials are re-aligned to compute a new
average. A critical aspect of latency jitter correction is the method
used to estimate single-trial latencies.  Many decisions need to be
made when such a latency estimation method is applied, and the
implications of these decisions under different experimental conditions
are quite variable.
We report here a simulation study in which the accuracy of several
latency estimation methods was assessed. A data set of 15 "subjects"
was simulated, each subject's data consisting of 50 trials with a
component of fixed amplitude embedded in Gaussian noise.  Five latency
estimation algorithms were tested: peak-picking, cross-covariance,
cross-correlation, cross-covariance with wavelet transform, and
cross-correlation with wavelet transform.  We varied in the simulation
the following factors: signal-to-noise ratio, filter cutoff frequency,
template width, and number of wavelet coefficients.  Complex
interactions were observed among these factors, and will be described
in detail.

Receiver operating characteristic analysis of psychophysiological
indices in schizophrenia
Scott R. Sponheim1, Sean M. Nugent1, William G. Iacono2, and John W. 
1Veterans Affairs Medical Center, Minneapolis, Minnesota, 2University of 
Minnesota, 3National 
Computer Systems

Psychophysiological indices have been used to characterize biological
abnormalities associated with schizophrenia; however, few studies have
examined the diagnostic utility of these indices.  We employed receiver
operating characteric (ROC) analyses to determine the extent to which
psychophysiological indices discriminate between schizophrenia patients
and normal controls.
Subjects completed a battery of measures on which schizophrenia
patients have been shown to exhibit abnormalities.  Ocular motor
functioning was assessed during a smooth-pursuit eyetracking task (20
cycles at .4 Hz).  Electrodermal activation was measured during a
habituation task consisting of soft (85 dB) and loud (105 dB) tones.
The visibility of the capillary plexus at the base of the nailfolds was
measured using Maricq's Scale of Plexus Visibility (Maricq, 1970).
Electroencephalograms (EEGs) were collected from subjects in a resting
Univariate ROC analyses yielded area under the curve (AUC)
values that were greatest for indices of ocular motor function
and nailfold plexus visibility.  AUC values for indices of
electrodermal activation and EEG low-frequency and alpha band power
yielded AUC values that were greater than chance group assignments but
lower than AUC values for indices of ocular motor functioning and
nailfold plexus visibility.  Multivariate analyses revealed that
measures of ocular motor functioning and nailfold plexus visibility
together yielded a greater AUC value than either measure alone.
Indices of electrodermal activation and EEG low-frequency and alpha
band power failed to add significantly to the AUC value of ocular motor
functioning and nailfold plexus visibility taken together.

Context- and intensity-effects on psychophysiological emotion responses
Gerhard Stemmler
University of Marburg

Autonomic activity and self-reports of emotion were collected from
N=3D158 female Ss under a crossed Group (Ss blind or informed about
impending emotion inductions) x Emotion (fear, anger) x Emotion
Intensity (high, medium) x Emotion Induction Context ("real-life",
imagery) design,with Context presented in two separate sessions. The
real-life context per se elicited larger autonomic and subjective
responses than imagery.  Blind compared to preinformed Ss exhibited
much larger physiological responses in the real-life context, but not
during imagery. Self-reported emotions, however, were similarly large
in both contexts. During imagery, autonomic activity of fear and anger
was practically indiscriminable, but highly differentiated during
real-life inductions. Intensity influenced several autonomic variables.
Results are discussed in terms of current theorizing about "basic"
emotions and emotional specificity.

Is the EGG a marker for infant colic?  Results from a prospective study
Cynthia Stifter, Dave Zelis, Jane Mihailoff, and Ken Koch
Pennsylvania State University

Several hypotheses exist regarding the etiology of colic.  What is
undebatable is that infants who have colic appear to be experiencing
discomfort that seems tocenter around the GI tract.  Studies of adult's
gastric myoelectrical activity (EGG) suggest that dysrhythmia or
tachygastria is associated with nausea, vomiting, and motion sickness.
To examine whether infants with colic present with different gastric
myoelectrical activity, a prospective study was conducted in which
EGG's were recorded at 2 weeks of age before the onset of colic, and at
6 weeks of age during the period colic emerges.  Attempts were also
made to record EGG at 8 weeks of age during a colic bout.  EGG's were
recorded postprandially for 30 minutes using a portable computer.
Special procedures designed to detect and remove movement artifact were
applied to the data.  Spectral densities in the following frequencies
were calculated: bradygastric (1-2.5cpm), normal (2.5-3.75cpm),
tachygastric (3.75-10cpm), and duodenal (10-15 cpm). Recordings were
available for 61 two week olds and 68 six week olds.  Colic was
identified in 7 of the infants.  Results revealed that at 2 weeks colic
and noncolic infants exhibited significantly different EGG's.  Infants
with colic had less bradygastria and more tachygastria.  At 6 weeks,
only the bradygastric range was different for the two groups.  Only 1
colic infant was recorded during a colic bout but he exhibited large
amounts of tachygastria compared to the other frequencies.  The results
of the present study suggest that EGG may be a physiological marker for
the identification of infants susceptible to colic.

Ouch, that hurts!  Blood pressure and heart rate responses to a dental
Evelyn R. Sullivan, Bruce N. Cuthbert, Katherine Karpinia, Arthur Hefti, 
and Peter J. Lang.
University of Florida

This study is part of an ongoing project investigating the
relationship between psychophysiological responses to stressful and
potentially fear-provoking situations in both real-life settings and in
corresponding laboratory representations.  In the present study,
questionnaire and interview measures were used to assign 35 college
students to either low, medium, or high dental fear groups.  Subjects
underwent a periodontal exam including a manual gingival exam and an
invasive examination with an automated probe (Florida Probe) to measure
gum attachment and pocket depth.  Systolic (SBP) and diastolic blood
pressure (DBP), and HR were measured before, during, and after each
	Overall, analyses showed highly significant SBP and DBP
increases from an initial resting baseline for all exam
components.  BP increased during the gingival exam, although not
significantly relative to immediate pre-gingival exam levels.  BP
levels increased significantly in anticipation of the probe procedure
and continued to rise during the probe exam, finally plateauing during
the final recovery period.  Compared to low fear subjects, medium fear
subjects showed significantly larger SBP responses to the probe, with
high fear subjects falling in between the other two groups; these
increases were maintained during recovery.  No fear group differences
were present for HR.  Interestingly, HR levels did not change from
baseline during anticipatory periods, but did drop significantly during
the gingival exam and probe procedures, remaining below baseline levels
during recovery.  This pattern of results suggests a passive behavioral
set to the procedures in all groups, with BP increases likely modulated
by peripheral vasoconstriction.

Resting anterior EEG asymmetry predicts affect-related information
Steven K. Sutton1, Richard J. Davidson1, and Gregory M. Rogers2
1University of Wisconsin in Madison, 2Northwestern University

Previous research has shown that individuals with greater left (right)
anterior resting EEG asymmetry experience more positive (negative)
affect and behavioral approach (inhibition).  Using these data, it was
predicted that this brain activity measure would be related to the
processing of stimuli that varied in affective tone.  82 undergraduates
(44 F) completed two resting EEG sessions between 4 and 28 months prior
to performing a repeated, simple choice task where two word-pairs were
presented simultaneously.  The subject chose the word-pair that "went
together best."  Word-pairs were previously categorized as either
negative (e.g., criminal-prison), neutral (e.g., book-cover), or
positive (e.g., won-victory); with 48 exemplars in each category.
Subjects responded to a quasi-random sequence of negative-neutral,
negative-positive, and neutral-positive comparisons.  Individuals with
greater left anterior (FpF1/FpF2) resting EEG asymmetry were more
likely to select positive word-pairs (r = .25, p = .02); whereas
individuals with more right anterior resting EEG activity showed a
tendency to select negative word-pairs (r =
.20, p < .07).  Resting EEG asymmetry from central and posterior
regions did not predict word-pair selection (r's ranged from .06 to 
-.11, p's >.31).  Furthermore, the 26 individuals exhibiting relatively 
stable anterior EEG asymmetry showed a stronger significant relation 
between left anterior EEG asymmetry and the selection of positive word-
pairs (r= .40, p = .04).  These data suggest that anterior brain regions 
are related to the processing of affect-related stimuli.

ERP evidence for effects of imageability and semantic distance in word
Tamara Swaab1, Kathleen Baynes1, and Robert T. Knight1,2
1Center for Neuroscience, UC Davis, 2VAMC Martinez

In a previous study we found that non-aphasic right hemisphere (RH)
patients failed to show a modulation of the N400 to non- associatively
related target words that were from the same semantic category
(Hagoort, Brown, & Swaab, in press). The same RH patients showed
relatively normal associative N400 priming effects. This result might
indicate that the RH is important in the processing of distantly
related words (coarse semantic coding). However, the words that were
used in the semantic condition of this study were highly imageable. It
has been proposed that the RH plays a role in the processing of highly
imageable words. In the present study, we examined the separate
contributions of semantic distance and imageability on the amplitude of
the N400. Four stimulus lists of 80 word pairs were constructed, half
the pairs were unrelated. Two lists contained word pairs that had
strong associative relations, and two lists contained 40
non-associatively related word pairs that were from the same semantic
category. One of the associative lists and one of the semantic lists
contained highly imageable words, the other two lists contained weakly
imageable words.  The 320 word pairs were presented auditorily, in
random order. The results showed that the distance of the semantic
relation and the degree of imageability both influenced the size and
the topographic distribution of the N400 effects. These results will be
evaluated in terms of dual-coding theory versus context-availability
theory. Possible implications for the role of the RH in word processing
will be discussed.

When do we dream? Comparison of dreams at sleep onset between REM and
nREM periods
Tomoka Takeuchi1, Akio Miyasita1, Maki Inugami1, and Yukari Yamamoto2
1Tokyo Metropolitan Institute for Neurosciences, 2Waseda University

No one can answer conclusively whether REM dreams differ from NREM
dreams in quantitative and qualitative terms. In comparison of dreams
between REM and NREM periods, many researchers failed to control dreams
elicitation and/or analysis, in terms of a)uncontrol of the sleep
length or stages before the appearance of REM or NREM (stage 2)
periods, and b)contingent of extraneous variables ascribed to
experimenters or subjects in sampling or analysis of data.  Our
experiment achieved control of these problems. By a sleep interruption
technique (Miyasita et al., 1989, Electroenceph. clin. Neurophysiology,
73, 107-116.), we elicited sleep onset REM periods(SOREMP). This
technique enabled us to compare dreams at sleep onset between REM and
NREM (stage 2) sleep with their latencies controlled. Further, in
sampling dreams, we had subjects checked on standardized rating scales,
which evaluated properties of dreams. These procedures yielded the
following results. a)Subjects recalled their dreams more from SOREMP
than NREM period. b)The properties of SOREMP dreams differed from NREM
dreams. c)In case the subjects recalled dreams, physiological states
before the appearance of rapid eye movements and of sleep spindles
(K-complex) varied from each other.  The results allowed us to
speculate that the process of originating dreams varied between REM and
NREM periods.

Psychopathology and electrodermal response modulation in adolescent
Jeanette Taylor, Scott R. Carlson, William Iacono, David T. Lykken, and 
Matt McGue
University of Minnesota

Electrodermal response modulation to an imminent but predictable
aversive stimulus is viewed as adaptive.  Given their inability to
reduce the impact of aversive stimuli, poor modulators may experience
problems in certain areas of psychological functioning.  We expected
poor modulators to have more psychopathology than good modulators.  In
order to test this hypothesis, we assessed the skin conductance
response (SCR) of 150 adolescent males (aged 16-18 years) participating
in the Minnesota Twin Family Study (MTFS).  An aversive 90db white
noise blast was presented during five 100 second trials [two
unpredictable (UP) and three predictable (P)].  Response modulation was
indexed with the preception score [100 * (mean UP - mean P) / mean UP],
which reflected the percent increase or decrease in SCR when the blast
is made predictable.  From the pool of 150 subjects, three modulation
groups were derived based on their having poor, good, or moderate
preception scores.  The Poor group was comprised of subjects with the
lowest 25 preception scores (mean = -43.17), the Good group were those
with the highest 25 scores (mean = 60.25), and the Moderate group
consisted of a random sample of 25 from the middle of the distribution
(mean = 25.50).  The three groups were compared on symptom counts of
externalizing, internalizing, and substance use disorders.  Good
modulators had significantly fewer symptoms of alcohol use disorders
than both the Moderate and the Poor modulators (who did not differ
significantly).  No group differences were found for the other
disorders.  These results suggest that good modulation ability may
serve a protective function against alcohol use disorders.

Condition and modality effects on late positive ERP components obtained
during continuous performance tasks
Ayda Tekok-Kilic and David W. Shucard
State University of New York at Buffalo

Previous studies within our laboratory with a visual Continuous
Performance Task (CPT- AX) revealed that scalp-recorded Late Positive
Components (LPCs) were affected by different stimulus sequences or
conditions that were part of the CPT. Following these findings, we have
investigated the effects of condition variables and modality on scalp
recorded LPCs, specifically on the P300 and NoGo P3. Both auditory and
visual versions of the CPT-AX paradigm were presented to right handed
male and female volunteers between 18-35 years in a counterbalanced
order. CPTs were the same in terms of the number of stimuli (n=508),
type of stimuli (letters), and the stimulus sequences. ERPs were
obtained from 12 scalp sites. Here data are reported for midline scalp
sites. The results showed that the amplitude, latency and the scalp
topography of the LPCs (P300 and NoGo P3) changed as a function of the
cognitive processes required by the different stimulus sequences
(conditions) within the CPT task. The NoGo P3 in the A-not-X condition
had the highest amplitude and longest latency in the more anterior
sites. In the target detection condition (A-X), the P300 component
showed a maximum amplitude in the more posterior scalp sites.  These
condition-related topographic findings were similar for both auditory
and visual modalities. Results are discussed in terms of both modality
and condition effects on P300.

Analysis of correlated data: Multiple ANOVA's vs. MANOVA
K. A. Thayer and J. F. Thayer
University of Missouri-Columbia

Researchers often collect data on multiple, correlated dependent
variables. For example, multiple cardiovascular variables such as heart
rate, systolic blood pressure, and diastolic blood pressure, or
multiple channels of evoked potential data are collected in a given
experiment. However, these data are often analyzed one variable at a
time using multiple univariate ANOVA's. This approach fails to take
into account the relationships among the dependent variables. This is
particularly important given that subtle effects in the pattern of
response can be masked by this approach. In the present paper we
illustrate the effect of various degrees of correlation among the
dependent variables on the results obtained by multiple ANOVA's versus
MANOVA.  Data from two experiments, one involving prenatal exposure to
diethylstilbesterol (DES) in mice, and the other involving a
cardiovascular reactivity paradigm, were analyzed using multiple
ANOVA's and MANOVA. The correlations among the dependent variables were
assessed using Pearson correlations. Results indicate that when the
dependent variables were moderately correlated it was possible to
obtain significant ANOVA's for all of the dependent variables even when
the MANOVA was non-significant. Moreover, when the correlations were
low it was possible to obtain a significant MANOVA without a single
significant univariate test after correction for Type I errors. These
results suggest that MANOVA may provide increased sensitivity to
patterned responses in psychophysiological data and highlight the
different research questions addressed by ANOVA and MANOVA.

Stress responses and motivational systems
Joe Tomaka, Rebecca Palacios-Esquivel, Julie A. Penley, and Sandra D. 
University of Texas at El Paso

This study examined the relationship of stress-related patterns of
cardiovascular reactivity to underlying motivational systems.  We
hypothesized that stress responses consisting of high cardiac
reactivity coupled with a decline in systemic vascular resistance
(e.g., challenge, fight or flight, Pattern 1) primarily reflect the
activity of the behavioral activation system (BAS), and that stress
responses consisting of moderate cardiac reactivity coupled with an
increase in systemic vascular resistance (e.g., threat,
vasoconstrictive, Pattern 2) reflect co-activation of behavioral
activation and behavioral inhibition systems (BIS).  To activate the
BAS, all participants performed a simple but engaging computer math
task which allowed them to accumulate a very high point score for three
minutes.  After three minutes, and in order to co-activate the BIS,
half the subjects began receiving difficult/unsolvable math problems,
causing them to loose all of their previously accumulated points (i.e.,
punishment manipulation).  The punishment manipulation, within the
context of previous BAS activation, was expected to represent a state
of co-activation of BAS and BIS systems resulting in a decline in
cardiac activation (heart rate and pre-ejection period), and an
increase in vascular (total peripheral resistance) and electrodermal
(skin resistance) reactions.  Results supported the hypotheses for the
cardiac variables (F[1/25] = 9.19, p < .01, for PEP, and F[1/25] =
2.44, p = .08, for HR), but not for the vascular and electrodermal
variables (both ns).   The lack of significant effects for TPR may have
been due to the tendency for subjects in general to have strong
positive vascular reactions to the computerized math task.

Event related potentials (ERPs) in Obsessive-Compulsive Disorder (OCD)
during verbal and nonverbal directed-attention tasks
J.P. Towey1,2, G.E. Bruder1, C.E. Tenke1, P. Leite1, R. Fong1, and M. 
1NYS Psychiatric Institute, 2Mercy College

Our prior studies have found ERP abnormalities in OCD, including
enhanced attention-related negativities to nonverbal sounds suggestive
of overfocused attention and/or hemispheric dysfunction. The current
study investigated both verbal (consonant-vowel) and nonverbal (complex
tone)  directed listening  tasks in 18 OCD patients and 20 normal
controls.  ERPs were recorded from 26 lateral electrodes to monaural
stimuli matched in intensity, duration and repetition rates.  After
correcting for eye artifacts, ERPs were submitted to repeated measures
ANOVAs using averages over the latency windows of interest, such as
early  (160-250 ms) and  late  (500-1000 ms) components of
processing negativity  (PN).  PN components were clearly visible in
difference waveforms for nontarget data (attend minus unattend
conditions). The following Group differences were found: (1)  early
PN:  Unlike OCD patients, normal controls had larger early PN in the
nonverbal task over right than left hemisphere sites (Group X Condition
X Hemisphere:  p=.04). This asymmetry in normals, evident in the
difference waveforms at most sites, may reflect a right hemisphere
advantage for processing complex/musical tones.  OCD patients had an
ERP asymmetry in the opposite direction. Patients produced larger
attention-related negativity over left than right hemisphere sites,
consistent with our prior findings for nonverbal stimuli. (2)  late
PN:  For the verbal task, OCD patients differed from controls in  late
PN  asymmetry (Group X Condition X Hemisphere: p=.03). The control
group had a larger late PN over left than right hemisphere sites in
this task, possibly due to left hemisphere dominance for
verbal/linguistic processing.  There was a small asymmetry in the
opposite direction for OCD patients. These results can be interpreted
as consistent with hypothesized left-hemisphere dysfunction in OCD.

Comparison of CNV amplitude and P300 latency and amplitude in subjects
practicing the Transcendental Meditation Technique for less than 1 year
or more than 8 years.
Frederick Travis
Maharishi University of Management

To investigate possible long term physiological effects of
Transcendental Meditation (TM) practice, 9 subjects, average age 20.7
yrs with 1.1 yrs TM experience, and 12 subjects, average age 20.7 yrs
with 9.3 yrs TM experience, were given a S1-S2 oddball task.  The task
comprised a "+" (S1) followed by two single-digit numbers (S2), 2 s
later that subjects were asked to multiply together.  If the resultant
product was less than 39, they pressed a left-hand button (standards-
80%); if it was greater than 39, they pressed a right-hand button
(rare-20%).  EEG was recorded from Fz, Cz, and Pz, referenced to linked
mastoids.  Horizontal EOG was used for artifact rejection.
CNV amplitudes were significantly lower in the group with 9 yrs TM
experience (F(1,19) = 9.6, p = .006).  No significant group effects
were seen in P300 latency, P300 amplitude, reaction time or accuracy;
however significant task effects were seen for P3 latency (F(1,19)=4.1,
p=.057) and reaction time (F(1,19)=7.4, p = .016) as expected in an
oddball paradigm.  Tecce's two-process model relates CNV amplitude to
attention and arousal.  Reduced CNV amplitudes with similar P3
latencies and reaction times to targets might suggest lower tonic
arousal with sustained ability to respond to stimuli.  A similar
pattern of lower baselines with quick responses has also been reported
in skin conductance orienting and immune system functioning in subjects
with extensive TM experience.

The MMN is sensitive to perceptual changes in the absence of  physical
changes in auditory stimuli
Leonard J. Trejo and Tara M. Johnson
University of Illinois

The Mismatch Negativity (MMN) is believed to arise from comparisons
of deviant auditory stimuli with a sensory memory of a standard
auditory stimulus.  In the present study we sought to assess the
influence of higher-order perceptual effects on the MMN, using an
auditory illusion.  Our approach was to use a visual-auditory illusion
(the McGurk effect) to make the standard auditory stimulus sound like a
deviant auditory stimulus.
Subjects (n=7, ages 19-25 y, 4 females) watched movies in which an
actor recited a pseudorandom series (4800 trials) of syllables in an
oddball paradigm (e.g., 80% "MA", 10% "KA", 10% illusory "NA").  We
created the illusory syllables by combining visual cues from deviant
stimuli with auditory cues from standards (e.g., KA-lips + MA-sound =
illusory "NA").  Each subject performed the oddball task under two
conditions (MA-KA-NA; BA-GA-DA). All seven subjects reported hearing
the illusion in both conditions. Subjects pressed a button in response
to the auditory deviants (KA, GA) while ERPs were recorded from 30
scalp electrodes, referenced to nose. We computed average ERPs for each
subject x stimulus x condition and measured MMNs in difference waves
(deviant auditory - standard, illusory deviant - standard).
Preliminary analyses suggest that both the physical and illusory
auditory deviants elicit a MMN. This suggests that the MMN generator
does not merely rely on physical stimulus properties in comparing
deviant with standard auditory stimuli.  Such information appears to
include higher-order perceptual interactions such as the auditory
illusions of the McGurk effect.

Ethnographic notions of the relationship between physiology and
reported affect in Chinese and European cultures: A test of opposing
Jeanne L. Tsai and Robert W. Levenson
University of California, Berkeley

Ethnographic studies of Chinese culture lead to two opposing
predictions.  Compared to European cultures: (1) stronger Chinese
beliefs in the mind-body relationship predict a stronger relationship
between physiology and self-reported affect; and (2) lesser Chinese
reliance on internal states (and greater reliance on interpersonal
contexts) in determining affect predicts a weaker relationship between
physiology and self-reported affect.
We examined the relationship between autonomic and somatic
physiology and self-reported affect in 48 Chinese American and 50
European American dating couples while they discussed conflicts in
their relationships.
For each cultural group, 20 multiple regressions were conducted
using physiology to predict global and specific affects (overall
affect, extreme positive affect, extreme negative affect, amusement,
anger, anxiety, confusion, contempt, contentment, disgust, excitement,
fear, interest, pain, relief, sadness, satisfaction, shame, surprise,
tension). Physiology was related to specific positive and negative
affects for European Americans (greater autonomic activation related to
less contentment and amusement and more confusion and fear; greater
somatic activity related to more excitement and less confusion).  For
Chinese Americans physiology was related to global and specific
negative affects (greater autonomic activation related to more negative
overall affect, extreme negative affect, pain, and sadness; greater
somatic activity related to more positive overall affect and disgust).
Our findings support ethnographic notions of closer mind-body
relationships for Chinese in global affect, but not specific affects.
Opposing notions that Chinese base their affect less on internal states
was supported only by the lack of relationship between physiology and
specific positive affects.

Clinical certification in electrophysiology
Patricia Tueting
University of Chicago

The American Psychiatric Electrophysiology Association (APEA) has
appointed a committee to consider the issue of whether there is a need
to establish a certification mechanism for electrophysiologists engaged
in recording EEG and evoked potentials for the purpose of aiding
clinical judgments on abnormal neurological and behavioral function and
treatment.  Electrophysiologists conducting research in clinical
settings are often called upon to contribute to patient care, either as
a part of their clinical research position or in return for
remuneration that can be used to fund research.  As with any clinical
activity, it therefore becomes important to establish educational
processes and guidelines to ensure that the research data base is being
used appropriately and that patient safety and quality of care remain
paramount.  APEA members are writing position papers on the clinical
application in psychiatry of basic reserach in the areas of: 1)Clinical
and QEEG, 2) Event-related potentials (ERPs), 3) Polysomnography, and
4)Biofeedback.  Existing certifying boards in Clinical EEG,
Polysomnography, Biofeedback,  Neurophysiological Monitoring (in
operating rooms) are being reviewed by the committee. A list of
certifying programs with summaries of their scope, training, and
eligibility requirements is being compiled.

Event-related potential (ERP) measures of auditory sensory gating:
Pitch and interval parameters
Patricia Tueting1 and Nashaat Boutros2
1University of Chicago and 2Yale University

A systematic decrease in event-related (ERP) amplitude has frequently
been observed as a function of stimulus repetition and shortened time
interval between stimuli.  Dynamic processes of neuronal inhibition,
e.g., sensory gating, have been theorized to account for these
findings.  In order to investigate parameters related to this proposed
inhibition, ERPs were recorded in three different paradigms in the same
22 healthy normal subjects.  Paradigm 1:  Long sequences of pairs of
clicks separated by a 500 ms time interval between S1 and S2 were
presented with an interpair interval of 2 s.  S1 and S2 were either
identical or non-identical in pitch.  Paradigm 2:  Long sequences of
two alternative stimuli varying in pitch were randomly presented at
15%/85% relative probability with 2 s between stimuli and at least 4 s
between low probability stimuli.  Paradigm 3:  Blocks of trials at
varying interstimulus intervals - 0.5 s, 2 s, 4 s, and 8 s were
presented.  The gating effect observed in the paired stimulus paradigm
was more pronoucned when the clicks were identical in pitch than when
they were non-identical.  For the oddball paradigm, P50 was larger when
elicited by the low probability stimulus, both when the rare stimulus
was lower and when it was higher in pitch than the high probability
stimulus.  The counterbalanced design also allowed for an examination
of P50 amplitude across paradigms in the same subjects.  The findings
will be discussed in relation to theories of neuronal inhibition,
gating, and cognition.

Expanding the irrelevant-probe technique: Novel auditory probe
sensitivity to workload changes in a complex dual-task paradigm.
Peter Ullsperger1 and Darryl G. Humphrey2
1Bundesanstalt Arbeitsmedizin; 2Wichita State University

The goal of this study was to explore the utility of the novel
auditory probe technique of mental workload estimation in a complex
dual-task paradigm.  In addition to a baseline auditory oddball task,
fifteen subjects performed a mental arithmetic task and a gauge
monitoring task in single task and dual task conditions.  The
difficulty of the gauge monitoring task was manipulated to give
conditions of graded difficulty. Throughout task performance three
types of auditory stimuli were presented: standards (80%) targets (10%)
and, novel sounds (10%).  In the oddball task subjects counted the
number of target stimuli.  In the workload tasks the auditory probes
did not require a response.  ERPs were recorded from the occurrence of
the tones in both the baseline oddball task and in the three
combinations of the gauge monitoring and arithmetic task.  P300
amplitude elicited by targets was sensitive only to the introduction of
the workload tasks. The amplitude of the novelty P3 tended to follow
the gradations of mental workload.  To further characterize the novelty
P3 a scalp dipole source analysis was conducted.  = In the time range
of novelty P3 two equivalent dipoles were active.  Preliminary analyses
suggest these dipoles are differentially sensitive to changes in mental
workload.  Obtaining these results in a dual-task environment extends
the realm in which the irrelevant auditory probe technique has
demonstrated sensitivity to changes in mental workload. The use of the
novel probes in addition to the targets and standards allows a finer
grained analysis of this sensitivity.

Accessory stimulus (prepulse) effects on the lateralized motor
readiness potential.
Fernando Valle-Inclan and Steven A. Hackley
University of La Coruna and University of Missouri-Columbia

Comparisons at the Missouri lab of prepulse effects on voluntary and
reflexive reactions have suggested that divergent mechanisms are
involved. The present study attempted to identify whether a
task-irrelevant prepulse speeds choice reaction time (RT) by
facilitating low-level motor pathways (which might be shared with
reflexes) or central, decision-level structures (which reflexes lack).
On each trial, the subjects (N=20) made a speeded left- or
right-hand keypress to the letter "S" or "T" if that stimulus was of
one size ("Go" trials), but withheld their response if it was of
another possible size ("No-go" trials).  On half of the trials, a
task-irrelevant tone-pip accompanied the letter with a lead time of 30
On Go trials, choice RTs were reliably facilitated by these
accessory stimuli. Lateralized motor readiness potentials (LRPs) were
separately computed with time-locking to stimulus onset and to movement
onset. Comparisons showed that facilitation of RTs was due to
shortening of the time interval extending from stimulus onset until LRP
onset (Miller's Jack-knife test).  The interval from onset of
lateralization until keypress was unaffected.  A brief LRP was observed
on No-go trials, in the absence of EMG activation.  This abortive
activation of motor cortex was also facilitated by the accessory
stimulus.  The results converge with those of reflex-RT studies to
contradict the hypothesis that accessory stimulus effects on RT are due
to facilitation oflow-level motor pathways.  This research was
supported by the Ministry of Education (Spain) and the National
Institutes of Health (U.S.A.).

Optimal recording of electric, acoustic, and visual blink reflexes:
Effects of EMG signal bandwidth and inter-electrode distance
A. van Boxtel, A.J.W. Boelhouwer, and A.R. Bos
Tilburg University

Blink reflexes are usually measured by integrating the orbicularis
oculi EMG signal.  EMG signals are high-pass filtered to remove
low-frequency movement artifacts.  Filtered responses are sometimes
small and difficult to detect which can bias experimental outcomes.
Mathematical models of the surface EMG signal predict that increasing
the inter-electrode distance will produce smaller high-frequency and
larger low-frequency spectral components with a net increase in signal
power.  However, this may provide less reliable results because the
spectral shift to lower frequencies leads to a larger contamination by
low-frequency artifact components.  We have studied the optimal EMG
recording bandwidth and inter-electrode distance in order to maximize
signal amplitude and minimize artifact effects.  Electric, acoustic,
and visual blink reflexes were elicited in three different groups of 15
subjects.  Fifty reflexes were elicited at irregular intervals during
the presentation of a cartoon film.  Two pairs of electrodes were
applied on orbicularis oculi with inter-electrode distances of 12 and
36 mm.  EMG signals were recorded with a bandwidth of 0.4-520 Hz,
digitally high-pass filtered with the -3 dB cut-off frequency
successively increasing from 4 to 92 Hz in steps of 8 Hz, and subjected
to spectral analysis.  Low-frequency artifact components were
relatively larger for electric and acoustic than for visual blinks.
The optimal high-pass cut-off frequency was 28 Hz for electric and
acoustic reflexes and 20 Hz for visual reflexes, both for the large and
small inter-electrode distance.  The total signal power was
systematically larger with larger inter-electrode distance.

The use of reflexes in chronopsychophysiology and in the study of
response inhibition
G.J.M. van Boxtel12, R.H.A.H. Jacobs1, M.W. van der Molen2, J.R. 
Jennings3, and C.H.M. Brunia1
1Tilburg University, 2University of Amsterdam, 3University of Pittsburgh

Spinal reflexes are important for chronopsychophysiology because they
provide an additional time marker between the currently available
cortical (LRP) and peripheral (EMG) measures. They are particularly
important for the study of response inhibition, which is commonly
modelled by a "horse race", in which independent activation and
inhibition processes race for completion. This race is suspected to
finish very late in the information processing system, since response
inhibition is accompanied by cortical activation (as indexed by the
LRP), which does not lead to overt movement (as indexed by the EMG).
Hence it is interesting to study a neurophysiological system in between
the motor cortex and the effector muscles, such as the spinal cord.
Achilles tendon reflexes were evoked in an auditory Go/NoGo 50/50 task
involving a response with the right hand. In experiment I, reflexes
were evoked at 100, 200, 300, 400, 500, and 600 ms after the stimulus.
Experiment II consisted of two sessions, the first being equal to
Experiment I. In the second session, reflexes were evoked at latencies
relative to the mean reaction time of the first session. The results
indicated that reflexes increased both on Go and NoGo trials at 100 ms
after the stimulus. They then returned to baseline on NoGo trials, but
sharply increased on Go trials starting from 80-100 ms before the
response. These results show that Go and NoGo trials can be succesfully
distinguished at the spinal level. The use of reflexes for
chronopsychophysiology and response inhibition is discussed.

When does phonology meet semantics?  ERP evidence for early semantic
Cyma Van Petten, Susan Rubin, Elena Plante, and Marjorie Parks
University of Arizona

We established the "isolation point" for a set of auditory words - the
amount of acoustic information necessary to identify a word.  Subjects
were presented with the first 50 ms of each word, or the first 100,
150, etc. ms (up to the end of the word) and asked to guess what they
were hearing. Mean word duration was 600 ms, but accurate word
identification was achieved with 350 ms of input on average. These
words were then used to complete spoken sentences, and ERPs were
timelocked to both word-onset, and to the isolation point. The
sentence-final words were either 1) semantically congruous, 2)
incongruous, but sharing initial phonemes with the expected word, 3)
incongruous, but sharing final phonemes (rhyming words), or 4)
completely incongruous.  For example, "The highway was flooded so they
had to take a long DETOUR / DETAIL / CONTOUR / TABLE." All of the
incongruous conditions elicited larger N400s than congruous. The
congruity effect began at about 150 ms after the onset of completely
incongruous and rhyming words, but was delayed until about 350 ms in
the case of incongruous words with appropriate initial phonemes. The
congruity effect for words with appropriate initial phonemes began at
about the isolation point.  For words with inappropriate initial
phonemes, the congruity effect began well before the isolation point.
These results suggest that the N400 congruity effect is timelocked to
what can be called the "discrepancy point" - the first point at which
acoustic input diverges from semantic expectations. The results
additionally suggest that semantic processing begins before a word can
be uniquely identified.

An ERP study of the effects of cueing and switching on the processing
of compound stimuli.
J.L. van Velzen1, A.A. Wijers1, D.Vorberg2, A. Heinecke2, L.J.M. 
Mulder1, and G. Mulder1.
1University of Groningen, 2Technical University of Braunschweig

Visual information can be thought of as an hierarchy of properties: at
the most local level information about the smallest details is
represented, and at the most global level information about the object
as a whole. From several behavioral studies it appeared that attention
can be directed effectively to local or global features of a compound
stimulus (e.g. Kinchla, Solis-Macias & Hoffman, 1983, Perception and
Psychophysics, 33, 1-10).
In this study multichannel EEG recordings used to investigate the
temporal and topographical effects of precueing one level of a
compound stimulus and of switching attention between the two levels.
The behavioral results showed that more time is needed to switch
attention from the local to the global level than vice versa.
Furthermore, from the ERPs and the topographical distribution of
recorded activity it seemed that different brain structures were
involved, dependent on the direction of the switch. A parietal centre
of activation was found when subjects switched from local to global,
and more occipital activation when a switch was made from global to
local. The ERP-effects of orienting attention to the global level as
compared to orienting to the local level are evident in the
cue-stimulus interval as well as in the interval following the
presentation of the compound stimulus. The latter effects consisted of
a larger P1 amplitude and a smaller occipital and parietotemporal N1
amplitude for locally cued conditions.
From the results of this experiment it can be concluded that
zooming out involves a fundamentally different process than zooming
in.  An alternative explanation would be that an additional process is
involved in processing the compound stimuli in this task, that
contributes to the effects found in this experiment. We investigated
the role of a negative priming mechanism as a possible candidate for
this additional process.

Resting EEG predicts performance in a subsequent vigilance task
Andrei Vedeniapin, John Rohrbaugh, Erik Sirevaag, and John Stern
Washington University School of Medicine

Individual differences in resting EEG were studied with the aim of
identifying features that might be premonitory of performance in a
subsequent vigilance task.  Multi-channel EEG data were obtained from
13 subjects under eyes open conditions, after which subjects
participated in a continuous performance task lasting 50 min.  The task
required subjects to monitor visually-displayed numerals to detect and
respond to occasional runs of three consecutive odd or even digits.
EEG spectral power was derived by FFT from an artifact-free 30 sec
period, with power partitioned into delta (1-4 Hz), theta (4-7.5 Hz),
alpha 1 (7.5-9.5 Hz), alpha 2 (9.5 to 10.5 Hz), alpha 3 (10.5 to 13
Hz), and beta 1 (13-18 Hz) bands.  A coefficient of prevalence of slow
or fast alpha power was also calculated in terms of the normalized
difference between alpha 1 and alpha 3 power.  A subjective measure of
sleepiness was obtained using the Stanford Sleepiness Scale,
administered before and after the vigilance task.  Several features of
the initial resting EEG were correlated significantly with subsequent
performance.  Subjects who performed accurately showed a prevalence of
right occipital alpha 3 spectral power. Increased theta and delta power
was observed in subjects with the fastest reaction times. Right
occipital alpha 1 power (but not slower activity) was associated with a
large increase in sleepiness during the vigilance task. In sum, these
data indicate that features of the resting EEG can be used to predict,
to some extent, subsequent performance in a sustained attention task.
Supported by FAA Contract 94-G-015.

CNV is largest after partial cueing
Rolf Verleger, Bernd Wauschkuhn, Edmund Wascher, Torsten Niehoff, and 
Peter Trillenberg
Medical University of Luebeck

CNV was recorded in an S1-S2 task where S2 could signal one of four
alternative responses: right keypress, left keypress, looking right,
looking left. S1 provided full or no or partial information about the
response, with partial information specifying either the side of
movement (right or left) or the effector system (hand or eye). The
paradigm worked well: Responses were fastest after full infomation and
slowest after no information. If reflecting movement preparation, the
pre-S2 CNV should be largest after full information and smallest after
no information. If reflecting expectancy of information, the pre-S2 CNV
should be largest after no information. However, in contrast to both
assumptions, CNV was largest after partial information. This finding
was replicated in a second experiment. Topography (19 recording sites)
did not differ between conditions, always displaying a Cz maximum.
Therefore CNV appears to reflect the same process in each cueing
condition.  Assuming this process to be movement preparation, the large
CNV after partial information might reflect subjects' simultaneously
preparing two movements (even antagonistic ones, e.g. looking right and
left). This conclusion would support part of Miller's position about
partial cueing in his debate (J. Exp. Psychol. HPP 1982-85) with Reeve
& Proctor (cueing indeed affecting movement preparation) and part of
Reeve & Proctor's position (cueing having benefits for any pair of
movements that share a common code). Besides, a constant
posterior-parietal asymmetry was found (PO6 > PO5), perhaps due to the
eye-hand coordination needed in this task.

Order effects related to habituation in the central and peripheral
nervous systems
Lara Versluys1, Tomas Furmark1, Hakan Fischer1, Gustav Wik2, and Mats 
1Uppsala University, 2Karolinska Institute

Spider phobics and non-fearful subjects viewed spider and neutral
video's (parkscenes) presented twice while regional cerebral blood flow
(rCBF) was measured using positron emission tomography with [15-O]
butanol as tracer. Autonomic indices included heart rate and
electrodermal activity. Order effects were studied by substractive
imaging methodology.  An order effect was observed in non-fearful
subjects viewing spider but not park scenes. A significant rCBF
decrease was evident in the primary and secondary visual corticies (
Brodmann areas 17, 18 and 19 ), with a relatively greater decrease in
the right than the left hemisphere. This order effect might reflect
learning related to habituation processes since heart rate decrement
was also present.  This indicates parallel central and peripheral
processes at a group level. Correlations between rCBF order effects and
heart rate decrement suggest that individual differences in heart rate
habituation correspond to individual differences in central neural

Autonomic mechanisms underlying the cardiac defense response in humans
Jaime Vila1, M. Carmen Fernandez1, M. Nieves Perez1 and Gustavo Reyes2
1University of Granada, 2University of Jaen

The sympathetic and parasympathetic mechanisms underlying the heart
rate response to intense auditory stimulation (the so-called cardiac
defense response) were concurrently re-examined in three studies which
used different autonomic indices of cardiac control: Pulse Transit Time
(sympathetic), Stroke Volume (sympathetic) and Respiratory Sinus
Arrhytmia (parasympathetic). The results of the three studies clearly
confirm the description of the heart rate response as consisting of
four consecutive components: A first and short acceleration that peaks
around second 3, followed by a deceleration that generally surpasses
that of baseline, and a second and long acceleration that peaks around
second 30, ending with a second deceleration. Analysis of the autonomic
space underlying the observed heart rate changes shows a sequencial
pattern of sympathetic-parasympathetic interactions along the four
components of the response with co-inhibition (first acceleration),
co-activation (first deceleration), sympathetic
activation-parasympathetic inhibition (second acceleration) and
sympathetic inhibition-parasympathetic activation (second
deceleration). The finding of a vagal dominance during the first two
components of the response, restraining the sympathetic influences, and
a reciprocal interation during the last two components, with
sympathetic dominance, gives support to the idea that the defense
reaction in natural settings follows a sequential process with initial
phases in which attentional factors predominate -directed to detailed
processing of the stimulus- and later phases in which motivated actions
predominate -directed to overt behaviors of fight or flight-. The
implications of these results are discussed in the context of
traditional views on orienting and defense.

Imaging functional brain connections in stimulus processing revealed by
evoked coherence of EEG components
Hans-Juergen Volke, Boris M. Velichkovsky, Peter Dettmar, Peter Richter, 
Matthias Rudolf, and 
Torsten Klemm
Dresden University of Technology

Multichannel EEG was derived in three stimulus presentation
experiments: 1. Counting target tones within an oddball- paradigma; 2.
Solving different types of visually presented chess tasks. 3.
Categorization of written words at different levels of processing
(Craik & Lockhart, 1972). EEG records were continuously decomposed into
the "classical" components beta, alpha, theta and delta, using
time-variant spectral analysis (Volke & Dettmar, 1990).  For these
components evoked coherences were computed upon component- related post
stimulus intervals - especially from .5 to 2.5 mean wave lengths.
Results obtained suggest that stimulus processing of the brain implies
a subtle cooperation of specific areas which is highly structured in
frequency, space and time. In all three experiments, highly significant
task-specific patterns of evoked coherences emerged. As a tendency,
task-related differences of evoked coherences mostly occured within the
slower frequency bands (theta and delta), and more complex processing
modes involved more left-temporal and frontal areas of the cortex.
Generally, evoked coherences of the EEG seem to provide insights into
some new aspects of the functioning of the brain.  A basic
methodological problem will be discussed: As evoked coherences appeared
to be component-specific, the question arises which physiologically
defined components the EEG consists of, which is not answered as yet.
It will be shown that the normal EEG may be fully represented by only
four components of variable frequency and amplitude, which are in good
accordance with the classical components from delta to beta. Results
obtained on the basis of these components will be compared with those
obtained by alternative frequency resolutions.

Chronic work stress and the risk for cardiovascular disease in
sedentary males
Tanja Vrijkotte, Eco de Geus, and Lorenz van Doornen
Vrije Universiteit

Recent epidemiological research has shown that a high level of work
stress is associated with an increased risk for cardiovascular disease.
The pathophysiological mechanisms underlying this association remain
unclear. The present study tests to what extent work stress is
associated with elevations in the so-called Insulin Resistance
Syndrome, or syndrome "X" variables, i.e.
cholesterol/lipoproteinfractions, plasminogen activator inhibitor,
fibrinogen, glucose, and insulin.
From a group of 500 sedentary middle-aged males, the group of 50 men
with the
highest work stress was selected and compared to the group of 50 men
with the lowest level of work stress. Work stress was defined,
according to Siegrist and coworkers(Siegrist et al., Soc. Sci. Med.,
31, 1991), as a recurrent imbalance between high effort and low
reward.  Blood samples were taken at the beginning, in the middle and
the end of the workweek. Intra-week effects were found for glucose
only, with highest glucose levels at the start and lowest glucose
levels at the end of the week. On all three days, the high work stress
group showed higher triglyceride and lower HDL cholesterol levels than
the low work stress group. No group effect emerged on glucose, insulin
or the fibrinolytic parameters. However, subjects have only been
partially analysed (N=62) for these latter variables. Full results will
be presented at the conference. These will be discussed within the
framework of the prospective epidemiological connection between chronic
work stress and atherosclerosis and thrombosis,

Physiological indices of mental effort during a warned choice reaction
time task: A comparison between heart rate variability and corrugator
EMG activity
W. Waterink, A. van Boxtel, and I.J.T. Veldhuizen
Tilburg University

Physiological indices of mental effort may be sensitive to demands on
particular resources or may reflect aggregate demands on all
resources.  Heart rate variability (HRV) within the 0.07-0.14 Hz
frequency band and corrugator EMG activity have repeatedly been claimed
as measures with sensitiv- ity to aggregate demands.  In the current
study, these measures were compared during the performance of a
difficult warned visual choice reaction time task.  The reaction signal
was presented 4 s after a warning signal.  Repetition rate of trials
was varied across three 10-min trial blocks which were presented in a
counterbalanced order and preceded by a 5-min baseline period.  To
study the time-on-task effect, baseline period and trials blocks were
presented a second time.  Achievement motivation was manipulated by
giving 20 subjects a combination of knowledge of results and monetary
reward whereas 20 control subjects did not receive incentives.  It was
expected that working rate, time-on-task, and incentives would induce
higher compensatory mental effort reflected by larger corrugator EMG
responses and reduced HRV.  The task had a significant overall effect
on EMG activity but left HRV unaffected.  Working rate, time-on-task,
and incentives significantly facilitated EMG responses but had no
effect on HRV.  The experimental manipulations produced several
tendencies to increasing rather than decreasing HRV.  This might be
related to the relatively strong parasympathetic control of cardiac
responses during warned reaction time tasks.  We conclude that during
such tasks corrugator EMG responses better reflect the mobilization of
aggregate mental resources than does HRV.

EEG and caffeine: A comparative spectral and dimensional analysis
Paul A. Watters
Mental Health Research Institute of Victoria, Australia

Research into the observed effects of pharmaoclogically active agents
on EEG power spectra has yielded inconsistent results regarding
relative and absolute changes in power as a function of dosage. A
recent study (N=10) examined the effects of graded doses of caffeine
(100 mg, up to 600 mg) on EEG power, and on EEG dimensionality (as
determined from the application of the Grassberger-Procaccia
algorithm). Caffeine appears to have an optimising effect on caffeine
dimension, decreasing initially (up to 400 mg), then increasing
(400-600 mg). Results are interpreted in the context of both classic
theories of arousal (e.g., Yerkes-Dodson Law), and current thinking on
the dynamics of cortical activation.

Lateralized cortical activity due to preparation of saccades and finger
movements: A comparative study
Bernd Wauschkuhn, Edmund Wascher, and Rolf Verleger
Medical University of Luebeck

The topography and time course of event-related asymmetries of the EEG
associated with horizontal saccadic eye movements and finger movements
was compared in a four-choice response task, where the subjects had to
respond to the imperative stimulus (S2) by moving the right or left
index finger or by making a saccade to the right or the left.  The cue
stimulus (S1) contained full, partial, or no information about the
direction and the effector. In case of finger movements three distinct
lateralisations were found: 1) increased negativity over the motor
cortex contralateral to the future movement direction, 2) increased
contralateral negativity at temporo-parietal sites beginning 200 after
delivery of the directionalinformation and 3) increased ipsilateral
negativity at temporo-parietal sites beginning 350-500 ms after
delivered direction and effector information. The Early Temporo-
Parietal Lateralisation was also visible in case of saccadic eye
movements and in case of effector-unspecific directional information.
Before saccadic eye movements no other distinct lateralisation could be
observed at any recording site. In sum, lateralised cortical activities
due to preparation processes for finger movements and due to
effector-unspecific processing of directional information for motor
preparation by the posterior parietal cortex could be demonstrated,
whereas no distinct lateralisation due to preparation for saccadic eye
movements was visible.

Respiratory sinus arrhythmia from child- to adulthood; What happens in
E.J.M. Weber, R.J.M. Somsen, and M.W. Van der Molen
University of Amsterdam

Developmental changes in respiratory sinus arrhythmia were examined in
two experiments in which heart rate and respiration were recorded from
subjects in rest and attention demanding conditions. Three age groups
participated in the first experiment; 5-, 7-, and 9-year olds. These
children performed standard oddball and reaction time tasks. The
results from this experiment revealed a depression of spectral power in
the RSA band during task performance relative to baseline. The
reduction in RSA did not discriminate between age groups. In the second
experiment, the age range was expanded by recording heart rate and
respiration from two groups of children, 5-, 12-year olds, and adult
volunteers. Measurements were taken in rest and during a habituation
task. The results were similar to the data from the first experiment in
showing a reduction of spectral power in the RSA band during task
performance relative to rest. In contrast to the first experiment,
however, RSA decreased significantly with age. In addition, RSA peak
amplitude shifted to lower frequencies with age. Interestingly, RSA
peak frequency corresponded to respiratory frequency for the older
children and adults but not for the youngest age groups. Subsequent
coherency analysis yielded lower values for the 5-year olds compared to
the older age groups.  These findings may suggest that the
respiratory-heart rate coupling does not reach mature levels during
childhood. This finding might have important implications for the use
of RSA as a noninvasive tool to examine the status of the nervous
system in clinical groups.

Prepulse inhibition and habituation of skin conductance responses in
schizophrenics: Neuroleptic drug effects
Almut Weike1, Jutta Globisch1, Alfons Hamm1, and Ulrike Bauer2
1University of Greifswald, 2University of Giessen

Startle response magnitude is substantially reduced when a
non-startling stimulus (prepulse) shortly precedes the startle
eliciting stimulus. This so called prepulse inhibition is impaired in
schizophrenic patients. The present study was designed to replicate and
extend the findings, focussing on the influence of neuroleptic
medication on this process. Moreover, it was investigated, how
habituation of the electrodermal orienting activity was influenced by
different drugs. Thirty-six schizophrenic patients (12 unmedicated)
participated in the experiment. Twelve healthy controls were recruited
from hospital and laboratory staff. The prepulse was a 85 dB(A) 1000 Hz
tone (20 ms duration), which preceded the startle stimulus (50 ms burst
of 105 dB(A) white noise) by 30, 60, 120, or 240 msec.  As expected,
while control subjects showed a clear reduction in startle magnitude
following prepulse presentation, prepulse inhibition was significantly
impaired in unmedicated schizophrenics. Unmedicated patients also
showed significantly less habituation of skin conductance responses. By
contrast, medicated patients showed prepulse inhibition and
electrodermal response habituation comparable to that shown by the
controls. This medication effect was also apparent when some of the
patients from the unmedicated condition were tested after receiving
neuroleptic medication. Once medicated, these patients showed larger
prepulse inhibition and increased habituation of electrodermal

Role of spatial abilities in motion sickness susceptibility
S.E. Weinstein, E.R. Muth, J.T. Andre, E. Jarret, R.S. Stern, H.W. 
Leibowitz, and W.J. Ray.
Pennsylvania State University

The primary objective of this study was to determine whether
motion sickness susceptibility is related to individual
differences in spatial abilities. Motion sickness susceptibility was
determined by exposing 45 subjects to a rotating optokinetic drum and
recording their subjective symptoms.  Electrogastrograms (EGGs) and
cardiac interbeat intervals (IBIs) were also recorded.  During a second
session, subjects were tested on three types of spatial tasks; the
water level test, a mental rotation test, and the rod and frame test.
No systematic relationship was found between mental rotation or the rod
and frame task and symptoms of motion sickness.  However, subjects
susceptible to motion sickness did significantly worse on the water
level test than those who were not susceptible (p < .01).  EGG results
indicated that susceptible subjects showed a significantly greater
increase in percentage of gastric tachyarrhythmia, abnormal gastric
myoelectric activity usually associated with nausea, from baseline to
drum rotation than did the non-susceptible subjects.  Vagal activity,
as estimated from variability in the IBIs using the algorithm of Porges
and Bohrer (1990), showed a significantly greater decrease from
baseline to drum rotation for susceptible subjects compared to
non-susceptibles (p < .05).  In conclusion, subjects who performed
poorly on the water level task, which has been considered a measure of
an internal sense of horizontality, were also more susceptible to
vection induced motion sickness.

Cardiovascular patterns associated with threat and challenge
appraisals: Individual responses across time
S. Weinstein1, K.S. Quigley1, and L. Feldman Barrett2
1Pennsylvania State University and 2Boston College

The purpose of this study was to examine individual cardiovascular
patterns associated with cognitive appraisals over four arithmetic
tasks.  Previous studies demonstrated distinct patterns associated with
threat and challenge appraisals across groups of subjects.  Threat
appraisals (high stress; low coping ability) are associated with modest
increases in heart rate (HR), cardiac output (CO), and pre-ejection
period (PEP), and increases in total peripheral resistance (TPR).
Challenge appraisals (mild to moderate stress; high coping ability) are
associated with larger increases in HR, CO, and PEP, and decreases in
TPR (Tomaka et al., 1993).   In the current study, we extend these
results by examining threat and challenge appraisals within individuals
across time.  Subjects completed four verbal arithmetic tasks with
ratings of stressfulness and coping ability made before and after each
task.  Appraisal ratios were computed by dividing stressfulness ratings
by coping ratings.  The relationships between changes in individuals'
appraisals and cardiovascular patterns over the four task epochs were
analyzed using hierarchical linear modeling (HLM) which permits
separation of between- and within-subjects variance.  Change scores
from baseline for the final minute of each task were computed for PEP,
TPR, interbeat interval (IBI; inverse of HR), and CO.  Post-task threat
appraisals were associated with smaller decreases in PEP (p < .02), and
larger increases in TPR (p < .05).  Changes in IBI and CO were
nonsignificant.  These data extend previous results for PEP and TPR
using individual-level analyses of task appraisals across four tasks,
and demonstrate the use of HLM for psychophysiological studies.

EEG measures of differential brain activity: Before, during, and after
the perception of apparent as opposed to actual movement
Erik D. Welch, John Lagomarsino, and John M. Morgan 
Humboldt State University

The Autokinetic effect is a visual phenomenon of illusory movement,
experienced when a small, stationary spot of light is seen to move
across a contrasting and patternless background.  Outflow theory states
that autokinesis is due to a mismatch between the retinal image signal
and the eye motion, outflow, signal.  It is theorized that involuntary
directional shifts of the eye are not monitored by the brain.
Voluntary motor commands generate sensory expectations of movement of
the visual field which are efference copies or corollary discharges
This expectation cancels out the afferent signals which result when the
visual field is moved However proprioceptors, which signal the amount
of stretch in muscle, have been found in human extra-ocular muscles
which weakens the outflow theory.
	Event Related Potentials (ERPs) were recorded from four
conditions: darkness (no visual stimulus), autokinesis (illusory
movement), horizontal movement (2 cm/sec), and vertical movement (2
cm/sec).  These conditions were presented in blocks of ten, in random
order 6 times each, for a total of 240 trials per subject.
	A P3 component was found in all of the conditions, along with
	the Bereitschaftspotential (BP) and Motor Potential (MP).
Differences were found however, in the latency to the subjects
response, suggesting subjects confidence level acting as a confound in
the amplitude of the components.  Differences were found between the
darkness condition and all other conditions, supporting the validity of
Outflow theory.  Results are discussed in reference to relevant
perceptual and ERP theories.

The effects of stress and muscle activity on P50 suppression
Patricia M. White, Cindy M. Yee, Maria Nazarian, Halle Jones, and 
Valerie Gilman
University of California, Los Angeles

The ratio of P50 amplitude to paired clicks has been suggested as a
preattentional gating response that is impaired in schizophrenia.
Previous work assessed the relative effects of attention and anxiety on
the P50 suppression ratio in unselected controls, suggesting that P50
suppression is unaffected by attentional manipulation but diminished
during an oral mental arithmetic stressor.  In this study, several
factors which might explain reductions in P50 suppression by the oral
mental arithmetic stressor were considered.
Ten tasks which manipulated muscle activity, auditory competition,
cognitive demand, and levels of effort and anxiety were administered
while subjects were presented with paired clicks. Preliminary results
suggest that reductions in P50 suppression during performance of the
oral mental arithmetic stressor are unlikely to result from either
competing cognitive or auditory tasks. Competing cognitive tasks
performed without muscle activity showed no effect on P50 suppression
while competing auditory tasks reduced P50 amplitude but did not
influence the P50 suppression ratio.  Tasks which varied muscle
response were associated with an enlarged P50 suppression ratio
although diminished P50 suppression did not occur in all subjects.
Within tasks in which effort was manipulated, self-ratings of effort
correlated positively with the P50 suppression ratio. Results of tasks
rated as anxiety-provoking suggest that anxiety also may reduce P50
suppression.  The post-auricular reflex and heart rate were found to
correlate with P50 suppression and with self-ratings of effort and
anxiety.  These findings are discussed in relationship to P50
suppression research on schizophrenia.

Heartbeat detection and the experience of emotions
Stefan Wiens, Elizabeth Mezzacappa, Stephen Palmer, Anil Chacko, Marc 
Lingat, and Edward S. 
State University of New York at Stony Brook

Self-perception of physiological activity has been postulated to play
a crucial role in the experience and expression of emotion (James,
1884, Mind, 188-205; Schachter & Singer, 1962, Psychological Review,
379- 399).  However, there is scant empirical work on the relation
between self-perception of visceral activity and the expression and
experience of emotion (Ferguson & Katkin, 1996, Biological Psychology,
131-145; Katkin, 1985, Psychophysiology, 125-137).  In the present
study, 52 undergraduates (19 males) were classified as good (n = 10) or
poor (n = 42) heartbeat detectors using a modified Whitehead
discrimination procedure (S+ = 200, S- = 500).  Subjects were then
presented with two video clips representing each of three different
emotional valences (amusement, anger, fear). After each of the six
clips, self-reported affective responses to the clips were measured on
two orthogonal 9-point scales indicating arousal level and valence,
respectively.  Electrodermal activity was measured at baseline and
during each video clip.  It was hypothesized that good heartbeat
detectors would experience more intense emotional arousal than poor
heartbeat detectors, irrespective of actual arousal level, but would
not differ in ratings of the valence of the experienced emotions.  As
predicted, good detectors reported higher levels of arousal than poor
detectors across all three emotional valences, F(1,46) = 8.89, p < .01,
using electrodermal activity for each clip as a covariate to control
for physiological arousal level. No differences were found on valence
ratings between the two groups on any of the emotions. These results
support the view that heartbeat detection plays a role in the
experience of emotional intensity.

Ambulatory assessment of self-report, autonomic, and respiratory
responses during phobic anxiety
Frank H. Wilhelm and Walton T. Roth
Stanford University School of Medicine and VAPA Health Care System, Palo 

Although physiological measures are recognized as theoretically
important in assessing anxiety, their practical utility has been
questionable.  We undertook to evaluate physiological indices of
anxiety and to test specific anxiety theories in 14 flight phobics and
15 sex- and age-matched controls who were willing to take a short
commercial flight.
Subjects wore a light, compact monitor that recorded
cardiovascular, electrodermal, and respiratory activity. They
rated themselves on a battery of self-reports of mood and symptoms
before, during, and after the flight.  The benefits of baseline
adjustment and transformation for all variables, and the adjustment of
heart rate by ventilation to give "additional heart rate" were
The variables that best distinguished the groups during flight
were self-reported anxiety and additional heart rate.  The first
classified subjects with an accuracy of 90% and the second with an
accuracy of 79%.  Although respiratory rate and minute volume,
indicators of hyperventilation, did not distinguish groups, phobics
paused during inspiration more than controls.  Skin conductance
fluctuations were increased and respiratory sinus arrhythmia was
decreased in the phobics.  Awareness of physiological changes was
indicated by significant correlations between skin conductance level
and reported sweating, but not between respiratory measures and
reported shortness of breath or between heart rate and reported heart
pounding or racing.
Our results demonstrate that ambulatory monitoring of multiple
physiological systems is feasible in phobic situations outside the
laboratory.  Several measures of heart rate proved to be powerful
indices of anxiety.  Specific breathing irregularities characterized
phobics when anxious.

Additional heart rate: Application and validation under ambulatory
Frank H. Wilhelm and Walton T. Roth
Stanford University and VAPA Health Care System, Palo Alto

Physical activity produces physiological activation like that
produced by anxiety.  One way of avoiding confusion is to exclude
epochs containing physical activity.  Alternatively, the anxiety
component of heart rate (HR) can be calculated as the HR above that
predicted by oxygen consumption. Our innovation in this study was to
substitute minute ventilation (Vmin) for oxygen consumption in an
ambulatory setting, calculating additional heart rate (HRadd) from the
relation between Vmin and HR during an exercise test.
26 participants with flying phobia and 15 nonanxious controls were
physiologically monitored while bicycling under 3 different loads,
while walking (leaving the hospital, entering a plane), and during
flight.  Individual regression equations relating HR to Vmin were
derived from the bicycling (mean R square=.80).  Vmin did not differ
between groups at any point.  Unadjusted HR was not significantly
different between phobic participants and controls leaving the hospital
(119/114 bpm, p>.38) or entering the plane (117/110 bpm, p>.25).
However, although HRadd was not different leaving the hospital (7.9/8.6
bpm, p>.75), it was entering the plane (18.0/9.9 bpm, p=.01),
reflecting the greater anxiety in the phobic participants.  Even though
physical activity was minimal during flight, HRadd was also superior to
HR for indexing group differences there.  In general, our results
suggest that this methodology could be useful in assessing

Psychophysiological correlates of multiple task performance
Glenn F. Wilson1, Lisa Fournier2, and Carolyne R. Swain3
1Armstrong Laboratory, 2Washington State University, 3Logicon Technical 
Services, Inc.

Since multi-task paradigms approximate real world environments, the
nature of psychophysiological responses to the cognitive demands of
complex operations is of both theoretical and applied importance. By
the same token, multiple psychophysiological measures provide an
exceptional means for delineating the relationships between multi-task
demands and performance.  In this study, the Multi-Attribute Task
Battery was used to generate single and multi-task scenarios.  An
auditory communication task served as the single task and was combined
with visual monitoring and tracking to provide a multi-task with four
levels of difficulty.  Performance and subjective workload ratings as
well as ECG, EEG, EOG and respiration were recorded from ten subjects
(4 male, 6 female).  EEGs recorded from 58 active sites were corrected
for eye artifacts and submitted to current source density and FFTs
procedures.  The relative power in five bands were statistically
compared.  Heart rate (HR), heart rate variability (HRV), eye blink
rate (BR) and respiration rate (RR) were also evaluated.  Performance
and workload ratings evidenced the predicted changes due to number of
tasks and difficulty.  HR increased and HRV decreased significantly
between single and multiple tasks and in response to the most difficult
condition. BR decreased and RR increased significantly during
multi-task conditions. With regard to the EEG, multi-task conditions
resulted in increased theta (central and parietal) and beta (frontal)
and decreased alpha (parietal and occipital) compared to single task
conditions. Results are discussed in terms of the sensitivity of
psychophysiological measures to different aspects of mental workload.

Effects of tone-cued fear and pleasant imagery on reaction times to
early probes, heart rate, and skin conductance
Charlotte vanOyen Witvliet, Jason R. Robinson, Georgia Panayiotou, and 
Scott R. Vrana
Purdue University

This study examined the effect of imagery cues on reaction times
(RTs) to auditory probes.  Imagery effects on heart rate and skin
conductance were also assessed.  In each of 6 blocks, participants (24
males, 22 female) imagined themselves in fear and pleasant relaxation
situations for 6-second trials when cued by high or low tones.  Imagery
trials were separated by 3 to 6 six-sec intervals of a "think 'one'"
control condition cued by medium frequency tones.  Materials were
counterbalanced across subjects.  For the secondary task, RT probes
(100 ms of 64 dB white noise) were presented 120 ms, 500 ms, or 2000 ms
after tone (600 ms, 69 dB) onset.  Participants were instructed to
press a button with their dominant thumb as soon as they heard the RT
probe and immediately resume their primary task.  Overall, as RTs were
probed further away from tone onset, RTs were significantly faster.
When probes occurred 2000 ms after tone onset, imagery and "think
'one'" control RTs did not differ.  However, at 120 and 500 ms, RTs
were significantly slower for probes that occurred during fear and
pleasant imagery tones than during control tones, suggesting that
imagery cues demanded more processing resources than the control
condition cue.  At 120 ms, RTs were significantly slower during fear
than pleasant imagery tones, indicating that fear cues demanded more
processing capacity at this early interval.  Paralleling previous
studies, heart rate was significantly faster during fear, and skin
conductance tended to be higher during fear than pleasant conditions.

The emotional impact of instrumental music on affect ratings, facial
EMG, autonomic measures, and the startle reflex: effects of valence and
Charlotte van Oyen Witvliet and Scott R. Vrana
Purdue University

Participants (31 males, 31 females) listened to prerecorded
instrumental music that varied in emotional valence and arousal.  Music
was selected based on ratings by an independent sample.  Each subject
first listened to and rated 16 pieces of music in terms of valence and
arousal.  Then the subject listened to 51 counterbalanced trials.  Each
trial consisted of 26 sec of music followed by 10 sec during which the
subject was instructed to continue to experience the emotion expressed
in the music.  A tone then signaled the subject to relax and "think
'one'" to clear the emotion and music from the subject's mind.  After
the last trial, subjects re-rated the music.  Facial EMG, heart rate,
and skin conductance were measured during and after the music.  The
acoustic startle reflex (for 25 males, 23 females) was probed at 3.5,
5, or 6.5 sec after music offset.  Participant ratings clearly
differentiated the negative and positive as well as the low and high
arousal materials both at the beginning and end of the experiment.
Corrugator EMG was higher during negative music and zygomatic EMG was
higher during positive music.  Orbicularis oculi EMG, heart rate, and
skin conductance were greater during high arousal music.  The acoustic
startle magnitudes in this music paradigm did not match the patterns
typically obtained in emotion research.  During imagery, high arousal
facilitates the startle.  However, in this study, startle magnitudes
were significantly larger for low compared to high arousal affect
conditions, and valence did not reliably influence startle magnitudes.

Emotional imagery and the visual startle reflex: Negative valence and
high arousal independently increase magnitudes
Charlotte van Oyen Witvliet and Scott R. Vrana
Purdue University

This study used visual (light flash) probes to prompt the startle
reflex during emotional imagery that systematically varied in affective
valence and arousal.  Participants (24 males, 22 female) completed 6
blocks of tone-cued imagery with their eyes closed.  In each block,
they imagined themselves in two different emotional situations during
8-second trials when cued by either a high or low tone.  Imagery trials
were separated by 3 to 6 eight-sec periods of tone-cued relaxation.
The results replicate findings obtained with an acoustic startle probe
(Witvliet & Vrana, Psychophysiology, 1995).  Here, visually-probed
startle magnitudes were significantly larger both during negative
(fear, sadness) compared to positive (joy, pleasant relaxation) imagery
and during high arousal (fear, joy) compared to low arousal (sadness,
pleasant relaxation) imagery.  The influence of affective arousal was
even more potent than valence in modulating startle magnitudes.
Participants were also asked to estimate how often the visual startle
(flash) occurred during each type of imagery (fear, sadness, joy,
pleasant relaxation).  Paralleling startle results, participants
estimated that the flash occurred significantly more often during
negative versus positive and high arousal versus low arousal imagery,
even though all types of imagery were probed with the light flash
equally often.  Consistent with previous findings, heart rate was
significantly faster during high compared to low arousal imagery.

Heart rate variability in PTSD: Mid-frequency power is inversely
related to trauma severity and nightmares
Steven H. Woodward and M. Michele Murburg
National Center for PTSD

We measured heart rate variability (HRV) during baseline sleep in 46
combat-related posttraumatic stress disorder patients and eight
combat-exposed controls.  All subjects were unmedicated and, in
particular, were free of beta-blockers and related hypertension and
cardiac medications.  Three nights of continuous ECG were continuously
digitized at 125 Hz and heart period spectra calculated from demeaned,
detrended and windowed instantaneous heart period time series.  The
latter were 30 seconds in duration and corresponded to manually-scored
sleep epochs.  Consistent with earlier reports, REM sleep was observed
to be associated with elevated power at low and mid-frequencies and
reduced power in the high frequency band.  HRV at all frequencies
tended to be lower in PTSD patients as compared to combat controls, but
these differences did not achieve significance.  HRV was independent of
subjective sleep quality and sleep continuity.  In contrast, low and
mid-frequency heart rate variability during non-REM sleep was reduced
in patients reporting 1) more combat exposure (Combat Exposure Scale;
Keane et al, 1989)  (Pearson r's = -0.299 to -0.381, p < 0.04 to
0.009), and 2) more nightmares (assessed by combining the two
nightmare-related items from the Mississippi Scale of Combat-Related
PTSD, Keane et al, 1988; Pearson r = -0.379 to -0.391, p < 0.009 to
0.007).  Within the control group, a similar relationship was observed
between and HRV and trauma exposure (Spearman r = -0.81, p < 0.05), but
not nightmare symptomology.  These data suggest that heart rate
variability offers a valuable perspective on biological adaptations to
extreme stress in humans.  More specifically, they suggest that trauma
exposure and nightmare symptomology may together be associated with
long-term modifications of sympathoneural cardiac regulation.

Effects of prehabituation of the prepulse on startle eyeblink
Jonathan K. Wynn 1, Anne M. Schell 1, and Michael E. Dawson 2
1 Occidental College, 2 University of Southern California

It is generally believed that prepulse inhibition of the startle
eyeblink in passive attention tasks, where the subject is not
instructed to attend to the prepulses, is a truly automatic process. If
this is so, then prepulse inhibition should not be affected by
prehabituation of the prepulse, which reduces orienting to the
prepulse. Thirty-nine college student participants were randomly
assigned to two groups, one which received 30 presentations of a 5 s
tone stimulus and one which received 3 presentations of the tone
intermixed with 27 presentations of a light.  Immediately following
these trials, the tone was used as a prepulse in a startle  blink
modificationImmediately following these trials, the tone was used as a
prepulse in a startle blink modification noninstructed attention
paradigm with lead intervals of 60, 120, and 2000 ms, using an auditory
startle stimulus. Prehabituation was highly effective in reducing the
skin conductance OR to the tone by the beginning of the startle blink
modification trial series. Significant prepulse inhibition was observed
in both groups at the 60 and 120 ms probe positions, and significant
facilitation was observed in both groups at the 2000 ms  probe
position. However, the prehabituation group and the no-prehabituation
group did not differ in either the amount of prepulse inhibition or
later facilitation that was seen. These results suggest that in a
passive attention paradigm, both prepulse inhibition and long lead
interval facilitation  reflect the action of automatic (preattentional)
as opposed to controlled processes.

Effects of attention on P50 gating in schizophrenia
Cindy M. Yee, Patricia M. White, and Keith H. Nuechterlein
University of California at Los Angeles

Abnormalities in attention and sensory perception have characterized
models of schizophrenia since its first clinical description.  Recent
advances suggest that the cognitive fragmentation and stimulus overload
observed in schizophrenia may be related to an inability to filter or
gate information.  Freedman et al. (1987, Schizophrenia Bulletin, 13,
669-678) demonstrated that in normal subjects, the P50 component of the
auditory ERP to a second click shows gating or suppression relative to
the response elicited by the first click when the two audio clicks are
delivered 500 ms apart.  In contrast, schizophrenic patients often fail
to exhibit P50 suppression, suggesting a loss of normal inhibition or
gating.  For gating to be inferred accurately, changes in P50 amplitude
must result from inhibition rather than from fluctuating attention.  To
examine such a possibility, Jerger, Biggins, & Fein (1992, Biological
Psychiatry, 31, 365-377) manipulated the salience of the second click
and found that while N100 was profoundly influenced by attention,
neither P50 amplitude nor its suppression were affected.  The present
study adapted the paradigm developed by Jerger et al. to examine the
effects of attention on P50 suppression in schizophrenic patients.
N100 amplitude followed the predicted pattern with a diminished
response to the second click when attention was directed to the first
click but an enhanced response to the second click when attention was
directed there.  Analyses of P50 suppression data suggest that a subset
of patients may be responsive to attentional manipulations depending
upon sub-classification of patient diagnosis (i.e., paranoid,
disorganized, or undifferentiated schizophrenia).

The role of psychophysiological traits in the process of the
biobehavioral therapy in patients with labile and stable hypertension
V.V.Zakharova, E.Sokhadze, O.Trofimov, and S.Kasjanova

Siberian Branch of Russian Academy of Medical Sciences

The goal of this study was the specification of the role of personal
psychophysiological  traits in the pathogenesis of the stages and forms 
of development of hypertension. 
Biobehavioral therapy with temperature and EMG biofeedback (BFB) was
administered to 42 hypertensive patients.  Before and after the course
of BFB they were tested by Mini-Mult inventory, Cook-Medley hostility
scale  and Spilberger's test of anxiety.
Patients with labile hypertension (high systolic blood
pressure with elevated stroke volume and heart rate and normal
peripheral resistance) were characterized by higher tone of the somatic
muscles and normal skin temperature. EMG BFB training was more
effective as compared to temperature, and was accompanied  by
decreased anxiety and fear.
Patients with stable hypertension (high diastolic blood pressure with
increased total peripheral resistance with normal or even decreased
cardiac output) demonstrated comparatively lower baseline temperature
of the skin with mild EMG tone. Successful temperature BFB training led
to decrease of diastolic blood pressure, and also lowered aggressiveness 
and hostility.
The results of this study show that peculiarities of the personal
profiles determine variations in psychophysiological reactivity and
affect development of hypertension.  Perfectly selected
biobehavioral therapy enhances correction of the pathological changes
and prevents their further progress.