1997 SPR Meeting Poster Abstracts: A - L

Differences between ERP conditions: Different sources or different activity of the same sources?
Andre Achim1 & Gilles Plourde2
1University of Quebec-Montreal, 2McGill University
A difference between spatio-temporal event-related potentials (ST-ERPs)
across conditions could reflect different sets of activated sources
or simply different responses from the same sources. To establish
how many principal components (PCs) are required to explain each ST-ERP,
a statistical signal detection test (e.g., method CP1 form; Achim,
1995) may be applied to the individual cases composing it, removing
PCs one at the time until no signal remains detected. The following
is proposed to test whether two or more conditions may be considered
as simply differing in responses from the same N sources: derive spatial
PCs as if the conditions were attached end to end in time, select
the first N common PCs, filter out those topographies from all data
matrices, and finally apply the signal detection test on each condition
separately. If the data represent no more than N active sources, no
residual signal should be detected. This technique is applied to 40Hz
auditory steady-state ERPs from seven channels in the coronal plane,
obtained from ten healthy volunteers under four dosages on isoflurane
anaesthesia (0, .28%, .38%, .50% end-tidal), with nine replicated
averages per condition. The results show that some subjects differ
significantly from one another, even at baseline, and that in many
subjects the topographies required to explain the signal under awake
and anaesthetised conditions are not the same. Simulation studies
with comparable synthetic data demonstrate that the method does not
exceed its nominal false alarms rate.

Evaluation of mental concentration under different lighting conditions using eye fixation related potential
Yukio Akashi1, Chie Umeno2, & Akihiro Yagi2
1Matsushita Electric Industrial Co., Ltd., 2Kwansei Gakuin University
We can obtain a specific ERP called the eye-fixation-related-potential
(EFFORT) with averaging EEG time locked to onset of eye fixations.
Using the EFFORT as a physiological index, two experiments were carried
out to evaluate the mental concentration of subjects under different
lighting conditions. These experiments were based on the hypothesis
that the varieties of lighting in time and space would increase the
mental concentration. In the first experiment, the ratio of the ambient
illuminance to the task illuminance was varied from 0.02 to 1.0, while
the task illuminance was kept at 1000lx. Under each lighting condition,
each subject performed a searching task for an hour. During the task,
the EFFORT of each subject was recorded. Fourteen subjects aged 20-30
participated in the experiment. The results showed that there was
a close correlation between the subjective appraisals and the measured
EFFORT, and that the optimum concentration was obtained under the
lighting conditions where the ambient / task illuminance ratio was
0.02-0.3. In the second experiment, the illuminance level of the task
was changed from 750 to 1500 lx every five minutes in three manners.
Under each lighting condition, each subject performed the searching
task for 20 minutes. During the task, the EFFORT of each subject was
recorded. Eleven subjects aged 21-36 participated in the experiment.
The results of the experiments showed that the amplitude of EFFORT
was correlated to the subjective appraisal, and that the change in
lighting level had a positive affect on the mental concentration of

Manipulation of frontal EEG asymmetry alters self-reported affect and facial EMG
John J.B. Allen1, Eddie Harmon-Jones2, & James H. Cavender1
1University of Arizona, 2University of Wisconsin-Madison
Resting individual differences in frontal EEG asymmetry have been
shown to predict emotional responses across a variety of contexts.
Although it has been shown that frontal asymmetry can be manipulated
through biofeedback training, it remains to be determined whether
such manipulation will alter subsequent emotional responses. We therefore
investigated emotional responses before and after biofeedback training
in 18 right-handed female college students, nine randomly assigned
to receive biofeedback training designed to increase right frontal
alpha relative to left frontal alpha (UP), and 9 assigned to receive
training in the opposite direction (DOWN). Five consecutive days of
biofeedback training provided signals of reward (300 Hz tones) or
nonreward (150 Hz tones), depending on whether the difference between
right (F4) and left (F3) frontal alpha exceeded a criterion value.
Asymmetry scores (F4-F3) varied as a function of training, with UP
subjects demonstrating significantly (p < .05) greater scores than
DOWN subjects during the third and fourth days of training. Self-reported
affect in response to film clips showed significant training-related
changes, with UP subjects reporting significantly greater positive
affect (interest, happiness, and amusement) after training than did
DOWN subjects. Facial EMG from zygomatic and corrugator muscle regions
corroborated the self-report findings, with DOWN subjects showing
a significant decrease in zygomatic (but not corrugator) activity
from before to after training, and UP subjects showing a significant
decrease in corrugator (but not zygomatic) activity. It remains to
be determined whether these effects of training will persist outside
the laboratory environment.

When it is and when it is not: The heritability of frontal EEG asymmetry
John J.B. Allen1, Jennifer Reiner1, Joanna Katsanis1, & William G. Iacono2
1University of Arizona, 2University of Minnesota
Individual differences in resting frontal EEG asymmetry have been
shown to be related to differences in temperament and to predict emotional
responses across a variety of contexts. Although two abstracts in
recent years have suggested that frontal EEG asymmetry may be, in
part, heritable, these studies used a family study design, in which
both environmental and genetic influences can contribute to the child-parent
similarity. We therefore used a twin design to investigate the extent
to which resting frontal EEG asymmetry is heritable. Resting EEG was
collected with eyes closed for 5 minutes from 30 pair of MZ and 30
pair of DZ twins, all 17 year- old females. After rejection of ocular
and other artifacts, alpha power (8-13 Hz) was extracted via an FFT
with a Hamming window. The standard asymmetry score was calculated:
Ln(F4)- Ln(F3). Internal consistency reliability, based on treating
each of the individual minutes as a separate "item," indicated that
this frontal asymmetry score had a coefficient alpha internal- consistency
reliability estimate of .97. Across all five minutes, Falconer's estimate
of heritability was a modest .33 . When examined by minute, however,
heritability varied systematically such that heritability was very
low during the first two minutes (.11 & .13), was modest during the
next two minutes (.42 & .44), and was extremely high during the final
minute (.97). These findings suggest that the immediate environmental
context influences the extent to which environmental or genetic influences
predominate in determining resting frontal EEG asymmetry.

Music, cardiovascular reactivity, and cognitive task performance among college students: A focus on mathematics, psychology, chemistry, and music majors
Karen Allen
State University of New York at Buffalo
Recent studies have demonstrated a role for music in reducing blood
pressure (BP) and heart rate (HR) among surgical patients and surgeons.
Other research has focused on the neurobiology of learning and has
demonstrated that listening to a Mozart sonata can improve spatial
reasoning skills. The current study extends these two areas of inquiry
and considers the effects of three experimental conditions on HR,
BP, and task performance in response to two cognitive stressors (serial
subtraction and completion of GRE questions.) Participants were 120
senior-level undergraduates majoring in mathematics, music, chemistry,
or psychology (n = 30 in each major). Experimental conditions were:
(1) participant-selected music, (2) experimenter-selected music (Mozar
Sonata for Two Pianos in D Major), and (3) no music. The study design
was within-subjects and conditions were counterbalanced. ANOVA analyses
revealed main effects (p< .01) for all within-subject variables (music
condition, task, and rest vs. task period) as well as music condition
by task and academic field. Psychology and music students had responses
similar to each other, i.e., best performance and lowest reactivity
when listening to the music they chose themselves. Math and chemistry
students had responses similar to each other that followed another
pattern: best performance and lowest reactivity when listening to
the Mozart selection. For all students the condition with no music
was associated with worst performance and highest reactivity. This
study demonstrates evidence that music can be related to improved
autonomic responses and performance during stressful tasks, and suggests
that, for some individuals, control over selection of music is an
important factor.

The D4DR gene and platelet MAO are markers for human fear-conditioning
Peter Annas, Hakan Garpenstrand, Jonas Ekblom, Lars Oreland, and Mats Fredrikson
Uppsala University
Fear-conditioning, a basic form of associative learning, seems partly
modulated by the dopaminergic and serotonergic systems in animals.
To examine if these systems also regulate fear-conditioning in humans
we studied polymorphisms in the dopamine 4 receptor gene (D4DR) and
platelet monoamine oxidase (trbc MAO) in subjects showing good (n=20)
or poor (n=20) acquisition of fear-conditioning. The experimental
procedure consisted of three phases. First, during habituation 2 slides
were presented for 8 seconds on the screen 2 times each. Second, during
acquisition each slide was presented 8 times. At stimuli offset one
but not the other slide was paired with an electric shock. Third,
during extinction slides were presented 8 times each with shocks withheld.
The sample was divided into one group with only short D4DR repeat
sequences (2-5 repeats) and another with long D4DR repeat sequences
(6-8 repeats) present. Using a median split approach, the sample was
also divided into one group high and one low in trbc MAO activity.
Subjects low as compared to those high in trbc MAO activity showed
better learning during acquisition and also retained learning better
during extinction, t(35)=2.1, P<.05 and t(35)=2.0, P<.05respectively.
Subjects with long D4DR repeats showed better retention during extinction
than those with short t(38)=2.3, P<.05. We conclude that the dopaminergic
system and serotonergic capacity are markers for fear-conditioning
in humans.

Two studies examining the relationship between social support and cardiovascular reactivit
Jennifer L. Anthony & William H. Obrien
Bowling Green State University
Researchers have suggested that social support may operate at a psychophysiological
level, as a moderator between environmental stressors and cardiovascular
reactivity (CVR), which has been linked to the development of coronary
heart disease and hypertension. Recently, experimental studies have
examined the relationship between social support and CVR in laboratory
manipulations. The purpose of these two studies was to further evaluate
this relationship. In the first study, sixty-seven subjects were randomly
assigned to one of three conditions: alone, supportive, and nonsupportive
conditions. In the second study sixty female subjects were randomly
assigned to one of three conditions: nonsupportive, low supportive,
and high supportive. The second study incorporated methodological
modifications which included: reducing the evaluative nature of confederates
and incorporating a new condition to include higher levels of support.
It was hypothesized in both studies that the more supportive the condition,
the less reactivity subjects would experience. A speech was utilized
as the stressor in both studies and cardiovascular reactivity was
examined using cardiac impedance and blood pressure. Results for the
first study showed that no significant differences in cardiovascular
reactivity were found between the supportive and nonsupportive conditions.
For the second study, a significant difference was found on one measure
between the low supportive condition and the high supportive and nonsupportive
conditions, with the low-support condition exhibiting significantly
less cardiovascular reactivity. Analyses comparing the results of
the two studies were conducted. The findings for both studies are
discussed in terms of evaluation apprehension and methodological considerations.

Resting autonomic control of the heart is related to 24-hour blood pressure variability
K.M. Ashley, D.W. Bytnar, M.S. Friederich, R.S.Gard, J.A. Kozisek, M.R. Rist, T.C. Sanford, J.J. Sollers, A.C. Stratman, & J.F. Thayer
University of Missouri-Columbia
Recent studies have implicated blood pressure (BP) variability as
a risk factor for heart disease. Sloan et. al.(1997) have shown that
short term heart period (HP) variability was inversely related to
short term diastolic BP variability during laboratory tasks. However,
to date, no studies have examined the relationship between resting
autonomic control of the heart in the lab and measures of long term
BP variability. Eight participants (6 female; 2 male) underwent a
10-minute laboratory assessment of heart period variability and wore
an ambulatory BP monitor for 24-hours. Pearson correlations assessed
the relationship between time and frequency domain measures of HP
variability, and the mean and standard deviation of the 24-hour B
recordings. As evidence of convergent validity, HP standard deviation
in the lab was significantly correlated with the standard deviation
of the 24-hour ambulatory heart rate (r(8)=.74, p<.05). Importantly,
the mean HP in the lab and a measure of vagally mediated HP variability,
the number of interbeat interval differences greater than 50 milliseconds
(%BB50), were inversely related to the standard deviation of the 24-hour
diastolic BP, (r(8)= -.83, p<.05 and r(8)= -.76, p<.05, respectively).
These results suggest that decreases in vagally mediated HP control
are inversely related to long term diastolic BP variability. These
findings could have an impact on our understanding of the etiology
of cardiovascular disease.

Sensorimotor gating: Can a prepulse inhibit the inhibition of a startle blink?
Bridgette C. Avery1, Michael C. Avery2, & E. Evan Krauter1
1The University of South Carolina-Spartanburg, 2The Timken Company
A weak stimulus presented just prior to a startle-eliciting stimulus
can strongly inhibit the auditory startle blink. One suggestion is
that this prepulse inhibition reflects a mechanism that gates out
incoming stimuli. We tested that notion in four studies (N=87). In
each study, participants received trials in which pairs of startle
stimuli separated by 1 - 2 s were presented so that the first startle
stimulus (S1) depressed the eyeblink to the second stimulus (S2).
On other trials, a prepulse (a visually presented word or word with
tone) preceded S1 by 150 ms. If a prepulse reduces the sensory impact
of a subsequent stimulus (S1), then the blink to S2 should be less
depressed by S1, and participants should judge S1 as less intense
than for a no-prepulse condition. As expected, prepulses inhibited
the blink reflex to S1, and S1 depressed the blink to S2. The size
of the blink response to S1 and the depressive effect of S1 on the
blink to S2 decreased with reductions in the sound level of S1. However,
S1 had the same depressive effect on a subsequently elicited blink
reflex whether or not the blink to S1 was inhibited by a prepulse.
Generally, estimation of the magnitude of S1 was strongly related
to its actual sound level. However, S1 was perceived as slightly less
intense when preceded by a prepulse. This perceived change in loudness
only occurred if the prepulse included a stimulus in the same modality
as S1, suggesting the operation of loudness assimilation. These data
fail to support the idea that a prepulse can gate out the sensory
attributes of a startle stimulus necessary to suppress a blink elicited
a second or two later, and provide little evidence for the notion
that a prepulse can diminish the perceived intensity of a startle

Gamma-band enhancement during Necker cube reversals
Canan Basar-Eroglu, Daniel Strueber, Michael Miener, Katrin Offe, & Michael Stadler
University of Bremen
Our research group recently reported results indicating strong increases
in the gamma-band during observation of a dynamic reversal pattern,
which is called stroboscopic alternative motion (SAM), Basar-Eroglu
et al., 1996, Int. J. Psychophysiol., 113-125. In case of viewing
reversible figures EEG correlates are of a special kind, because no
external trigger exists. The real stimulus is the switching, i.e.
an endogenous process. Therefore we assume that such a paradigm which
activates internal events is very important in order to understand
the functional role of different frequencies in the EEG. The aim of
the present study was to find out whether the increase in the gamma-band
during viewing the dynamic SAM can also be observed during reversals
of the static Necker cube. EEG was recorded from frontal, central,
parietal, temporal, and occipital locations from both hemispheres
in 10 subjects. There were 4 experimental conditions: A- Recording
of spontaneous EEG as baseline. B- Viewing of nonreversing cubical
figures. C- Naive observation of the Necker cube reversals. D- Active
observation of Necker cube reversals. In this condition, subjects
were instructed to press a button immediately following the switching,
whereas during naive observation no motor task was required. The highest
increase of the gamma-band amplitude was observed during naive observation
of the Necker cube, where the subjects reported to discover the yet
unknown event. These results are in accordance with our earlier results
with the SAM. According to these results we will discuss the functional
role of the gamma-band during viewing reversible patterns.

Aerobic fitness in the elderly: Effects on protein kinase activity and processing speed
Theodore R. Bashore1,2, H.-Y Wang2, & E. Friedman2
1University of Northern Colorado, 2Allegheny University of the Health Sciences
Platelet protein kinase (PKC) activity and translocation are reduced
in elderly men, but these reductions are mitigated by sustained aerobic
fitness (Wang, Bashore, & Friedman, 1995). We examined the relationship
of PKC activity to neurocognitive processing speed in the men studied
by Wang et al.. 8 young and 18 older nonexercisers (mean ages=29 and
70, respectively), and 12 young and 19 older aerobic exercisers (mean
ages=30 and 68, respectively) were screened medically, given an exercise
stress test, and matched on IQ. They completed a choice RT task in
which stimulus discriminability and S-R compatibility were varied:
LEFT or RIGHT was embedded in a matrix of number signs (#) or letters
chosen randomly from A-G or A-Z and subjects responded with either
the same or opposite hand indicated by the word. The choice reaction
elicited a series of ERP components at the midline scalp: the N60,
N160, P200, N260, and P300. Young exercisers and nonexercisers differed
on none of the measures. Age-related differences were observed for
the latencies of the N60, P200, N260, and P300 components, and RT
However, the only measure of processing speed that differed among
older exercisers and nonexercisers was reaction time: Older exercisers
had significantly faster RTs than older nonexercisers (842 vs 961
ms). These results suggest that the fitness-related changes evident
in older men at the molecular level may be manifest in a reasonably
specific influence on processing speed that is restricted to the later,
perhaps response-related, end of processing.

Hemispheric asymmetries in the recognition of familiar and unfamiliar faces
Gordon Bazana, Kimberly Cote, Robert Stelmack, & Kenneth Campbell
University of Ottawa
A simple discrimination task was employed to assess hemispheric asymmetries
in face recognition. Five subjects were required to differentiate
familiar target faces from unfamiliar non-targets. In the learn phase,
subjects were presented with 6 target faces, each of which were presented
6 times. In the subsequent practice phase, subjects saw the target
faces embedded in a series of 24 non-targets. The learn phase was
repeated. In the test phase, subjects saw the 6 targets embedded in
a series of 24 faces which were different than the non-targets in
the practice phase. ERPs were recorded from 19 scalp electrodes (Fz,
Cz, Pz, Fp1, Fp2, F3, F4, C3, C4, P3, P4, F7, F8, T7, T8, P7, P8,
O1, O2). Analyses were conducted on N1, P2, N2 and P3 ERP components.
ERPs elicited by familiar faces were significantly larger than those
elicited by unfamiliar faces at 11 electrode sites (Fz, Cz, Pz, F3,
F4, C3, C4, P3, P4, P7, P8). This effect was observed for the P3 component
only. No latency effects were observed, indicating that it did not
take significantly longer to classify the unfamiliar faces. Amplitudes
were significantly larger in the target condition at Cz, Pz, C4, and
P4 electrode sites. Amplitudes in the non-target condition were largest
at Cz, Pz, and P4. Results confirmed previous findings of right hemispheric
superiority for face recognition.

Electrocortical correlates of picture processing: The influence of task demands
Gabriele Becker1, Stefanie Maier1, Ewald Naumann1, Oliver Diedrich2, & Dieter Bartussek1
1University of Trier, 2University of Tuebingen
In our studies about the processing of emotional slides, a frontal
positive slow wave was observed which seems to reflect specific aspects
of information processing which focus on the content of complex stimuli.
Positive and neutral pictures were presented in a categorical oddball
task. Demands on information processing were varied in three experimental
conditions through the number of distinct stimuli which had to be
categorized (group I: one positive, one neutral slide; group II: fiv
positive, five neutral slides; group III: 30 positive, 30 neutral
slides). Subjects had to count the number of neutral or positive pictures.
The frontal positivity should be elicited in group III, where each
picture had to be processed in detail to decide whether it belonged
to the target category. In group I, the frontal positivity should
not appear, because subjects can just perceptually discriminate the
two pictures for categorization. Pictures were presented for 6 to
7.5 sec. EEG was recorded for 5000 ms at F3, Fz, F4, C3, Cz, C4, P3,
Pz, and P4. The results show an effect of processing complexity on
the frontal positivity: The frontal positivity is elicited only in
group III, where each picture had to be processed according to its
emotional content. In group I and II, where slides were presented
repeatedly, no frontal positive slow wave appears. The results indicate
that the frontal positivity is only elicited during complex processing
of stimuli with regard to their content, but not during the processing
of their structural features. Supported by Deutsche Forschungsgemeinschaft.

Effects of exposure of relevant information to innocent subjects on the efficiency of the GKT: An attempt to reduce false-positive outcomes by introducing target stimuli
Gershon Ben-Shakhar1, Eitan Elaad2 and Nurit Gronau2
1The Hebrew University of Jerusalem, 2Israel Police Headquarters
The Guilty Knowledge Test (GKT) is a method of psychophysiological
detection of concealed information that has been extensively researched
and discussed in the literature. It is based on a series of multiple-choice
questions, each having one relevant alternative (e.g., a feature of
the crime under investigation) and several neutral (control) alternatives,
chosen so that innocent suspects would not be able to discriminate
them from the relevant alternative. Although many studies have demonstrated
that the GKT can be used quite efficiently, leakage of relevant information
to innocent suspects can undermine its validity and increase the rate
of false-positive outcomes. The present study attempts to reduce false-positive
outcomes, due to leakage of relevant items, by introducing a set of
target items (i.e., items known to all examinees, but unrelated to
the crime), to which subjects have to respond (e.g., by pressing a
key), while answering the standard GKT questions. The results were
consistent with those obtained in previous studies, in demonstrating
that informed innocent participants showed relatively larger electrodermal
responses to the critical items than non-informed participants, but
smaller than guilty ones. When a respiration measure was used, no
differences between informed and non-informed innocent participants
were obtained. The use of the target items tended to reduce the differences
between informed and non-informed innocents. These results indicate
that knowledge of the relevant information is not in itself sufficient
for differential autonomic responsivity.

Orienting response reinstatement and dishabituation: The effects of substituting, adding, and deleting components of nonsignificant stimuli
Gershon Ben-Shakhar & Itamar Gati
The Hebrew University of Jerusalem
This study examined predictions derived from a feature-matching theory
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 in two experiments. Both used a simple habituation paradigm
with sequences of nonsignificant verbal and pictorial compound stimuli.
A test stimulus (TS) was presented after 9 repetitions of a standard
stimulus (SS), followed by 3 additional repetitions of SS. In Experiment
1, TS was created by substituting 0, 1, or 2 stimulus components of
SS, and stimulus change was manipulated both within and between categories.
In Experiment 2, TS was created by either adding or deleting 0, 1,
or 2 components of SS. The dependent measure was the electrodermal
component of the OR to both TS (OR reinstatement) and SS that immediately
followed TS (dishabituation). The results of Experiment 1 (N=168)
supported our predictions that substituting components of neutral
stimuli affects OR reinstatement, and that between-categories substitution
is more effective than within-categories substitution. Experiment
2 (N=152) demonstrated that both adding and deleting components of
neutral stimuli affects OR reinstatement, with no differences between
these two manipulations. In both experiments, OR magnitude was not
significantly affected by whether stimulus change was created by adding,
deleting or substituting two or just a single component, and no dishabituation
effects were obtained.

Using time-frequency to model multiple dimensions of P300
Edward Bernat, Tzu-Hsien Sang, William Williams, Scott Bunce, & Howard Shevrin
University of Michigan
This study investigates time-frequency representations of the P300
component. A specially developed algorithm makes it possible to reduce
high resolution time-frequency P300 representations to multiple dimensions
carrying unique information. Specifically, we hypothesized that: 1)
P300 could be measured in terms of stable constituent features, and
2) that these features carry different category information. P300
time-frequency representations showed a distinguishable radial 'arm'
structure extending from the time-frequency location of the P300 amplitude
peak on the time-frequency plane. An algorithm to parameterize these
arms was designed, which rotates an arm anchored at the peak of P300
through 180 degrees. Regression was performed on log transformed dat
for each step through the 180 degrees, and the mean squared error
(MSE) was computed. Peak MSE and associated step measures were then
entered into a discriminant analysis for pleasant and unpleasant words
as indexed by Osgood's Semantic Differential scale. Results demonstrated
that the model significantly differentiates ERPs to pleasant and unpleasant
words. Importantly, in this model, features representing the rise,
peak, and decay of the P300 response are measured separately. Peak
and decay features were primarily responsible for correct categorization.
These analyses of supraliminal word presentations are contrasted to
analyses of subliminal presentations of the same stimuli, which showed
categorization primarily in the features representing the P300 rise.
This model is discussed as a method for parameterizing the P300 component
into a multidimensional model, in which the dimensions carry different

Inhibiting facial expressions: Voluntary control under explicit social demand
Francois Bherer, Arvid Kappas, & Melanie Theriault
Laval University
Recently, Fridlund (e.g., 1994) and others suggested there is no relationship
between emotional state and facial activity. These authors argue that
"facial expressions of emotion" can be completely accounted for by
"social motivation". To clarify the influence of social motivation
on the production of facial displays, we created an explicit motivation
by using facial inhibition instructions. Thirty-seven participants
saw 6 funny and 3 neutral video stimuli while facial EMG was taken
at three sites (Zygomaticus Major, Orbicularis Oculi, Corrugator Supercilii).
Funny stimuli were randomly assigned to two conditions (spontaneous,
inhibited), while the neutral stimuli served as control in a spontaneous
neutral condition. Thus, each condition was repeated 3 times in a
complete counter balanced sequence in a fully within-subjects design.
Results indicated that while participants were able to significantly
reduce activity at the Zygomaticus (F(1,35) = 93.56, p < .001 and
Orbicularis oculi (F(1,35) = 76.09, p < .001) sites when inhibiting
amusement, they still displayed a significantly higher level of Zygomatic
activation compared to watching the neutral stimuli F(1,35) = 5.84,
p < .021. Furthermore, they displayed less Corrugator activity when
inhibiting than in the neutral condition (F(1,35) = 38.24, p < .001).
Our results suggest that explicit motivation by instruction is not
sufficient to mask the effects of spontaneous facial activation linked
to emotional stimuli. Obviously, voluntary control can determine facial
activation to a large degree, but there are non-voluntary components
which are not purely reflexive in nature, that can also contribute
in specific circumstances, such as when watching funny videos. We
will discuss whether these non-voluntary influences are better attributed
to the concept of social motivation or to emotio

Acute changes in carbon dioxide levels do not affect cognitive function but do affect EEG in healthy subjects
Elisabeth Bloch-Salisbury, Robert W. Lansing, & Steven A. Shea
Harvard School of Public Health, University of Arizona, & Harvard Medical School
Arterial PCO2 (PaCO2) is tightly regulated in healthy individuals.
Yet, there are marked variations in PaCO2 among individuals (i.e,
normal range approx. 32-44 mmHg) as well as acute changes in PaCO2
that occur within an individual (e.g., during speech; environmental
conditions). Acute increases and decreases in PaCO2 affect brain blood
flow and neuronal excitability. Large acute increases have been shown
to adversely affect performance on cognitive tasks. We tested whether
such cognitive deficits are a direct effect of PCO2 on brain state,
and unrelated to ventilatory changes and respiratory discomfort that
typically occur with acute increases in PCO2. To control for changes
in ventilation and discomfort subjects were passively ventilated on
a mechanical ventilator at fixed, high ventilation. Nine healthy subjects
performed five cognitive tasks selected from the Walter Reed Performance
Assessment Battery (PAB), a standard 'oddball' auditory ERP task,
and an EEG task (Alpha Attenuation Test) during three levels of end-tidal
PCO2 (PETCO2; a non-invasive estimate of PaCO2): Low (mean=30 mmHg);
Normal (mean=38 mmHg); and High (47 mmHg). There were no effects of
PETCO2 on 1)any PAB test (reaction times or error scores); 2)N1, P2,
or P3 (latencies or amplitudes); or 3)Alpha Attenuation Ratio (index
of alertness). However, we did find that EEG centroid frequency decreased
with increases in PETCO2 (p<.05). These results suggest that in healthy
individuals, acute modest changes in PETCO2 do not seem to affect
cognitive function, but do subtly alter CNS excitability. Supported
by: spinal cord research foundation/paralyzed Veterans of America.

Habituation versus the effect of a social encounter: A discriminative decrease of the startle blink in humans
A.J.W. Boelhouwer1 & T.D. Blumenthal2
1Tilburg University, 2Wake Forest University
A social encounter decreases startle eyeblink magnitude (Britt & Blumenthal,
1993). Since decreased reflex magnitude can be seen due to either
habituation or to other experimentally induced effects, the present
study was designed to distinguish between habituation and social encounter
effects. Four groups of 25 subjects each participated in four experimental
stages: Baseline (subject alone); Anticipation (subject alone and
told that an observer would enter the room soon); Encounter (observer
entered and sat in front of the subject (Group 1), or sat behind th
subject (Group 2), or the subject was told that the observer would
not be coming (Group 3); Recovery (subject alone in the room). The
habituation group (Group 4) received no instructions before each stage,
sitting alone throughout. Startle eyeblinks were electrically elicited
approximately twice a minute, for a total of 20 trials per stage.
The positioning of the observer made no difference, arguing against
a visual prepulse effect. R1 blink magnitude decreased during anticipation
of a social encounter, decreased further during the encounter, and
recovered after the encounter. When the encounter was canceled (Group
3), recovery of R1 started immediately after the anticipation stage.
R2 blink magnitude decreased when a social encounter was anticipated,
and decreased further during the encounter, but did not recover significantly
after the encounter. R2 magnitude was reduced more by the encounter
than by habituation. This supports the conclusion that different mechanisms
are involved in blink magnitude decreases during social encounter
and habituation.

Television and arousal: SCR responses to pacing and content arousal
Paul D. Bolls, Robert F. Potter, & Annie Lang
Indiana University
What makes television arousing? Obviously the content of the message
(sex, news, violence) is one important element. In addition, it is
often argued that the form of the message (i.e. how the message is
produced) may independently elicit arousal. For example, television
viewers rate fast paced messages as more arousing than slow paced
messages. This finding, however, is confounded by the fact that messages
about arousing contents are often the same messages which use fast
paced production techniques. This study explores the independent effects
of content arousal and production pacing on the frequency and amplitude
of spontaneous SCRs. It is expected that content arousal will primarily
affect viewers' sense of emotional arousal which should be reflected
in the amplitude of spontaneous SCRs occurring during viewing. Pacing,
on the other hand, is expected to impact orienting behavior, reflective,
perhaps, of an attention dimension of arousal, and as a result impact
the frequency (rather than the amplitude) of spontaneous SCRs. 35
adult subjects viewed one of four orders of 12 30-second commercials
that varied on production pacing and content arousal. SC was measured
time locked to message presentation. Frequency of spontaneous SCR,
amplitude of the largest SCR, and average amplitude of SCR were computed
for each message. Results show that messages with arousing contents
elicited more frequent SCRs with larger amplitudes than did calm contents.
Fast paced messages elicited more frequent SCRs than slow paced messages.
Pacing did not impact the amplitude of the SCRs.

Lack of control of fixation or focussing does not explain variation in photic blink amplitu
A.R. Bos, A.J.W. Boelhouwer, & C.H.M. Brunia
Tilburg University
Photic blink amplitude sometimes is and sometimes is not found to
be influenced by the direction of attention. This inconsistency might
stem from a lack of control of fixation or focusing of the eye. If
control is not adequate, attentional effects on the blink reflex are
confounded with effects of changes in the location of the flash on
the retina or focus changes. Therefore, the effect of systematic variations
in fixation and focus on the photic blink were studied. Blink reflexes
were elicited in 26 subjects by a photo flasher positioned on top
of a computer screen at 1 m distance, while subjects looked at the
upper or lower edge of the screen or at the center. No significant
influence of gaze direction on blink amplitude was found. Subsequently,
subjects looked at fixation dots 70 and 140 cm in front of them, while
reflexes were elicited by a flasher at 1 m distance 11o above them.
Again, no significant differences in blink amplitude were found. Since
large systematic variation in fixation and focus did not lead to significant
differences in photic blink amplitude, it was concluded that attentional
effects on the photic blink reflex cannot be attributed to floppy
control of peripheral receptor adaptation.

Investigation of response organization in Attention Deficit Hyperactivity Disorder using Lateralized Readiness Potentials
M. Bourassa, P. Robaey, J. Renaud, F. Karayanidis, G. Geoffroy, & G. Pelletier
Ste-Justine Hospital
Attention Deficit Hyperactivity Disorder is one of the most prevalent
childhood psychiatric disorders. Although the name of the disorder
suggests a deficit in the input stages of information processing,
recent behavioral and pharmacological data locate the deficit and
the effect of medication (methylphenidate) in response organization
stages. Two experiments were performed to test this hypothesis. First,
hyperactive and age-matched control boys were compared in a stimulus-response
compatibility task. Secondly, a different group of hyperactive subjects
was included in a double-blind cross-over methylphenidate-placebo
design using the same task. EEG was recorded over electrodes C1 and
C2 (motor cortices) while subjects responded to arrows with the compatible
(same direction as arrow, 66% of trials) or the incompatible hand
(opposite to arrow, 33%). Stimulus/hand compatibility was determined
by stimulus color. For both studies, preliminary reaction time analyses
performed for subgroups of subjects showed a significant GROUP (or
MEDICATION) X HAND interaction which tended to differ across conditions.
Given these results, LRPs amplitudes were computed for the two hands
separately for subjects included in the first study. On all successfu
trials, an activation of the incorrect hand preceding the preparation
of the correct hand was observed in a 600-200msec window prior to
the response, relative to a pre-stimulus baseline. Although preliminary
analyses of these potentials are not significant, the differences
obtained indicate a GROUP X HAND interaction that could reflect impairment
in left hand inhibition processes. To further investigate this tendency,
analyses with the full groups of twenty subjects will be performed
and presented.

Effects of directed attention and social encounter on startle eyeblink reactivity
M. Scott Bovelsky & Terry D. Blumenthal
Wake Forest University
The startle reflex has been used to explore such diverse topics as
emotion and attention. A social encounter, which is often considered
to be unpleasant, has been shown in previous research to attenuate
the blink response. The present research explored this social encounter
effect more closely in order to determine if, in fact, attention might
play a key role. Participants (N=18) were exposed to a series of 105
dB noise bursts, and eyeblink EMG responses were recorded in two sessions,
one week apart. Each session included five blocks of trials, an initial
baseline block, three task blocks (attend to nothing, attend to acoustic
startle stimulus, attend to visual task, in counterbalanced order),
and a recovery block. In one of the two sessions (counterbalanced
for order), an observer entered the testing chamber and sat directly
behind the subject during the three task blocks, ostensibly to evaluate
the subject's non-verbal behavior. As attention was drawn away from
the eliciting stimulus by the visual task, blink amplitude, duration,
and probability decreased, and blink latency increased. Directing
attention towards the startle stimulus resulted in increased blink
amplitude, duration, and probability, and decreased blink latency.
The presence of a social encounter resulted in reduced blink amplitude
and duration, similar to, and additive with, the effect of directing
attention away from the startle stimulus (e.g. in the visual task).
These results are consistent with an attentional explanation for the
effects of a social encounter on startle reactivity.

What's new? What's exciting? Novelty and emotion in perception
Margaret Bradley, Misty Kolchakian, Bruce Cuthbert & Peter Lang
University of Florida
A number of physiological systems appear to respond to both novelty
and affect. For instance, both heart rate and skin conductance sho
characteristic responses to novel events, and differentially to those
that are emotionally salient. The startle reflex also reflects both
attentional and emotional processes. In this study, we assessed effects
of novelty and emotion on heart rate, skin conductance, corrugator
EMG responses, and the startle reflex by presenting a series of affective
and neutral pictures for two contiguous processing periods. The duration
of the first period was either .5 or 4 s (between subject); the second
presentation was 6 s. A central goal was to assess effects of pleasantness
and arousal on responses in the presence and absence of novelty. Results
indicated that heart rate primarily reflected processes related to
sensory intake: Deceleration was predominant when the picture was
perceptually available. Affect modulated this effect most strongly
during the second presentation, in which larger decelerations were
obtained for emotionally arousing pictures. Skin conductance responses
were larger during initial processing, demonstrating a clear novelty
effect, but this difference varied with emotional valence: Whereas
responses were equally large to aversive materials in both intervals,
pleasant pictures prompted less activity in the second viewing, especially
for women. The startle blink response was primarily sensitive to emotional
valence and, in particular, blinks were larger during unpleasant,
compared to pleasant, pictures even for probes presented very early
in the second period. These data suggest a method for distinguishing
effects of novelty, emotion, and sensory intake on physiological responses,
with ramification for theories of orienting and emotion.

Social support may mitigate hostility's influence on blood pressure in young men
Edith E. Bragdon1, Kimberly A. Brownley2, Sheila G. West1, Susan S. Girdler1, Andrew Sherwood3, & Kathleen C. Light1
1University of North Carolina at Chapel Hill, 2University of Miami, 3Duke University
Although hostility has been proposed to contribute to heart disease
and hypertension by affecting cardiovascular stress responses, investigators
have not consistently detected a link between hostility and blood
pressure or heart rate. Clinic systolic and diastolic blood pressure
(SBP,DBP) were obtained from 98 men, ages 25-33, 63 of whom also underwent
24-hour ambulatory monitoring o f blood pressure (BP) and heart rate
(HR) during a typical work day. In addition, subjects completed the
Buss-Durkee Hostility Inventory and the Sarason Social Support Questionnaire
(SSQ). Controlling for body mass index, high hostility score was associated
with high HR when at work either alone (r=.38, p=.005) or with others
(r=.28, p=.04). For the group as a whole, hostility was not significantly
correlated with BP measured in clinic, at work, at home, driving or
sleeping. However, in subjects reporting low satisfaction with their
level of social support (by median split of the SSQ Satisfaction scale),
higher hostility was significantly associated with higher clinic SBP
(r=.41, p=.007) and DBP (r=.41, p=.008), as well as with higher SB
while driving a car (r=.43, p=.05). For subjects reporting high satisfaction
with social support, hostility was not related to higher BP during
any activity. These findings suggest that hostility may influence
cardiovascular measures more in some life situations than in others,
and that a satisfactory level of social support may buffer deleterious
effects of hostility on BP.

Relationship between complexity and predictability of alpha EEG
Michael E. Brandt1, Ahmet Ademoglu1, & Walter S. Pritchard2,3
1University of Texas-Houston Medical School, 2R.J. Reynolds Tobacco Company, 3Bowman Gray School of Medicine
Two prediction techniques were used to investigate the dynamical complexity
of the alpha EEG; a nonlinear prediction method using the K-nearest
neighbor local linear (KNNLL) approximation, and a linear prediction
method based on autoregressive (AR) modeling. Generally, KNNLL has
more predictive ability to detect nonlinearity in a chaotic time series
under noise-free and moderately noisy conditions versus AR prediction
as demonstrated using increasingly noisy realizations of the Henon
(a low-dimensional chaotic) and Mackey-Glass (a high-dimensional chaotic)
maps. However, at higher noise levels KNNLL prediction performs no
better than AR prediction. For linear time series, such as a sine
wave with added Gaussian noise, prediction using KNNLL is no better
than AR even at very low signal-to-noise ratios. Both prediction techniques
were applied to resting EEG scalp recordings (O2 site, 10-20 system)
from 10 subjects each having 3 separate recordings collected under
both eyes closed and eyes open conditions. Each recording was 1024
points corresponding to 8 sec. The first 512 points were used for
training and the second 512 for prediction. The cumulative relative
mean squared error (RMSE) for one-step-ahead prediction was calculated
and the model parameters were optimized to yield the best prediction
performance. In all recordings tested, KNNLL prediction did not yield
a lower RMSE than linear AR prediction. This result more closely resembles
that obtained for prediction of noisy sine waves as opposed to prediction
of chaotic time series with added noise. This lends support to the
notion that these signals have a linear-stochastic nature. However,
the possibility that some of our EEGs, particularly those with high
prediction error, were produced by a noisy nonlinear system cannot
be ruled out.

Peak (but not end) ANS reactivity to aversive episodes predicts bracing for anticipated re-experience
Christine A. Branigan1, Jessica F. Moise1, Barbara L. Fredrickson1, & Daniel Kahneman2
1University of Michigan, 2Princeton University
People's retrospective evaluations of emotional episodes have been
shown to correspond to the affect experienced at the most intense
(PEAK) and final (END) moments, with other features of the episode
(e.g., duration) largely inconsequential. We proposed that support
for this PEAK+END Rule would best predict retrospective evaluations
registered in self-reports of subjective experience, and that a simpler
PEAK Rule would best predict retrospective evaluations registered
in ANS changes. Fifty-two Ss (46 % female) viewed two matched series
of slides selected for highly negative valence. Each series was 8
or 11 slides long, and was framed in Blue or Green. Half the time
each series also included a "better end" -- three additional slides
that were less negative. Later, Ss were informed that they would view
the Blue or the Green series again. Continuous measures of ANS activity
were obtained throughout, including finger pulse amplitude (FPA) and
skin conductance response (SCR). During the period of anticipated
re-viewing, Ss showed decreased FPA (mean = -.22; t(51) = 5.15, p
< .001) and increased SCR (mean = +.44; t(51) = 4.42, p < .001). These
changes in FPA (SCR) at anticipated re-viewing were predicted by the
Peak -- but not End -- FPA (SCR) values obtained during the initial
viewing (FPA: peak beta = .41, p < .01; end beta = -.03, ns; SCR:
peak beta = .27, p = .08; end beta = .01, ns). Findings support a
simpler PEAK Rule for capturing the body's "memory" for past aversive

Confidence volume of a single dipole location in EEG
Christoph Braun1, Stefan Kaiser1, Wilhelm-Emil Kincses1, & Thomas Elbert2
1University of Tuebingen, 2University of Konstanz
Single dipole models have been proved to be useful for the localization
of neural sources of EEG activity. For the interpretation of the source
localization results, the knowledge of the statistical accuracy of
dipole locations is important. The accuracy of localization depends
essentially on the noise of the measured potentials. The aim of this
study is to compare four different strategies for estimating the localization
accuracy of EEG data. Besides Monte Carlo simulations (MCS), confidence
volumes were also estimated using two methods with different assumptions
about a linear transfer function between dipole location and potential
distribution. Finally, the confidence volume was estimated through
a procedure which used no linearization of the transfer function.
The four different methods were evaluated using a variety of manipulations
such as different levels and types of noise , i.e. correlated and
uncorrelated noise between electrodes. Confidence volumes were calculated
for simulated distributions of the electrical potential and for experimental
data including somatosensory evoked responses to stimulation of lower
lip, thumb, and little finger. The most important result for the simulate
data indicated that only the MCS yielded reasonable confidence volume
sizes if noise was correlated. Since the noise in experimental data
is highly correlated, only the MCS method appears to be useful in
estimating the size of the confidence volume of the dipole locations.
Thus, using real data with the MCS method, we easily distinguished
separate and distinct representations of the different stimulation
sites in the somatosensory cortex. Supported by the DFG (BI195/24).

Maturational changes in quantitative EEG in Attention Deficit Disorder
Susan M. Bresnahan1,2, John W. Anderson2, & Robert J. Barry1,2
1University of Wollongong, 2Neuroscience Psychological Services
Attention Deficit Hyperactivity Disorder (ADHD) in children and adolescents
is characterised by excessive restlessness and an extremely poor concentration
span resulting in impulsive and disruptive behaviour. Clinical observation
of Attention Deficit Disorder (ADD) in adults suggests that hyperactivity
diminishes in terms of its impact on social and academic function
while impulsive behaviours increase. EEG studies of children and adolescents
with ADHD have reported significantly higher levels of low frequency
activity (predominantly theta) and lower levels of beta activity compared
with normal controls. Other studies examining EEG activity in normal
children and adolescents have reported a decrease in theta activity
and an increase in beta activity with maturation. We examined the
relationship between age related changes reported in clinical observation
and the maturational changes in EEG occurring in a group of ADD patients
ranging in age from 6 - 42 years. Quantitative EEGs were obtained
from the midline sites of 25 adults, 25 adolescents and 25 children
diagnosed with ADD, and compared with those of age and sex matched
normal controls. The difference between relative beta activity in
the ADD groups and normal controls reduced with increasing age while
relative theta activity remained elevated across age in the ADD groups
compared with normal controls. Given that the hyperactivity component
in ADD reduces with age and the impulsivity component increases or
remains unchanged, these data suggest that beta activity may be linked
to hyperactivity and theta activity to impulsivity.

Covert and overt attention to threat: Effects of trait anxiety
Niall Broomfield and Graham Turpin
University of Sheffield
Broomfield and Turpin (Psychophysiology, 1996, Vol.33, No.S1, p.S
26) reported that high trait anxious subjects showed an uninstructe
eye movement bias toward threat cues during completion of a modified
Posner cue-target task. In addition, analysis indicated an absence
of the usual cue-facilitation of reaction time (RT). This was explained
in relation to a possible masking effect. The present study explored
these findings further. Subjects were allocated to one of two (N=20)
groups: high and low trait anxious. Each participant completed 96
reaction time (RT) trials. Half of the subjects were instructed to
maintain fixation and respond to a target appearing in the cued or
uncued hemifield. The remaining half were required to fixate only
at the start of each trial. Masking was manipulated by the relative
spatial position of the cue to the target. The remaining methodological
details are as detailed in Broomfield and Turpin (1996). Analysis
of unmasked trial data demonstrated a facilitation effect for non-threat
trials (p < 0.03). RT data also indicated that high anxious subjects
were significantly faster to disengage attention from threat cues
(p < 0.05). Electrooculographic data demonstrated that all subjects,
regardless of instructional set, made significantly more eye movements
towards threat cues (p < 0.005). Significant higher order interactions
(p < 0.05) for the heart rate analysis revealed differences in a secondary
acceleration to threat cues between the high and low anxious groups.
The results are discussed with respect to current theories of attention
and anxiety.

Estrogen replacement and vascular responses at rest and during stress: Preliminary findings of a 6-month randomized, placebo-controlled trial
Heather J. Brown, Susan S. Girdler, Sheila G. West, Traci L. Lamothe, Jean Ranc, John Steege, Catherine L. Stanwyck, Sunny H. Chung, & Kathleen C. Light
The University of North Carolina at Chapel Hill
Observational studies suggest that estrogen replacement lowers cardiovascular
risk in postmenopausal women. In animal models, estrogen enhances
vasodilatation and reduces agonist-induced vasoconstriction. In a
6-month intervention, 72 naturally and surgically postmenopausal women
ages 40-69 have been randomized to estrogen replacement or placebo.
In 37 women completing the full 6 months to da te, 28 women received
estrogen replacement and 9 women received placebo (mean ages 51.3
vs. 53.7, weight 165.8 vs. 166.1 lb.). Prior to treatment, groups
did not differ in mean arterial pressure (MAP) or total peripheral
resistance (TPR) at baseline (93.4 vs. 93.5 mmHg, 1651 vs. 1647 dyne
-sec*cm-5) or during behavioral stressors (speech, speech playback,
Stroop, and cold pressor). After 6 mont hs of treatment, repeated
measures MANOVA showed a Treatment Group X Time (Pretreatment vs.
Posttreatment) effect (p<.019). Estrogen subjects showed significant
decreases in TPR from pretreatment to posttreatment at baseline (1651
to 1490) and during tasks, while Placebo subjects showed significant
increases in TPR at baseline (1647 to 1759) and during tasks. Diastoli
pressure was also decreased only in Estrogen subjects by 4 mmHg at
baseline and by 5, 8, 5 and 3 mmHg during the stressors, respectively.
MAP showed similar significant decreases across conditions in the
Estrogen group only, while systolic pressure was unchanged. Thus,
these preliminary results, based on a randomized, placebo-controlled
trial, indicate that estrogen replacement reduces vascular responses
at rest and during stress in postmenopausal women.

Adrenocortical responses during positive and negative affective states
Tony W. Buchanan1, Thomas M. Donath2, Mustafa al'Absi1, Richard J. Davidson3, John T. Cacioppo4, & William R. Lovallo1
1Veterans Affairs Medical Center & University of Oklahoma Health Sciences Center, 2Heinrich-Heine-Universitaet, 3University of Wisconsin, 4The Ohio State University
The impact of emotions on endocrine outflow is important for understanding
relationships between personality characteristics, stress responses,
and health. Prior studies have reported secretion of cortisol to be
primarily influenced by negative affect, but less is known about cortisol
activity during states of activation involving positive affect. Accordingly,10
healthy men ages 21-28 years, each experienced: positive affect while
viewing an activating and humorous video; negative affect during a
public speaking stressor; and neutral affect during a period of rest.
Affect was documented by the Positive Affect/Negative Affect Schedule,
which measures the separate dimensions of positive affect (feelings
of enthusiasm, activation, and alertness) and negative affect (distress
and unpleasant engagement). Sessions began at 13:00 on separate days.
Cortisol concentrations were measured in saliva taken before and after
each 30-min mood induction. Changes in affect and cortisol were tested
using repeated measures ANOVAs which showed significant differences
across conditions in both variables (p's < .002). Positive affect
was increased to a similar degree by both the video and the speech
task compared to rest (p's < 0.008). Negative affect increased during
the speech task (p's < 0.012) compared to rest and the video, which
in turn were not different from each other. Cortisol output increased
only during the speech (p's < 0.015). The lack of cortisol response
during the video suggests that positive affect is not associated with
activation of the hypothalamic-pituitary-adrenocortical axis. Therefore,
glucocorticoid responses appear to be particularly linked to central
mechanisms underlying negative emotional states.

Subliminal (but not supraliminal) presentation of affective words elicits differential corrugator EMG
Scott C. Bunce1, Edward Bernat1, Howard Shevrin1, & Stephen Hibbard2
1University of Michigan, 2Pacific Graduate School of Psychology
Priming studies have demonstrated that the subliminal presentation
of affectively-valent stimuli can bias subsequent affective judgments
of neutral stimuli (Murphy & Zajonc, 1993; Journal of Personality
& Social Psychology, 64, 723-739). These same stimuli, when presented
at supraliminal durations, do not bias subsequent affective judgments.
This study examined the ability of affectively-valent mood adjectives
to elicit mood congruent responses in facial musculature under both
supraliminal and subliminal conditions. 32 mood adjectives were presented
to 8 right-handed subjects at both supraliminal (40 ms@5 ft/lamberts)
and subliminal (1 ms@5 ft/lamberts) durations. Analyses were based
on each individual's 10 most positively and 10 most negatively rated
words according to the Osgood Semantic Differential Evaluative Scale.
Negative words elicited greater corrugator EMG than positive words
at subliminal durations, a facial response consistently associated
with the experience of negative affect. Negative words did not elicit
differential effects at supraliminal durations. However, negative
words did elicit greater P300 amplitude than positive words in the
supraliminal condition at both frontal (F3, F4) and parietal (P4)
electrode sites. In the subliminal condition, negative words elicited
greater P300 amplitude at electrode P3 than P4, and greater P200 amplitude
than positive words at electrode F3. ERP results confirmed that differentiation
of word meaning occurred at both subliminal and supraliminal durations.
These results suggest that affectively-valent stimuli have the ability
to elicit affect-relevant systems when perceived out of awareness,
but may be subject to different regulation processes when consciously

Startle reflex modulation in a non-clinical PTSD sample
Stephanie D. Burks, Kakanang Maneesri, Ellen L. Spence, & Edwin W. Cook III
University of Alabama at Birmingham
Exaggerated startle is one diagnostic criterion for Posttraumatic
Stress Disorder (PTSD); however, empirical studies have not consistently
observed elevated startle in this disorder. In non-clinical samples,
a range of PTSD-related characteristics, including fearfulness and
positive schizotypy, have shown reliable correlations with affective
startle modulation. The present study examined relationships among
PTSD, related traits and symptoms, and startle responding in varying
affective contexts. In large screening sessions, undergraduates completed
a series of questionnaires, including the PTSD Checklist. Participants
were selected for the laboratory study if they reported either few
or no PTSD symptoms (n=27), or met diagnostic criteria for PTSD with
varying levels of severity (n=34). To manipulate emotion, subjects
were presented with 3-sentence scripts in 3 affective categories
fear, neutral, and joy. Acoustic startle probes were presented while
subjects read the scripts from a video screen and subsequently imagined
the situations described. Startle blink, heart rate, and skin conductance
level were measured. PTSD symptomatology was not reliably related
to overall startle blink magnitude, either during the inter-trial
intervals or averaged across affective conditions. However, individuals
reporting more severe PTSD symptoms showed greater potentiation of
startle magnitude while reading and imagining fear compared to joy
situations. Further regression analyses suggested that this relationship
was mediated in part by fearfulness, positive schizotypy, and a tendency
toward dissociative experiences. These results provide further evidence
that exaggerated startle in PTSD depends on an aversive context. They
also suggest characteristics of PTSD that may contribute to enhanced
affective startle modulation in this disorder.

Long-term test-retest reliabilities for autonomic, neuroendocrine, and immune variables in older women
Mary H. Burleson1, Kirsten M. Poehlmann2, Louise C. Hawkley2, John M. Ernst2, William B. Malarkey1, Gary G. Berntson2, Janice K. Kiecolt-Glaser1, Ron Glaser1, & John T. Cacioppo2
1The Ohio State University College of Medicine, 2The Ohio State University
Reliability of physiological measures over long time periods (e.g.,
years) is crucial in the design of longitudinal studies of physiological
responses to stress. In the current study, 19 older women caring for
a spouse with progressive dementia and 26 controls who were category-matched
for age and family income performed speech and math stress tasks on
two occasions approximately one year apart. Cardiovascular, neuroendocrine,
and immune measures were taken before and immediately following stressor
exposure during both years. Year-to-year test-retest reliability's
were calculated for baseline and change scores for the following variables:
heart rate; pre-ejection period; respiratory sinus arrhythmia (RSA);
systolic and diastolic blood pressure (DBP); respiration amplitude
and rate; plasma epinephrine, norepinephrine, adrenocorticotropic
hormone, cortisol, and antibody to Epstein-Barr virus; natural killer
(NK) cell cytotoxicity; lymphocyte blastogenic response to concanavalin
A and phytohemagglutinin; circulating numbers and percentages of T
cells, T4 cells, T8 cells, NK cells; and the T4/T8 ratio. Baseline
reliabilities were highly significant (p<.01) for all variables except
DBP (p<.05) and respiration amplitude (p<.03). Reliability of change
scores was more variable, but in general, cardiorespiratory and neuroendocrine
reactivities were significantly reliable across time. An exception
was RSA, which was marginally significant, possibly due to characteristics
of the stress tasks. In addition, the overall patterns of reliability
across variables were very similar between caregivers and controls.
These results suggest that individuals can be reliably characterized
psychophysiologically in terms of basal activation and stress-reactivit
of the autonomic and neuroendocrine systems, as well as in some aspects
of immune function.

EEG asymmetry, salivary cortisol, and affect expression: An infant twin study
Kristin Buss, Isa Dolski, Jessica Malmstadt, Richard Davidson, and H. Hill Goldsmith.
University of Wisconsin-Madison
As part of an ongoing longitudinal twin study, we collected EEG, salivary
cortisol, and extensive behavioral and parent report measures of temperament
and emotions. To date, 30 6-month-old pairs of twins have been run
through these procedures. Infants were seen twice with approximately
one week separating the visits. EEG was recorded during 5 1-minute
baseline trials and during a m ale stranger approach episode. EEG
data was averaged across the two sessions. Saliva was collected after
each session (i.e., 'stress' measure) and on three consecutive days
in the home (i.e., baseline). Temperament was assessed with Rothbart's
Infant Behavior Questionnaire (IBQ) filled out by the mothers. Behavioral
data coded from videotape of the stranger approach episode included
facial an d bodily fear, facial and bodily sadness, attempts to escape,
and distress vocalizations. Given the small sample size, data was
not analyzed with respect to behavior genetics. Baseline salivary
cortisol values were positively correlated with escape behavior during
the stranger episode (r = .44, p = .08). Higher fear scores on the
IBQ were related to more baseline right frontal EEG activity (r =
-.28, p<.05). Furthermore, EEG asymmetry during the stranger episode
was correlated with soothability scores on the IBQ (r = -.49, p< .05).
In the stranger approach episode, right frontal EEG activity was associated
with more sadness facial expressions (r = -.40, p< .05), more distress
vocalizations (r = -.38, p= .06), and more bodily and facial sadness
combined (r = -.40, p< .05).

Contingent negative variation: Clarifying the course of development
Dana Byrd, Gregory J. Austin, & W. Keith Berg
University of Florida
Previous research (Austin, Berg, & Fields, 1996; Warren & Karrer,
1992) has indicated differences between children and adult slow cortical
wave patterns during a fixed-foreperiod interval. These findings have
not addressed the developmental changes of the slow cortical wave.
In order to investigate these issue, 29 children (4 years to 15 years)
and 9 young adults were tested in an fixed-foreperiod anticipatory
paradigm with an inter-stimulus interval of 6 seconds. Slow brain
potential from Fz, Cz, and vertical EOG were recorded with a 18 s
time constant as was second-by-second HR. Trials with evidence o
excessive eye movements were eliminated from averages. Results were
restricted to the initial block of 10 trials since the youngest subjects
showed evidence of fatigue on later trials. The grand average results
show positive, high amplitude waves during the warning interval in
5- to 7-year-old children, a result consistent with Austin, et al.,
1996. In 8 to 10 year olds, the response was similar, but less positive,
but the 11 to15 year group was again strongly positive. Adults showed
the typical negative slow potential ("CNV") . At Fz responses were
less consistent, except for being clearly positive in the youngest
group. Examination of the HR data indicated a typical anticipatory
deceleration in all four age groups. This study strongly suggests
that major developmental changes in slow brain potentials extend well
into the adolescent period. It remains to be determined if these reflect
differences in the cognitive processes occurring during anticipation,
in brain development, or both.

The practical effect of routine data transformations on absolute EEG power derived from spectral analysis
John A. Caldwell & Kristi A. Roberts
U.S. Army Aeromedical Research Laboratory
The fact that absolute EEG power from spectral analysis often is not
normally distributed has lead to the recommendation that log(x) transformations
be applied. This is because many statistical techniques including
analysis of variance (ANOVA) rely on normal distributions. However,
the practical extent of the problem is clouded because ANOVA is known
to be robust to violations of the normality assumption. Thus, it is
unclear whether or not transformations inevitably will lead to divergent
conclusions about the effects of independent variables on the EEG.
In the present study, EEGs were collected from 18 volunteers in a
sleep-deprivation paradigm. Subjects were evaluated at 1030, 1430,
and 1830 prior to deprivation; and at 0230, 0630, 1030, 1430, and
1830 during deprivation. Data from Fz, C3, Cz, C4, Pz, O1, and O2
were subjected to spectral analysis, and results were analyzed as
untransformed absolute power, log-natural transformed absolute power,
and relative power. Although the appearance of relationships among
various data points was similar regardless of the transformations,
there were differences in the number and location of statistically
significant effects. For instance, of 36 significant time effects,
9 were found on the absolute power data, 12 were found on transformed
absolute power, and 14 were found on relative power. Of the instances
in which significance was obtained on either of the absolute power
measures, there was agreement across both only 40 percent of the time.
Agreement across all 3 measures occurred only 16 percent of the time.
Thus, the decision to transform or forego transformations of EEG data
clearly can affect research conclusions.

P300 amplitude in male adolescent monozygotic twins discordant and concordant for alcohol disorders
Scott R. Carlson1, William G. Iacono1, Joanna Katsanis2, Matt McGue1, & Rana S. Altan1
1University of Minnesota, 2University of Arizona
Numerous studies have shown that the offspring of alcoholic fathers
have smaller P300 amplitudes than the offspring of non-alcoholic fathers.
This finding along with evidence indicating that P300 amplitude is
heritable suggests that this characteristic is a good candidate for
identifying genetic risk for alcoholism. To further explore this possibility,
we examined the P300 amplitude of four groups of male adolescent monozygotic
twins who were A) discordant for alcohol disorders (n = 18 pairs),
B) concordant for alcohol disorders (n=7 pairs), C) randomly selected
from a large community based sample of twins without alcohol disorders
(n=21 pairs), and D) concordant for the absence of substance, externalizing,
or major depressive disorders (n=9). In the discordant twin pairs,
no significant difference in P300 amplitude was found between the
alcohol abusing and non-alcohol abusing cotwins. Moreover, the P300
amplitude did not vary significantly across the twin groups discordant
for alcohol abuse, concordant for alcohol abuse, and randomly selected
from non-alcohol abusers. These three groups were found, however,
to have significantly smaller amplitudes than the pairs screened for
the absence of psychiatric disorders. The results of this study, suggest
that P300 amplitude may be suitable as a genetic marker for a group
of possibly etiologically linked forms of commonly co-occurring psychopathology.
An alternate possibility is that a general absence of psychiatric
problems may be associated with large P300 amplitude.

ERP reactivity to socially-relevant stimuli: A test of Gray's model
Helene Chevalier, Sid Segalowitz, Jane Dywan, Linda Rose-Krasnor, & Robert Nadon
Brock University
According to Gray's model (1972/82/87) there exist 2 primary independent
motivational systems: a behavioural inhibition system (BIS) and a
behavioural activation system (BAS). Grey proposes that activation
of the BIS produces increased attention, arousal and behavioural inhibition
characterized by neuroticism and introversion. Activation of the BAS
relates to stability and extraversion. Based on this distinction,
we predicted that those high on BIS-related characteristics would
respond to socially relevant stimuli with high levels of arousal and
attention. Conversely, those high on BAS-related characteristics will
be less physiologically reactive to such stimuli. We administered
a number of personality scales (e.g., EPQ, NEO-PI, STAI) and a scale
designed to measure BIS and BAS reactivity (Carver & White,1994) to
a cohort of 36 female undergraduates. Factor analysis yielded 3 factors:
Neuroticism, Situational-Inhibition, and Extraversion. We examined
the relationship between these factors and subjects' ERP respons
to target faces depicting positive emotions. Controlling for state
anxiety, we found a negative correlation between extraversion and
the amplitude of the early portion of the ERP (300-900 ms), the higher
the level of extraversion the less initial reactivity to face stimuli
(sr2 = .19 and .13 at Cz and Pz respectively). However, a high level
of neuroticism was positively related to reactivity in the later portion
of the waveform (900-1500 ms) suggesting greater physiological response
(sr2 = .17 and .22 at Cz and Pz respectively). Thus, individuals who
show high levels of inhibition as a general personality trait are
likely to be more physiologically reactive to socially relevant stimuli
than those who are more generally uninhibited.

Attention Deficit Hyperactivity Disorder: Subtypes or severity differences?
Adam R. Clarke1, Robert J. Barry1, Rory McCarthy2, & Mark Selikowitz2
1University of Wollongong, 2Private Paediatric Practice
Attention Deficit Hyperactivity Disorder (ADHD) has undergone significant
changes in classification over recent years, with the DSM-IV identifying
three subtypes depending on the clustering of symptoms. While previous
studies have consistently found EEG differences between children broadly
labeled as ADHD and normals, little attention has been paid to neurological
differences between the subtypes. We investigated differences in the
EEG between groups of children (each N = 40) with ADHD of the Combined
Type, ADHD of the Predominantly Inattentive Type and normal controls.
All subjects were between the ages of 8 and 12 years and groups were
matched on age and gender. EEG was recorded during an eyes-closed
resting condition from 21 monopolar derivations. These were clustered
into nine regions prior to analysis. One minute of trace was analyzed
using Fourier transformation to obtain both absolute and relative
power estimates in the delta, theta, alpha and beta frequency bands.
The patient groups were found to have greater levels of relative theta
and deficiencies in relative alpha and beta in comparison to the control
group. Quantifiable EEG differences were found between children with
ADHD of the Inattentive type and those of the Combined Type in relative
theta, alpha and beta activity, with the Inattentive Type intermediate
between the Combined group and normals. These results generally support
a maturational lag model of ADHD. However, the differences between
the diagnostic subtypes of ADHD appear to be quantitative rather than
qualitative, and suggest the need for a reappraisal of classification
in this disorder.

A re-examination of the paradoxical P300 laterality effect
Jeffrey M. Clarke, Voyko Kavcic, & Christina M. McCann
University of North Tex
Several researchers employing lateralized visual presentations have
observed larger P300 amplitudes to ipsilateral, than to contralateral,
visual field (VF) - hemisphere combinations. This finding is paradoxical
since it appears to reflect an advantage for the longer, indirect
anatomical pathway than for the shorter, direct pathway. In some of
these studies, response hand and VF of presentation are confounded,
which has made interpretation of these findings difficult. We examined
lateralized P300 responses from six lateralized tasks (variants of
lexical decision or Stroop-like tasks) conducted in our lab. For parietal
(P3 & P4) and temporal (T5 & T6) sites over a given hemisphere, whenever
we found a significant VF difference in P300 amplitudes, it was always
larger for the ipsilateral VF - hemisphere combination, regardless
of response hand. This effect tended to be present over both hemispheres
for temporal sites, but only over the left or right hemisphere, depending
on the task, for parietal sites. P300 amplitudes at temporal sites
appear to reflect direct/indirect pathways, in a paradoxical fashion,
while at parietal sites there appears to be influences of hemispheric
specialization. Kutas et al. (1990; J. Cognitive Neuroscience, 258-271)
also observed paradoxically-lateralized P300 effects in normals, but
not in split- brain patients where there tended to be an advantage
for the contralateral VF - hemisphere combination. This suggests that
the paradoxical P300 laterality effect depends on interhemispheric
interactions, and is consistent with unilateral cortical lesion findings
which demonstrated that the normal P300 response relies on interhemispheric
functions (Knight et al., 1989; Brain Research, 109-116).

As the stomach turns: Autonomic reactions to unpleasant film stimuli
Maurizio Codispoti1, Daniela Palomba2, Giovanni Tuozzi1, Michela Sarlo2, Bruno Baldaro1, & Luciano Stegagno2
1University of Bologna, 2University of Padova
In a previous study in our laboratory, larger parasympathetic cardiac
activity was evoked by the presentation of a surgery film compared
to a threat film. The present study was aimed at extending these findings.
In addition to heart rate and respiratory sinus arrhythmia, gastric
myoelectrical activity (EGG) was assessed as an extra-cardiac measure
of autonomic activity. Three film clips (300 seconds each) depicting
surgery, threat, and neutral material respectively, were presented
to 42 subjects. EGG, heart rate, and respiratory sinus arrhythmia
(RSA) were recorded during baseline and film viewing. EGG spectral
density was calculated in the following epochs: 1-2.5 cpm (bradygastric
rhythm), 2.5-3.75 cpm (normogastric rhythm), and 3.75-10 cpm (tachygastric
rhythm). RSA was defined as heart period variability within .12 -
.40 Hz, and computed according to the Porges method. A modified Differential
Emotions Scale was used to measure subjects' affective responses t
film sequences. Results showed that the surgery film evoked larger
HR deceleration, RSA and EGG activity compared to the threat film.
These results confirm our previous findings and support the notion
of parasympathetic influence on psychophysiological reactions to specific
unpleasant events.

P3a and P3b from auditory and visual stimuli
Marco Comerchero1, Jun'ichi Katayama2, & John Polich3
1University of California-San Diego, 2Hokkaido University, 3The Scripps Research Institute
Target (.10), nontarget (.10), and standard (.80) stimuli were presented
in three-stimulus oddball tasks employing either auditory or visual
stimuli, with participants responding only to the target stimuli.
Auditory stimulus pitch and visual stimulus size between the target
and standard stimuli (Easy vs. Difficult discrimination) and the same
attributes between the nontarget and standard stimuli (Small vs. Large
difference) were manipulated orthogonally. Similar ERP findings were
obtained for both modalities: When the target/standard discrimination
was Difficult, target P300 amplitude was smaller and demonstrated
longer peak latency than when the task was Easy. In contrast, nontarget
P300 amplitude varied according to target/standard discrimination
ease and whether the nontarget was similar to the target. In the Easy/Large
and Difficult/Small conditions, nontarget stimuli elicited a parietal
maximum but a smaller P300 (P3b) than did the target stimuli; in the
Easy/Small condition, nontarget stimuli elicited ERPs similar to those
observed for the frequent-standard stimulus. However, in the Difficult/Large
condition, nontarget stimuli elicited a P300 (P3a) that was larger
at the anterior site and shorter in latency than the P300 (P3b) from
the target stimuli. These findings suggest that in the three-stimulus
oddball paradigm, target P300 (P3b) components are not affected by
the nontarget stimulus configuration, whereas the nontarget P300 (P3a)
outcomes are determined directly by the stimulus context. Thus, distinct
P3a and P3b components can be generated in both auditory and visual
modalities, but the paradigm requires appropriate control over stimulus
feature attributes.

N400 and expectancy for semantically associated contexts in schizophrenia
Ruth Condray, Stuart R. Steinhauer, & Daniel P. van Kammen
University of Pittsburgh School of Medicine/VAMC
Language and attentional processes are typically disrupted in individuals
with schizophrenia, but the relationship between these cognitive functions
has not been well-delineated for this population. Pharmacologic modulatio
of these processes has also not been determined. Mechanisms of attention
engaged during language processing are frequently examined by manipulating
proportions of stimulus classes during semantic priming tasks. High
proportions of associated words are assumed to engage strategic or
focused attention; low proportions activate automatic processes. N400
and P300 elicited during lexical decision were compared for 35 male
controls tested once, and male schizophrenics evaluated during double-blind
haloperidol (n=41) and placebo (n=28) conditions. Primes preceded
target words or nonwords at interstimulus intervals of 250 and 850
ms. Word-word pairs varied in associative strength, and word frequency
was a nested factor. Expectancy or attention set was induced by presenting
high (strategic attention) or low (automatic activation) proportions
of associated word-pairs (high expectancy=67% associated; 16% unassociated).
Unassociated targets reliably evoked an enhanced negativity at approximately
400 ms. This priming effect did not differ between controls and patients
during haloperidol treatment. When drug-free, however, patients'
ERPs did not discriminate between linguistic contexts. Expectancy
exerted differential effects on N400 and P300 latency in controls
and patients under both ISIs. When drug-free, patients exhibited a
prolonged N400 latency under the strategic attention condition, and
a shortened P300 latency under the automatic activation condition.
Results suggest attentional mechanisms engaged during the processing
of linguistic context are disrupted in schizophrenia, and haloperidol
may moderate these dysfunctions.

The effects of respiratory maneuvers on EEG: Gender differences
Mary R. Cook1, Mary M. Gerkovich1, Kirsten E. Kakolewski1, Kasey L. Goings1, & Kathleen A. O'Connell2
1Midwest Research Institute, 2Kansas University Medical Center
As part of research on smoking, we designed respiratory maneuvers
(RMs) that, by altering carbon dioxide (CO2) levels, might mimic the
effects of smoking on brain arousal systems. Six RMs (three to increase
and three to decrease CO2) were tested on a sample of smokers (10
men, 10 women). CO2 was measured with an Ohmeda 4700 Oxicap. F3, F4,
P3, P4 and Cz were referenced to linked mastoids; Fp1 was used to
detect eye artifacts. Subjects were trained to perform each of the
maneuvers to criterion. EEG was digitized at 128Hz for spectral analysis.
Brain activity was recorded before and after each of the maneuvers;
two orders of maneuver presentation were used. Significant differences
in CO2 between men and women were found only for one maneuver (breath
hold); CO2 increased more for women than for men (p = .03). None of
the maneuvers designed to increase CO2 had differential EEG effects
in men and women, although there were significant gender main effects.
Gender effects were observed for two of the maneuvers designed to
decrease CO2. When subjects performed an overbreathe maneuver, alph
peak frequency increased for men and decreased for women (p = .03).
A maneuver that mimics puffing a cigarette caused Beta 1 amplitude
to increase in women and decrease in men. A maneuver that mimics laughing
altered EEG similarly for both sexes. The specific patterns (left
v right hemisphere, frontal v parietal) will be presented, and implications
for the use of respiratory maneuvers in applied psychophysiology will
be discussed.

EOG correction: Towards a simple solution
Rodney J. Croft & Robert J. Barry
University of Wollongong
Correction techniques have been developed to remove eye movement related
artifact from the electroencephalogram (EEG). Such techniques estimate
the proportion (Beta) of the eye movement related voltage (EOG) present
in the EEG, and subtract it from the EEG. Research has consistently
found that different correction coefficients are required for blinks,
saccades, and/or different frequencies associated with eye movement.
Limitations of the regression procedure suggest that such differences
in Beta may be purely artifactual, due to the lower power EOG used
for the non-blink Beta calculation. Two experiments explored this
hypothesis. The first used simulated data to investigate whether the
equivalent of low power EOG would inflate Beta when forward propagation
of the EEG to the EOG and coherent interference in the two channels
occurred. The second experiment used real data to determine whether
the patterns observed in the simulated data were realistic. The first
study found that forward propagation and coherent interference did
inflate Beta at simulated low EOG power, and that as power increased,
the degree of inflation decreased. The second study found the same
pattern as in the first study. Further, no difference was found between
blink and saccade Betas, and both eye movement types yielded significantly
smaller Betas than fixations. It was concluded that the results are
consistent with the thesis that eye movement related fields propagate
similarly for a range of EOG types and frequencies, and that the differences
usually reported may be caused by the regression approach introducing
significant artifact at low EOG power.

Recognition memory for novel environmental sounds is affected by memory instructions
Yael M. Cycowicz & David Friedman
New York State Psychiatric Institute
This study examined the effect of memory instructions on the ERPs
elicited by novel environmental sounds during a subsequent sound recognitio
memory task. Subjects (32) were assigned either to incidental or intentional
memory instructions. Prior to a novelty auditory oddball task, subjects
in the incidental group were told to press a button for the rare tones
and were not informed about the occurrence of the novel sounds. Subjects
in the intentional group were told not only to press a button for
the rare tones but to memorize the novel sounds for a subsequent memory
test. Following the oddball task, subjects made old/new recognition
judgments to randomly intermixed old and new sounds. Brain activity
was recorded during both novelty oddball and sound recognition tasks.
Recognition performance was similar in both groups (~60%), and was
significantly above chance. For both groups, ERPs to correctly recognized
old sounds showed larger positivity than those to correctly identified
new sounds. However, ERPs to false alarm (new) and miss (old) trials
differed between the groups. For the intentional group, ERP amplitudes
associated with false alarms and misses did not differ, while for
the incidental group ERPs associated with misses elicited significantly
larger amplitudes than these associated with false alarms. Thus, the
ERP waveforms may have indicated that the similar recognition performance
of the two groups was due to different brain mechanisms. Recognition
for the intentional group was probably based on explicit retrieval
processes, whereas for the incidental group it may have relied on
an implicit retrieval mechanism.

Heart rate reactivity during the Stroop test better predicted by previous night's sleep than pulse reactivity to mild stressors
J. Alexander Dale, Jodi Dworkin, & Jeffrey D. Cross
Allegheny College
From a sample of 129 male subjects who completed 2 stressor tasks,
serial subtraction and H-word generation, 14 subjects were selected
as low pulse rate reactors and 12 subjects were selected as high pulse
rate reactors from the top and bottom of the distribution. These subjects
had been trained to take their own pulses for one min prior to and
one min after the stress tasks administered in a group situation.
The groups were divided at the median number of hours of sleep (between
6.5 and 7 hours) reported the previous night before the completion
of the Stroop Test. Electrocardiogram and systolic blood pressures
were assessed for a one min baseline before and for one min at the
midpoint and one min the end of a 6 min computer administered Stroop
Test. Participants with relatively less sleep the night before the
Stroop Test were significantly more reactive in terms of their heart
rate than participants who slept seven hours or over (p >.05). Participants
in the "High Pulse Rate Reactors" were not significantly higher (p>.05)
than participants in the "Low Pulse Rate Reactors". Pulse rate change
to the group stressors was low in all participants and the group nature
of the tests was interpreted as ameliorating the normally stressfu
nature of these tasks. Systolic blood pressures were not available
for analysis at the time of preparation of this abstract. Participants
with relatively less sleep did not differ from their more rested counterparts
on Jenkins type A or Spielberger's anger indices.

Cardiovascular reactivity during re-lived emotion as a function of individual differences in anger-expression tendency
Kim M. Dalton, Brian D. Ostafin, & Richard J. Davidson.
University of Wisconsin-Madison.
This study assessed whether specific cardiovascular profiles associated
with the imagining of anger, fear, happiness and sadness were qualified
by individual differences in response tendencies during anger. Sixty
two under- graduates were selected for extreme scores on Anger-In,
Anger-Out, or Low-Anger tendencies, (Speilberger, 1988). Impedance
cardiography measures were obtained during 1 min baseline and guided
imagery trials of four emotional events from participants recent past.
Significant main effects were found for Group on resting CO, SV, and
TPR, (p < .03). The Low-Anger group had significantly higher CO than
the Anger-In and Anger-Out groups and larger SV than the Anger-In
group. The Anger-In group had a significantly larger TPR than the
Low-Anger group. Significant main effects for Group were also found
for HR and SBP reactivity during the Emotion trials, (p < .02). The
Anger-Out group showed a significantly larger increase in HR than
the Low-Anger group. The Anger-In group showed a significantly larger
increase in SBP than the Low-Anger and Anger-Out groups. Significant
main effects for Emotion were found for HR and SBP, (p < .05). Fear
and Anger had significantly larger HR increases than Happy. Fear also
had a significantly larger HR increase than Sad. Sad and Anger had
significantly greater increases in SBP than Happy. The only significant
Group x Emotion interaction was for DBP reactivity, p =.02. The Low-Anger
group showed significantly larger increases in DBP during Sad and
Fear relative to Anger. The Anger-Out group showed significantly larger
increases in DBP during Anger relative to Happy.

An innovative assessment of language comprehension with event-related potentials (ERPs)
Ryan C.N. D'Arcy & John F. Connolly
Dalhousie University
Previous research on receptive vocabulary has demonstrated that standardized
neuropsychological vocabulary measures can be adapted for event-related
potential (ERP) recordings. Accordingly, we have extended this work
to examine both speech and reading comprehension. The Token Test was
adapted for computer presentation to assess auditory comprehension
(De Renzi, E., & Vignolo, L.A., 1962. Brain, 85, 665-678.). In the
computer Token Test, an animated token sequence was presented visually,
followed by a spoken sentence that either correctly or incorrectl
described the animation. Spoken sentences that terminated with incongruent
words elicited ERP components associated with speech perception (i.e.,
N400). Further, the written sentence comprehension section of the
Psycholinguistic Assessment of Language Processing in Aphasia (PALPA)
was adapted for computer presentation to assess reading comprehension
(Kay, T., Lesser, R., & Coltheart, M., 1992). In the computer PALPA
test, sentences were presented on a computer screen along with three
line drawings, one of which corresponded with the target sentence.
Following the study phase, each line drawing was presented individually
and subjects were asked to indicate the drawing that they had selected.
Drawings that an individual had selected as 'correct' were differentiated
from alternative drawings on the basis of the P300 component which
is thought to reflect context updating and recognition memory. For
each of the computer adapted tests, normative data was obtained from
a sample of university students and behavioral results were contrasted
with the respective traditional test for validation. This research
prefaced ongoing investigations of language impairments in individuals
with limitations in communicative and motoric responses.

Time trend of attentional performance as an individual trait
Dmitry M. Davydov
Serbsky National Research Center for Social and Forensic Psychiatry
As discussed in the literature (Filion et al., 1994) the attention-orienting
task requires subjects attention to be engaged tonically on a given
channel. The objective of the present research was to examine individual
differences in the trend over time in the allocation of controlled
processing resources during performance of attentional tasks. Two
studies were conducted using tachistoscopic and reaction time methods
to assess attention. Cluster analysis was used to distinguish patterns
of performance over time. In the first study, the time trend in performance
was assessed as a function of situational stress; in the second study,
it was assessed relative to neuropsychiatric disorders. Cardiovascular
response to the tasks was assessed using forehead pulse volume measured
with a photo transducer or an impedance plethysmograph. Both studies
demonstrated a link between the dynamics of attention and other behavioral
and physiological processes (verbal threshold, overall task performance,
pulse volume, physiological intercorrelations). The time trend in
performance on the tasks was not affected by situational stress or
by neuropsychiatric disorders in between-subjects analyses. Both studies
showed that resting forehead pulse volume and its left/right asymmetry
were associated with the different patterns of performance over time
on the tasks. In general, results of the two studies suggest that
these differences in performance on attentional tasks may be a relatively
stable characteristic of the individual. This deduction is concordant
with earlier work indicating that the dynamics of allocation of controlle
processing resources is primarily determined by genetic factors (Lykken
et al. 1988).

Startle eyeblink modification in asymptomatic schizophrenia patients and controls
Michael E. Dawson1, Anne M. Schell2, Jonathan K. Wynn1, & Keith H. Nuechterlein3
1University of Southern California, 2Occidental College, 3University of California-Los Angeles
This report follows up and extends our previous finding of abnormalities
in startle eyeblink modification (SEM) in schizophrenia. Twenty-four
young, relatively asymptomatic recent onset schizophrenia patients
and 18 matched controls participated in a selective attention task
in which one of two tones in an intermixed series was attended and
the other was ignored. Judgements were made of the duration of attended
tones. Auditory startle probes were presented at lead intervals of
120 and 2000 ms during both attended and ignored tones and during
the intertone intervals. Consistent with previous findings, control
subjects showed significant attentional modulation of prepulse inhibition
at 120 ms, with greater inhibition occurring during the attended than
during the ignored tone. They also showed significant modulation of
long lead interval facilitation, with greater facilitation occurring
at 2000 ms during the attended than the ignored tone. Patients showed
normal levels of prepulse inhibition and late facilitation during
ignored tones, but they failed to show attentional modulation of SEM
at either 120 or 2000 ms. Among the patients, deficits in prepulse
inhibition during the attended tone correlated positively with Brief
Psychiatric Rating Scale (BPRS) factors of Thought Disturbance and
Positive Styptoms, as well as with the BPRS total score. These correlations
were traceable to the fact that patients who showed poorest prepulse
inhibition also scored highest on BPRS items reflecting unusual thought
content (including delusions) and conceptual disorganization, symptoms
which could be interpreted in terms of a sensorimotor gating dysfunction.

An experimental study of heart-beat detection in subjects with high vs. low somatic symptom index under conditions favouring interoception
Jasminca Dedic & Bernhard Dahme
University of Hamburg
Hartl & Strian (1995) reported considerable situational fluctuations
in the precision of heart-beat detection in somatoform patients with
autonomous cardiovascular symptoms ("heart neurosis"). In a new experiment
we expected that somatizers would have a higher sensitivity, but also--because
of their catastrophizing tendency--would tend to attribute a wide
range of events to the perception of heart-beats, thus maximizing
their false alarm rate. Method: 10 somatizers (SSI score > 11), an
10 matched controls without somatization tendency (SSI < 2), were
selected from a sample of 300 students. Subjects' attention was focused
on internal stimuli either by an "imaginative transformation", or
an "imaginative trip through the body", or an unstructured waiting
period. By the method of constant stimuli (Brener, Liu & Ring; 1993)
the subject was required to judge the simultaneity of heart beat sensations
and tones. SDT parameters were derived from the relative numbers
of hits (all tone sequences attributed to heart-beat within a perception
range from 150 to 350 ms after R-wave) and false alarms. Main results:
(1) The overall mean of the perceptual index was d' = .50; both main
effects were nonsignificant in a 2(groups) x 3(conditions) ANOVA repeated
measures design. (2) However, symptomatic subjects responded significantly
more often "yes" ("tones simultaneous to heart beat") and so produced
more false alarms than controls. Conclusions: (1) Both groups performed
moderately well in detecting their heart beats in this task. (2) The
high proportion of false alarms in somatizing subjects might be attributed
to a specific fear to miss interoceptive stimuli indicating certain
somatic symptoms.

Emotional information processing: The effects of depression on P300
Patricia J. Deldin1, Jennifer Keller2, John A. Gergen2,3, and Gregory A. Miller2,3
1Harvard University 2University of Illinois at Urbana-Champaign, 3Covenant Medical Center
P300 studies of depression have inconsistently found differences between
major depressives and controls. When a reduction is found it is usually
in simple oddball studies at Pz and interpreted as a reduction in
resources. More theoretically complex studies have been called for
to help clarify the inconsistent findings and to aid in the understanding
of the abnormal cognitive processing associated with depression. Given
the evidence for mood-congruent memory biases in depression, ERPs
were recorded in depressives and nondepressives during exposure to
emotionally valenced word and face stimuli (an implicit encoding and
categorization task) and then again during an explicit recognition
task. It was predicted that parietal P300 would be enhanced to negative
stimuli in depressives or enhanced to positive stimuli in controls
or both. We found no main effect for group; rather, our results suggest
a more specific, theory-related difference between groups. For nondepressives,
previously viewed faces elicited significantly larger P300 amplitudes
than novel negative faces, whereas the opposite was true for the positive
faces. The depressives showed similar novelty effects for positive
faces but failed to show the P300 differences if they had previously
viewed a negative face. Results suggest that depressives allocate
equal amounts of resources when recognizing positive and negative
faces whereas controls allocate fewer resources when recognizing the
positive faces. Therefore, depressives show an even-handed response
during recognition that is consistent with depressive realism bu
do not show the potentially protective bias of needing to allocate
less resources when recognizing positive stimuli.

Amygdala activation in social phobics during Pavlovian conditioning detected by functional MRI
Oliver Diedrich1, Wolfgang Grodd1, Ute Weiss2, Frank Schneider2, Herta Flor3, & Niels Birbaumer1
1University of Tuebingen, 2University of Duesseldorf, 3Humboldt University-Berlin
The neurobiological processes underlying the classical conditioning
of aversive responses were examined using a differential conditioning
paradigm, with slides of two different neutral faces serving as conditioned
stimuli (CS+, CS-) that were paired either with an aversive or a neutral
odor. Animal research has demonstrated that the amygdala is a key
structure for the processing of conditioned fear stimuli. Using functional
magnetic resonance imaging (fMRI), we investigated changes of the
regional cerebral bloodflow (rCBF) in amygdala and thalamus in social
phobics and healthy controls. While healthy controls exhibited an
increase of amygdala activation elicited by the slides after the conditioning
procedure, social phobics responded to the slides with enhanced rCBF
of the amygdala already before pairing these stimuli with the odors:
Initially, social phobics exhibited stronger amygdala activation after
the presentation of human faces than did the healthy controls; no
such difference was found for the odor stimuli. During acquisition,
the healthy controls' amygdala response to the slides increased, while
rCBF of the amygdala remained high in social phobics. No effects were
found for thalamus. The results are discussed with regard to the neurobiological
mechanisms of emotional processes in general and their disturbance
in phobic disorders.

Addressing misallocation of variance in principal components analysis of evoked potentials
Joseph Dien
University of Illinois at Urbana-Champaign
Interpretation of evoked response potentials is complicated by the
extensive superposition of multiple electrical events. The most common
approach to disentangling these features is principal components analysis
(PCA). Critics have demonstrated a number of caveats that complicate
interpretation, notably misallocation of variance and latency jitter.
This poster describes some further caveats to PCA as well as three
methods for addressing them: parallel analysis, oblique rotations,
and spatial PCA. To demonstrate these techniques, spatial PCA is used
to obtain an estimate of the dimensionality of the late positive complex
in a 64-channel dataset. Temporal PCA is used to decompose the earl
potentials contributing to the N1 and posterior N2. It is concluded
that PCA is an essential statistical tool for ERP analysis, but only
if applied appropriately.

The effect of reward and punishment on ERN amplitude in high and low Socialized students
Ziya V. Dikman, John J. B. Allen, Jamie Gray, & Sohee Jun
University of Arizona
The Error-Related Negativity (ERN) is a response-locked brain potential
generated when subjects make mistakes during simple decision-making
tasks, and its amplitude appears to vary as a function of the awareness
people have of making errors. The present experiment was designed
to examine ERN amplitude in groups of subjects for whom error salience
might differ as a function of reward or punishment. Although ideally
psychopathic subjects would be examined, as a first step an analog
study was conducted examining subjects who scored low on the Socialization
scale of the California Psychological Inventory. The consequences
of making an error were manipulated for all subjects, such that in
one half of the trials subjects were monetarily rewarded (REW) for
making correct responses, and in the other half subjects were punished
(PUN) with a 95 dB tone for making incorrect responses. Fifteen high
socialized (SO) and 15 low SO introductory psychology students were
selected to participate in the current study. ERN amplitudes were
computed for REW and PUN conditions for sites Fz, Cz, and Pz. Analysis
of the data via a repeated-measures ANOVA (SO group x PUN-REW X site)
demonstrated a significant 2 way interaction between SO and PUN-REW
(p<.02). The data indicate that low SO participants produce smaller
ERNs during the punishment task than during the reward task, while
high SO participants produce similar ERNs in both conditions. RT and
EMG data essentially bolster the interpretation that the ERN effects
reflect differences in error salience for high and low Socialized

Is blood pressure related hypoalgesia related to stress induced analgesia?
Blaine Ditto1, Janis France1, & Christopher R. France2
1McGill University, 2Ohio University
Decreased pain perception has been observed in hypertensive patients
and normotensive individuals at risk for hypertension. It is possible
that this hypoalgesia is a variant of stress induced analgesia as
exaggerated cardiovascular reactivity to stress has been observed
in these same groups and some have proposed baroreflex stimulation
as the mediator of the effect. Electric shock and the cold pressor
test were presented to 24 healthy women with and 24 without a parenta
history of hypertension on two counterbalanced days. The pain stimuli
were preceded by a stressful 20 minute videogame on one day and by
20 minutes of a control activity on the other. Consistent with previous
research, the videogame often produced substantial increases in blood
pressure (mean SBP increase = 21 mmHg) while no increase in blood
pressure was observed prior to the pain stimuli on the control day.
Women with a parental history of hypertension and high blood pressure
reactivity to the videogame reported significantly less shock pain
on both days, suggesting an association between risk for high blood
pressure and decreased pain sensitivity in women. Since the same pattern
of results was observed on the control day, the actual degree of blood
pressure elevation prior to shock did not seem to be an important
determinant of pain. However, other analyses revealed an effect of
family history on shock pain only among women who reported relatively
high anxiety, suggesting that other aspects of the stress response
may be involved in this phenomenon.

Individual differences in P300 latency, reaction time, and movement time on a stimulus-response compatibility task
Cynthia Doucet and Robert Stelmack
University of Ottawa
ERPs and reaction time measures were recorded from introverts and
extraverts in a stimulus-response compatibility task (N=60). Stimuli
consisted of an arrow array in which a middle arrow was flanked by
congruent arrows (pointing in the same direction as the middle arrow),
or incongruent arrows (pointing in the opposite direction). Response
compatibility was cued by the word "Same" or "Opposite", preceding
the arrow array. By use of a home button and two target buttons, reaction
time (lift off from the home button) was distinguished from movement
time. Results indicated that both P300 latency and RT were faster
to congruent stimuli than to incongruent stimuli, and to compatible
responses vs incompatible responses. No effects with personality were
observed for either P300 or RT. Movement time, however, was faster
in extraverts than in introverts for all conditions.

Laser-evoked pain inhibits innocuous-related activity in the primary somatosensory cortex
Robert Dowman, Mose Herne, & Shamika Morris
Clarkson University
Our earlier studies strongly suggest that the amplitude of a positive
potential (SP4a: 160-180 ms) in the human sural nerve-evoked somatosensory
evoked potential reflects innocuous-related neural activity in th
primary somatosensory cortex (SI). We also observed that electrical
stimulus intensities that recruit the nociceptive peripheral afferents
(intensities which also elicit maximal activity in the innocuous afferents)
decrease SP4a amplitude, suggesting that activity in the pain pathways
inhibits innocuous-related SI activity. We investigated this hypothesis
in the present study by examining the effects of a painful 200 ms
CO2 laser pulse applied to the skin of the right foot on the SP4a
amplitude elicited by non-painful electrical stimulation of the right
sural nerve at the ankle. In each subject SP4a amplitudes were recorded
under 3 conditions: electrical alone (pre-laser control), followed
by electrical and laser presented together; followed by electrical
alone (post-laser control). In 2 separate groups (15 subjects/group),
painful laser stimuli whose onset was timed to elicit neural activity
that overlapped with that elicited by the non-painful electrical stimuli
yielded smaller SP4a amplitudes than the pre- and post-laser control
values (p<.01). The painful laser stimuli had no effect on SP4a amplitude
(p>.10) in a third group where there was no temporal overlap between
the neural activity elicited by the laser and electrical stimuli.
These results are consistent with our hypothesis that activity in
the pain pathways inhibits innocuous-related SI activity, and suggests
that this inhibition depends on the overlap between the pain-evoked
and innocuous electrical-evoked neural activity.

ERP indices of attention in absence and complex partial seizure disorders
Connie C. Duncan1, Allan F. Mirsky2, Christopher T. Lovelace2, & William H. Theodore3
1Uniformed Services University of the Health Sciences, 2National Institute of Mental Health, 3National Institute of Neurological Disorders & Stroke
Our previous research demonstrated impaired attention in patients
with absence epilepsy. In the current investigation, we used ERPs
to study sustained attention in matched groups of 9 patients with
absence epilepsy (AB), 13 patients with complex partial epilepsy (CP),
and 9 healthy control (HC) subjects. ERPs were recorded during two
visual and two auditory versions of the Continuous Performance Test
(CPT) of sustained attention. Increased rates of omission errors on
the visual and auditory CPT confirmed that AB patients are impaired
in their ability to sustain attention. The performance differences
were reflected in the ERP data: AB patients had significantly reduced
P300 and slow wave amplitudes on both visual and auditory tasks in
comparison with HC subjects. Although the CP patients and the HC subjects
did not differ in performance on the CPT, there were substantial differences
in brain activity. Thus, both AB and CP patients had smaller visual
P300 and slow wave components relative to HC subjects. However, such
differences occurred more than twice as often in the AB group than
in the CP group. Moreover, whereas the significant amplitude reductions
were almost equally divided between visual and auditory tasks in th
AB group, only a single significant reduction was seen on an auditory
task in the CP group. The results have implications for theories of
the pathophysiological processes underlying absence and complex partial
seizures, on the nature of the attentional capacities of patients
with seizure disorders, and on the accumulating body of evidence that
there exist separate visual and auditory attention systems in the

Cardiovascular responsivity during the phases of menstrual cycle
Nukte Edguer
Brandon University
The present study investigated the effects of menstrual cycle phase
on cardiovascular responsivity during a stressful task. Women were
tested either at the follicular phase (N=12), luteal phase (N=20)
or during menses (N=14) using two versions of Vandenberg Mental Rotation
Test. Stress was created by informing subjects at the end of the first
version that their performance was below average. Before starting
the second version of the test subjects were also told that because
they had some experience with the test they were expected to improve
their score. Subjects' cardiovascular responses were monitored throughout
the experiment, and they completed the Multiple Affect Adjective Check
List-R (MAACL-R) at the beginning of the experiment as well as after
completion of the two versions of the test. The results of the study
indicated that there was a significant difference in heart rate responsivity
as a function of menstrual cycle phase, but no significant differences
in blood pressure responsivity.

Limitations of extended source models: Measurement noise and model errors
Wilhelm-Emil Kincses1, Christoph Braun1, Stefan Kaiser1, & Thomas Elbert2
1University of Tuebingen, 2University of Konstanz
Extended source models using anatomical and physiological knowledge
as constraints provide more information on the characteristics of
neural sources than single dipole models. In the 'cortical patch model'
the cortical surface is taken as the anatomical constraint. This is
based on the fact that the EEG is generated by dendritic currents
of pyramidal cells perpendicular to the cortical surface. Accordingly,
the orientation of the dipole is constrained to be orthogonal to cortical
surface. The cortical surface is determined by segmentation and triangulation
of NMR images. In this approach the extended source consists of a
contiguous area with homogeneous activity density. Measurement noise,
errors in the head model and misslocations of electrode position
lead to inaccurate source extensions. To get an estimate of the impact
of these factors on the variability of the source extensions, several
variations of noise level and electrode position errors were simulated.
In addition the influence of different head model parameter was evaluated.
The simulations and analysis demonstrate that extended source models
can give detailed information on the size of active cortical regions,
assuming that data noise and measurement errors are below 5%. Supported
by DFG (BI195/24)

Anger Expression Style: Construct validity and exercise-induced ischemia
Tilmer O. Engebretson
Ohio State University
We recently advocated use of a new measure called Anger Expression
Style (AES) to quantify mode of anger expression and demonstrated
that rigid styles of anger expression, but not flexible styles, were
associated with elevated cholesterol (Engebretson & Stoney, 1995).
In this study involving cardiac patients and their spouses, we further
validate the AES construct and examine its association with exercise-induced
ischemia. Subjects were 36 middle-aged patients (and their spouses)
undergoing initial evaluations for heart disease. Ischemia was determined
from thallium-201 exercise stress tests. Patients' AES scores correlated
with spouses' reports of the patient's AES scores (r=.43, p=.02),
spouses' Anger Out scores (r=.53, p=.003) but not their Anger In scores
(r=-.14, p=.5), and with the statements; they "express their anger
" (r=.38, p=.04) and they "keep things in" (r=-.41, p=.02). Examination
of AES scores and extent of ischemia revealed that patient's own AES
scores were not related to ischemia (p>.2). Spouse reports of patients'
anger expression styles, however, were related to ischemia (p=.005).
Those with a rigid expressive style exhibited the greatest degree
of ischemia (mean = 5.0): those with flexible modes exhibited the
lowest degree of ischemia (means = 2.5 & .9); and those with a rigid
non-expressive style showed intermediate levels (mean = 2.6). These
data provide support for the AES construct and suggest that those
with a rigid expressive mode of managing anger may be at risk for
heart disease via heightened myocardial ischemia.

Validation of a new ambulatory impedance cardiograph
John M. Ernst, Ray B. Kowalewski, Louise C. Hawkley, Mary H. Burleson, David Lozano, Gary G. Berntson, & John T. Cacioppo
The Ohio State University
The impedance cardiograph (ZCG) is a non-invasive measure of cardia
function that is frequently employed in psychophysiology (Sherwood
et al., 1990). We report a validation study of an ambulatory ZCG.
Using a standard tetrapolar band electrode configuration (Sherwood
et al., 1990), we recorded electrocardiogram, basal impedance, and
the first derivative of basal impedance from 11 undergraduate students
during rest and during four mental stressors. We used both a new ambulatory
impedance cardiograph (AZCG) developed by World Wide Medical Instruments,
Inc. and the standard Minnesota Impedance Cardiograph Model 304B (MZCG)
built by Surcom Inc. Cardiac activity during half of each rest and
stressor period were collected with either the AZCG or MZCG (counterbalanced).
Heart rate (HR), pre-ejection period (PEP), and stroke volume (SV)
were derived using ensemble software (Kelsey & Guthlein, 1990). For
cardiac activity at rest, the results indicated no difference for
HR (p = .460), but significant differences for PEP (p <.001) and SV
(p < .05). However, for cardiac reactivity (stressor - rest), the
results indicated no difference for HR (p = .578), PEP (p = .434),
and SV (p = .816). In addition, correlations for the AZCG versus MZCG
were strong at rest for HR, PEP, and SV (all r values > .9), and the
correlations for reactivity were strong also for HR (r = .616) and
PEP (r = .825). The correlation was not strong for SV (r = .306).
Thus, these data suggest that for chronotropic measures such as HR
and PEP this AZCG is valid relative to the MZCG, but not for inotropic
measures such as SV.

Categorization of words and pictures in sentence contexts
Kara D. Federmeier & Marta Kutas
University of California-San Diego
Language comprehension involves the retrieval and integration of knowledge
stored in long-term memory. How is that knowledge structured in memory
and how does its structure affect on-line sentence processing? We
addressed these questions in two event-related potential studies using
words and pictures, respectively. We examined the relative contribution
of category membership (one type of long-term memory structure) and
contextual plausibility on the amplitude of the negativity in the
300-500 millisecond time window ("N400"). Volunteers read pairs of
sentences for comprehension. The first sentence (e.g., ``He caught
the pass and scored another touchdown.'') set up an expectation for
a particular exemplar of a particular category while the second sentence
(e.g. ``There was nothing he enjoyed more than a game of ...'') ended
with (1) the expected exemplar (``football''), (2) an unexpected exemplar
from the expected category (``baseball''), or (3) an unexpected exemplar
from an unexpected category (``chess''). All unexpected items yielded
N400s, but when the targets were words there was a clear effect of
category structure on processing: N400 amplitudes to unexpected items
from the expected category were significantly smaller than to unexpecte
items from an unexpected category. This influence of category structure
could be observed even when it went counter to contextual plausibility,
providing some evidence for its automatic use in sentence comprehension.
When the target concepts were represented as line drawings, however,
N400s to unexpected items were equally large regardless of category
membership, suggesting a greater influence of context than category
structure on processing.

Defensiveness in stress-induced asthma
Jonathan M. Feldman1, Paul M. Lehrer2, Richard E. Carr2, & Stuart M. Hochron2
1Rutgers University, 2UMDNJ--Robert Wood Johnson Medical School
This exploratory study examined the relationship between defensiveness,
laboratory stressors, and pulmonary function in 101 asthmatics, and
32 nonasthmatics. All subjects were young adults, ranging in age from
18 to 40 years. Participants who scored in the top quartile of the
Marlowe-Crowne Social Desirability Scale were defined as defensive.
Subjects participated in an experiment consisting of a rest period
and four tasks (mental arithmetic, reaction time, and two stressful
movies). There were no differences on pulmonary function between defensives
and nondefensives before the stressors. A series of ANCOVAs were performed
with post-stressor pulmonary measures as the dependent variable and
pre-stressor measures as the covariate. Asthmatics showed a greater
decline from baseline to the post-session period compared to nonasthmatics
on FEV1/FVC (p<.01) and % peak flow (p<.05). Nonasthmatics showed
no significant changes in pulmonary function or effects of defensiveness
and therefore were not included in subsequent analyses. Among asthmatics,
defensives demonstrated greater decreases from baseline to the post-stressor
period on FEV1/FVC (p<.01), %FEV1 (p<.05), and %FEF50 (p<.01) in comparison
to nondefensive asthmatics. Matched t-tests comparing baseline to
post-session revealed that defensive asthmatics declined in FEV1/FVC
(p<.01), while nondefensive asthmatics improved in %FEV1 (p<.05).
These results suggest that defensiveness may play an important role
in airway reactivity in stress-induced asthma.

Trait anxiety and individual differences in startle modification
Diane L. Filion & Catana E. Brown
University of Kansas Medical Center
Startle modification measures, both prepulse inhibition and emotional
modulation measures, are sensitive indicators of attentional and emotional
deficits in many psychiatric disorders including schizophrenia and
anxiety disorders. However, variability in startle modification patterns
also exists within the normal population and the distributions of
startle modification scores for control and clinical participants
often overlap. The purpose of this research was to examine the possibilit
that variations in trait anxiety within a non-clinical population
may play a role in determining individual differences in startle modification.
College undergraduates completed Speilberger's Trait Anxiety Inventory
and students with the highest and lowest scores were selected for
startle testing. The protocol involved presentation of positive, negative,
and neutral pictures from the IAPS with startle probes presented on
a subset of trials at either 120 ms or 3500 ms as well as during selected
inter-trial intervals to provide a baseline measurement. Preliminary
results from 10 high and 10 low trait anxiety participants revealed
that the high trait anxiety group exhibited significantly larger startle
responses during the baseline condition than the low trait anxiety
group. In addition, for the 3500 ms probe position, the high trait
anxiety group exhibited significantly larger startle responses during
the neutral-picture condition than the low trait anxiety group as
well as significantly greater startle potentiation in the negative-picture
condition. Results for the 120 ms probe position are currently being
analyzed. These preliminary results suggest that anxiety may be an
important factor in determining individual differences in startle
modification within the normal population.

Quantitative EEG and familial risk for alcoholism: Patterns of anterior and posterior slow alpha
Peter R. Finn & Alicia Justus
Indiana University
Although quantitative EEG studies of resting EEG have consistently
reported that alcoholics have reduced alpha power, studies of the
offspring of alcoholics have yeilded a mix, somewhat equivocal, set
of results. Issues in this literature include small sample sizes,
including only male subjects, not assessing the correlates of EEG
activity, and assessing EEG before or after significant experimental
manipulations. The purpose of this study was to assess the psychological
and familial correlates of resting slow alpha EEG power in young adult
men and women with a positive family history (FHP: n=68; 41 men, 27
women) and a negative family history of alcoholism (FHN: n=67, 37
men 30 women). Subjects participated in 3 sessions, two assessing
personal characteristics and family history of psychopathology, and
one involving only the assessment of resting EEG. Results indicated
that both male and female FHP subjects had significantly reduced absolute
and relative slow alpha power in posterior regions (T5,P3,O1,Pz,P4,O2),
and had significant differences in the laterality of slow alpha in
anterior, lateral regions (F7/F8 and T3/T: more slow alpha on left
relative to right) compared with FHN subjects. Reduced posterior slow
alpha was also associated with a family history of depression, but
not antisocial personality, anxiety, violence, drug abuse, or psychosis.
Laterality differences were also associated with a history of proband
drug abuse and childhood hyperactivity. Results are discussed in term
of the heterogeneity in familial-risk for alcoholism and different
vulnerabilities associated with familial-risk.

Abstraction ability, P300, and psychosis
Iris A. Fischer, Martha E. Shenton, Robert W. McCarley, Paola Mazzoni, Michele Harper, & Dean F. Salisbury
Harvard Medical School & McLean Hospital
Thought disorder is present in schizophrenia and bipolar disorder.
On the Johnston-Holzman TDI, thought disorder correlates with left
temporal scalp P300 amplitude. Another component of thought disorder
is abstraction ability, perception beyond concrete features. Abstraction
ability was measured by proverb interpretation in 24 schizophrenic
patients, 13 psychotic bipolar patients and 24 controls. P300 was
also recorded as subjects silently counted target tones (1.5kHz, 97dB,
15%) among standard tones (1 kHz, 97 dB) against 70dB white noise.
One-way ANOVA revealed that schizophrenics, bipolars, and controls
differed significantly in abstraction ability. Post-hoc tests revealed
that both psychotic groups had significantly worse abstraction ability
than controls, and schizophrenics had significantly lower scores than
bipolars. Both psychotic groups had significantly reduced peak P300
amplitude at Cz compared to controls. No significant correlations
were found between abstraction ability and P300 measures for schizophrenics
or bipolars. Abstraction scores in schizophrenics correlated negatively
with Brief Psychiatric Rating Scale (BPRS) scores (-.42, p=.042),
but there was no significant correlation in bipolars. No significant
correlations were found with BPRS and P300 in schizophrenics. However,
a negative correlation was found between P300 at Fz and BPRS in bipolars
(-.68, p=.015). These data suggest that schizophrenics have worse
abstraction ability than controls and psychotic bipolars. Thus, abstraction
ability in schizophrenia may be sensitive to underlying psychosis.
Abstraction ability did not correlate with P300, and thus may measure
a different facet of thought disorder than TDI. In bipolars, P3 from
frontal sites correlated with overall psychosis, which may reflect
selective frontal lobe abnormalities in this sample.

Effects of startle on a motor programming task
Denis F. Fitzpatrick1 & Stephen R. Paige2
1University of Kansas Medical Center, 2University of Nebraska at Omaha
It has been proposed that one function of the startle response is
to interrupt ongoing activity to prepare the organism for emergency
action. To examine this notion, electromyographic (EMG), electroencephalographic
and electrocardiographic activity were recorded as subjects (n=9)
performed memorized sequences of button presses involving the index
and middle fingers of both hands. In a given block of trials, subjects
chose from one of two sequences. Sequence choice was signaled by the
lighting of one of two horizontally-arranged light-emitting diodes
(LEDs). On 10% of all trials, 110 dB 50 ms duration white noise stimuli
were delivered 200 ms after the onset of the choice signal. Startle
reliably increased reaction time (RT) to the first button press in
the sequence by a mean of 220 ms. Subjects with larger orbicularis
oculi and sternocleidomastoid EMG amplitude responses to the startling
stimuli showed greater RT slowing. Within-subject (trial-by-trial)
variations in RT slowing due to startle were associated with N100
amplitude. Changes in heart rate (HR) evoked by the startling stimulus
were not related to RT slowing. The present study suggests that startle
may act to interrupt ongoing activity when uncertainty of action is
present or when subjects are committed to temporally extended sequences
of action.

Experimental investigation of the placebo and nocebo effects
Magne A. Flaten1, Terje Simonsen2, & Harald Olsen3
1University of Tromsoe, 2Tromsoe University Hospital, 3University of Oslo
Three groups of subjects (N=65) received Information that the drug
administered to them acted as a relaxant (Relaxant group) or a stimulant
(Stimulant group). The third group did not receive any Information
about the drug effect (Control group). Half of the subjects in each
group received 525 mg of Carisoprodol, that induces muscle relaxation
and tiredness, and dampens polysynaptic reflexes. The other half received
an inactive agent (lactose). Thus, responses to Information alone,
and responses to Information in combination with Carisoprodol were
obtained. Dependent variables were blink reflexes, skin conductance
responses, and subjective measures of relaxation and tiredness. It
was hypothesized that Information alone would induce responses that
matched the content of the Information. Furthermore, the Information-elicited
responses were expected to interact with Carisoprodol, so when Information
matched the effect of the Drug (Relaxant group), the drug response
should increase. The opposite result was expected when there was a
mismatch between Information and Carisoprodol (Stimulant group). The
results indicated that Information about drug effects induced physiological
and subjective responses that matched the Information. These responses
increased the relaxant effects of Carisoprodol in the Relaxant group.
There was no effect of Carisoprodol in the Stimulant group, which
indicates that drug-opposite Information canceled the drug response.
It is concluded that information about drug effects may elicit physiological
and subjective responses that strengthen or weaken the drug response.
Such drug-agonistic or -antagonistic responses may be interprete
as placebo- and nocebo-responses, respectively.

EEG alpha asymmetry in dysthymia and major depression: Effects of comorbidity with an anxiety disorder
Regan Fong, Gerard Bruder, Craig Tenke, Paul Leite, James Towey, Jonathan Stewart, Patrick McGrath & Frederic Quitkin
New York State Psychiatric Institute
Patients having both a major depressive disorder (MDD) and an anxiety
disorder have the opposite direction of alpha asymmetry at posterior
sites when compared to patients having a MDD alone (Bruder et al.,
1997, Biol. Psychiatry, 41, 939-948). The present study examined whether
the effects of comorbid anxiety on regional brain asymmetry extend
to patients having a dysthymic disorder (DD). Resting EEG (3 min eyes
open and closed) was recorded from 30 electrode sites in 63 unmedicated
outpatients having either a MDD (n=43) or a DD (n=20). Patients were
subgrouped by comorbidity, with some patients having only a depressive
disorder (MDD=26; DD=12) and others having both a depressive and anxiety
disorder (MDD=17; DD=8). An ANOVA of log alpha power for patients
in these groups revealed a significant difference in alpha asymmetry
between patients having an anxious depression and those having a nonanxious
depression (Group X Hemisphere interaction, p=.005), but no difference
between patients having a diagnosis of DD versus MDD. Patients having
an anxious depression showed a general pattern of greater activation
(less alpha) over right than left hemisphere sites, whereas patients
having a nonanxious depression showed less activation (greater alpha)
over right posterior sites (Group X Hemisphere X Region interaction,
p=.01). These findings indicate that the abnormal pattern of alpha
asymmetry in major depression and the opposing effects of comorbid
anxiety disorder on alpha asymmetry at posterior sites are also present
in dysthymia.

ERPs measure phonological recoding during silent reading involving homophones and non-words
Kelly A.K. Forbes & John F. Connolly
Dalhousie University
The degree to which phonological codes influence the word identification
process has been debated. Event-related brain potentials (ERPs) have
been used to investigate the processes involved in language. The N400,
a negative-going brainwave component, has been shown to be sensitive
to semantic priming and occurs reliably to semantically incongruous,
and therefore unexpected, terminal words of sentences. The phonological
representations of terminal words for highly contextually constrained
sentences were manipulated to observe the effects on the N400 response
Terminal words were semantically congruous or incongruous to the sentence
context. Incongruous endings were phonologically identical to or different
from the expected endings. Phonologically expected endings were either
words (i.e., homophone foils) or non-words that shared their phonological
representations with the semantically congruous homophone (i.e., pseudohomophone)
or non-homophonic word (i.e., pseudofoil). Phonologically unexpected
endings were words that were semantically incongruous to the sentence
context. Results indicated that the N400 responded differentially
to incongruous endings depending on whether they were phonologically
expected. Phonologically unexpected endings produced significantly
larger N400 amplitudes than did phonologically expected words and
non-words. N400 amplitudes for phonologically expected words and non-words
did not differ from each other. However, phonologically expected stimuli
produced significantly larger N400 responses than did the congruous
terminal words. These findings indicate that phonological codes of
both lexical and non-lexical stimuli influenced the word identification
process. In addition, the results suggest that the magnitude of the
phonological effect was influenced by the use of non-word stimuli.
Results are discussed in terms of current reading models.

P300 amplitude in schizophrenics ranging in clinical severity
J. M. Ford1,2, D. H. Mathalon1, L. Marsh1, W. Faustman2, D. Harris3, A. L. Hoff3, M. Beal3, K. Minn3, K. O. Lim1,2, & A. Pfefferbaum1,4
1Stanford University School of Medicine, 2Veterans Affairs Palo Alto Health Care System, 3Napa State Hospital, 4SRI International
Although P300 of the event-related brain potential (ERP) is reliably
reduced in patients with schizophrenia, its relationship to neuroanatomical
features seen in magnetic resonance images (MRI) and to various clinical
variables has been difficult to establish perhaps because of the restricted
data ranges in many samples. We report ERPs collected from two groups
of schizophrnenics (DSM-IV) from the Palo Alto Veterans Affairs Health
Care System (VA) and Napa State Hospitals (NSH) varying in clinical
severity (Brief Psychiatric Rating Scale=37.8 vs. 54.8), parietal
lobe gray matter volumes (46.4 vs. 42.8 cc), and age at onset (22.6
vs. 16.8 years) Severely ill chronically- hospitalized schizophrenic
NSH patients (n=25, age=34.8 years), moderately ill acutely-hospitalized
VA patients (n=15, age=37.1 years), and age-matched healthy men (n=33,
age=36.5 years) were compared on ERP oddball paradigms designed to
elicit attention automatically (loud noises) and effortfully (target
detection) (probability(rare)=.20; probability(frequent)=.80). Compared
to controls, both groups of patients showed reductions in both P300s;
NSH patients had an additional reduction in effortful P300. Although
the NSH patients were younger at disease onset, tended to have smaller
volumes of parietal cortical gray matter, and had more clinical symptoms,
only clinical severity accounted for the group difference in effortfu
attention. It is worth noting that in both groups, effortful P300
was related to parietal lobe gray matter volumes. Supported by: Department
of Veterans Affairs, National Institute of Health (MH30854, MH40052),
Norris Foundation, NARSAD, Stanley Foundation, and California Dept.
of Mental Health.

The relationship between anxiety, alcohol consumption, and ERP indices of memory in social drinkers
Allison M. Fox, Alistair M. Lethbridge, Susan Bresnahan, & Jason Bruggeman
University of Wollongong
When subjects memorize wordlists using elaborative learning strategies,
the amplitude of a Frontal Positive Slow Wave (FPSW) elicited by the
words predicts subsequent memory performance, leading to the suggestion
that the FPSW reflects the formation of inter-item associations. It
has been suggested that heavy social drinkers demonstrate reduced
abstracting and organizational abilities relative to light social
drinkers and, consistent with this view, they elicit a reduced amplitude
FPSW memory effect. These results were interpreted as suggesting that
heavy drinkers fail to integrate as much information during elaborative
learning conditions. However, one potentially confounding variable
linked to performance in research examining the effects of social
drinking on cognitive functioning is anxiety. Increased anxiety has
been associated with both increased alcohol consumption and reduced
working memory capacity, particularly during effortful tasks. The
present study examined the relationship between anxiety and alcohol
consumption in a verbal memory task. Subjects were instructed to memorize
wordlists by forming stories or pictures integrating the words. ERPs
elicited by words were recorded during the learning phase and sorted
according to subsequent memory performance. Consistent with previous
research, light drinkers showed an enhanced FPSW for subsequently
recalled words relative to those words which were not recalled and
this effect was absent in the heavy drinkers. The amplitude of the
FPSW memory effect was correlated with average quantity of alcohol
consumed after statistically controlling for self-reported anxiety
levels, providing further support for a relationship between alcohol
consumption and elaborative memory processes in social drinkers.

Assessing parental history of hypertension: Father (and mother) knows best!
Christopher R. France & Gary D. Page
Ohio University
Researchers interested in examining parental history of hypertension
effects often establish parental blood pressure history via interview
or questionnaires administered to offspring. This method is simple
and less costly than contacting parents, and has been reported to
provide greater than 90% accuracy when compared with written parental
or physician confirmation. The present study examined the two components
that contribute to overall accuracy in this assessment: sensitivity
(correct identification of high blood pressure) and specificity (correct
identification of normal blood pressure). Written parental blood pressure
history information was requested from undergraduate volunteers who
were in contact with their biological parents. Participants were asked
to indicate whether each parent had a history of high blood pressure
by responding yes, no, or don't know. Blood pressure history questionnaires
were then mailed to each parent. The following results were obtained
from 512 participants (293 women, 219 men) who did not select don't
know for blood pressure history for either parent, and whose parents
returned completed blood pressure history questionnaires. Consistent
with previous reports, we obtained an overall accuracy of group assignment
of 86% across participants. However, this reasonably high accuracy
rating was due to a combination of high specificity (92.6%) and low
sensitivity (69.2%). These findings suggest that reliance on offspring
reports results in an unacceptable degree of error with regard to
identification of offspring of hypertensives. Therefore, despite the
higher cost and effort involved, direct contact with biological parents
may be necessary to accurately assess parental history of hypertension.

Vagal tone moderates the undoing effect of positive emotions
Barbara L. Fredrickson
University of Michigan
The undoing model of positive emotions posits that positive emotions
speed recovery from the cardiovascular sequelae of negative emotions.
The present study examined whether parasympathetic vagal tone, as
measured by respiratory sinus arrhythmia (RSA), moderates the undoing
effect of positive emotions. Ninety-five participants (50% women)
prepared a self-relevant speech under time pressure, a task that reliably
elicits anxiety. Following this speech task, participants were randomly
assigned to view a film that elicited (a) contentment, (b) amusement,
(c) neutrality, or (d) sadness. Seven cardiovascular indices were
monitored continuously: RSA, heart rate, finger pulse amplitude, pulse
transmission times to the finger and ear, systolic and diastolic blood
pressure. The speech task elicited reactivity on all seven cardiovascular
measures. Duration of cardiovascular reactivity (CVR) was defined
as the time elapsed from film onset until CVR returned to within each
subject's own baseline range. A median-split on mean baseline RSA
divided Ss into Low and High Vagal Tone. A 2 X 4 ANOVA (Vagal Tone
X Film) yielded a main effect for Vagal Tone (F(1,87) = 5.59, p <
.05) and a two-way interaction (F(3,87) = 2.95, p < .05). On average
those with Low Vagal Tone showed longer-lasting CVR than those with
High Vagal Tone. Single-df comparisons revealed that the undoing effect
of positive emotions (i.e., shorter CVR when viewing positive films
compared to viewing a neutral or sad film) emerged for Low Vagal Tone
Ss only. Findings suggest that persons with low vagal tone may benefit
most when using positive emotions to regulate negative emotions.

Neural correlates of involuntary attentional shifts in young and old subjects. An event-related potential approach
Helen Gaeta, David Friedman, Walter Ritter, & Jeff Cheng
New York Psychiatric Institute
Involuntary shifts in attention to irrelevant stimuli were studied
in old and young subjects during a dichotic-listening task. Event
related potentials (ERPs) and behavioral measures were recorded. Pairs
of tones, with two different stimulus onset asynchronies (SOA, 200ms,
560 ms) were presented. Subjects ignored tones presented to the left
ear and attended tones presented to the right ear. Left ear tones
were a frequent standard (700Hz, p=.88), a small (650Hz, p=0.06) and
a large (500, p=0.06) deviant. Right ear tones (1500 Hz) were presented
with two equiprobable intensities. Subjects responded to the lower
intensity stimulus. It was predicted that unattended, large deviant,
left ear stimuli preceding target stimuli in the right ear would cause
involuntary attentional shifts. In line with prediction, large, unattended
deviants preceding attended targets caused impaired performance at
the short SOA that was greater for old subjects, who also showed impaired
performance at the long SOA. The mismatch negativity (MMN), indicative
of automatic change detection, and P3, indicative of active attention,
were elicited by large deviant stimuli which were associated with
impaired performance. P3 latency to large deviants was prolonged for
old subjects at both SOAs. Current source density analyses of the
MMN to these large, unattended deviants showed frontal and temporal
foci. These data suggest that a frontal lobe process associated with
the MMN plays a role in involuntary attentional shifts. Prolongation
of active attention induced by involuntary attentional shifts (indexed
by P3 latency) in older subjects can account for their impaired performance
at the long SOA.

Age-related changes in mismatch negativity elicited by tonal and non-tonal deviants
Helen Gaeta, David Friedman, Walter Ritter, & Jeff Cheng
New York Psychiatric Institute
The mismatch negativity (MMN) is an event related potential componen
that has been shown to reflect pre-attentive processing of stimulus
features and sensory memory processes, and to play a role in involuntary
attentional capture. MMNs to small and large deviant tones and to
non-tonal deviants (environmental sounds) were recorded from young
and old subjects during an ignore `oddball' task. For young subjects,
MMNs were elicited by both types of tonal deviants, with shorter latency
and greater amplitude for the large tonal deviant. For old subjects,
MMNs were also elicited by the larger, but not the smaller, tonal
deviant, which were smaller in amplitude than those for the young
subjects. To this extent, the MMN showed a graded response for both
young and old subjects. For both young and old, MMNs and robust P3s
were elicited by non-tonal deviants, but there were age-related differences
in P3 scalp distribution. For young subjects there was a central maximum,
whereas for old subjects there was a fronto-central maximum. In addition,
old subjects showed a posteriorly distributed slow positive shift.
Current source density (CSD) analyses of the MMNs showed bilateral
frontal foci and a right hemisphere temporal focus for young subjects
but not for old subjects. These results suggest that sensory memory
may be impoverished with aging resulting in reduced efficacy of pre-attentive
processing of auditory stimuli. However, the neural processes underlying
involuntary attentional capture are relatively preserved with aging,
possibly with an elevated threshold.

Amplitude changes in the brainstem frequency-following response during selective attention
Gary Galbraith, Bryan Chae, Jason Cooper, Mark Gindi, Timothy Ho, Benny Kim, & Diana Mankowski.
University of California at Los Angeles
Auditory perceptual asymmetries are presumably mediated in "higher"
neural centers, especially the cortex, due to a neuroanatomical dominance
of contralateral ear-to-hemisphere connections, and structural and
functional hemispheric asymmetries. Yet, descending efferent pathways
may modulate signals at all levels, including primary hair cells.
There is controversy, however, concerning the extent of such efferent
control in humans. We have previously reported attention effects on
the human brainstem frequency-following response (BFFR). In the present
study we show attention effects on BFFRs evoked by linguistic stimuli
(natural vowels). Twelve male and 12 female subjects heard the vowels
/a/ and /e/ presented simultaneously to the right ear. Stimuli were
identical in duration and rms amplitude. Frequency analysis of the
stimuli showed non-overlapping spectral peaks, permitting identification
of unique vowel components in the BFFR. Eighteen counterbalanced trial
blocks were divided among three tasks: (1) attend /a/, (2) attend
/e/, and (3) read text material (control). After each trial block
subjects rated how well they succeeded in attending to the task. Repeated
measures ANOVA showed a significant (F[2,40] = 8.45, p = 0.0009) Gender
x Rating x Attention x Vowel interaction. Subjects with the highes
ratings (above the median) showed larger amplitudes in the control
condition. Males in this group showed larger amplitudes to attended
vowels and smaller amplitudes to ignored vowels, while females showed
the opposite pattern. These results indicate that the BFFR reflects
subject differences and properties of linguistic coding, as well as
attention effects at the level of the human brainstem.

Startle modulation research requires simultaneous recording of both the orbicularis oculi electromyogram and the vertical electro-oculogram
Jean-Guido Gehricke, Edward M. Ornitz, Dale Song, & Patrik Gabikian
University of California-Los Angeles
In startle modulation research the orbicularis oculi electromyogram
(OO-EMG) has replaced the lever arm potentiometer and the electro-oculogram
(EOG) as the measure of reflex blinks. In this report, we document
the advantages of recording a vertical d.c. EOG along with OO-EMG.
There are at least two types of non-reflex eyelid movements that can
be confounded with reflex blinks to startling stimuli: spontaneous
blinks and lid saccades. The OO-EMG and a magnetic search coil provide
precise temporal and amplitude relationships for these various types
of eyelid movements but require absolute fixation of head and eyes
within a magnetic field (Evinger, Manning, & Sibony., 1991, Invest.
Ophthalmol. Visual Sci., 32(2), 387-400). Our study provides quantitative
data from OO-EMG and EOG recordings in 15 freely moving children,
aged 7 to 11, during an auditory startle experiment. The easily recorded
EOG shows the same relationships between lid position, velocity, and
duration of startle blinks as does the magnetic search coil. We report
examples of OO-EMG responses in which a simultaneous recording of
EOG revealed spontaneous blinks and lid saccades that would be misclassified
as reflex blinks if OO-EMG were used alone. It is concluded that spontaneous
blinks and lid saccades can be misidentified as reflex blinks in OO-EMG
recordings and that simultaneous recording of the EOG is necessary
to prevent such misclassification, provides a description of lid movement
mechanisms similar to that provided by magnetic search coils and can
be used with sensitive or less cooperative subjects, e.g. children,
in experiments requiring subject activity.

Sensory and cognitive shifts in olfactory event-related potential components across the lifespan
Mark W. Geisler1,2, Charlie D. Morgan3, James W. Covington2, Spencer Wetter2, Dennard Ellison2, & Claire Murphy1,2
1University of California School of Medicine, 2San Diego State University, 3SDSU/UCSD Joint Doctoral Program in Clinical Psychology
The objective of the present study was to assess the relationshi
between sensory and cognitive odor processing across the lifespan.
Sensory and cognitive odor information processing was measured with
olfactory event-related brain potentials (OERP's). Eighty subjects
from seven age groups (20-, 30-, 40-, 50-, 60-, 70-years of age),
with approximately equal numbers of males and females in each decade,
were tested. The amplitudes and latencies from the N1, P2, and P3
components were elicited in a single stimulus paradigm with an inter-stimulus
interval of 60 seconds using amyl acetate as the odorant. OERPs were
recorded monopolarly at the Fz, Cz, and Pz electrode sites and EOG
activity was monitored at the left eye, while subjects estimated the
intensity of the odor stimulus. The preliminary results showed a slowing
in sensory odor processing speed of approximately 20 ms per decade
as indexed by the N1 latency. The allocation of attentional resources,
as measured by the P3 amplitude during the odor information processing
task, remained relatively constant for the first three decades and
was followed by a downward shift of approximately 5 µV. The reduction
in P3 amplitude remained relatively constant for the last three age
groups. These results support previous findings that the ability to
process odors decreases with age. The results suggest that this decline
across the lifespan may be due to both a slowing of sensory function
and a decrease in cognitive resources. Supported by NIH grant DC02064

Going, going, gone: Habituation of sympathetic influences on cardiac responses to aversive auditory stimuli
Peter J. Gianaros & Karen S. Quigley
The Pennsylvania State University
There are unresolved questions about the extent to which cardioacceleratory
responses habituate to aversive auditory stimuli. Several studies
have suggested that a long-latency cardiac acceleration (peak at approximately
30 s) to an intense auditory stimulus is virtually absent after one
trial. This rapid habituation contrasts with prior theoretical accounts
of non-habituating defensive responses. In addition, Fernandez & Vila
(1989, Biol. Psychol., 28, 123-133) have reported that the autonomic
mediation of this long-latency cardioacceleration appears to be driven
primarily by sympathetic input to the heart. However, the habituation
of this autonomic component of the cardiac defense response has not
been investigated previously. In the present study, participants were
presented with twelve 100 dB bursts of white noise (20 ms rise with
a 60-90 s ISI) while heart rate and pre-ejection period (PEP), a noninvasive
estimate of sympathetic activation of the heart, were obtained using
impedance cardiography. Preliminary analyses revealed a robust and
statistically significant decrease in PEP, suggestive of increased
sympathetic activation, following stimulus onset during early trials.
Moreover, this decrease in PEP occurred at the approximate peak latenc
where cardiac defensive responses are typically observed. The decrease
in PEP habituated rapidly across trials and was absent by trial five.
These data are consistent with prior research that has shown both
sympathetic mediation and rapid habituation of the long-latency cardiac
response to aversive stimuli. These results extend prior research
by revealing the time course of sympathetic habituation to aversive
auditory stimuli.

A comparison of finger plethysmograph to EKG in the spectral analysis of heart rate variability
Nicholas D. Giardino1, Paul M. Lehrer2, & Robert Edelberg2
1Rutgers University, 2UMDNJ-Robert Wood Johnson Medical School
Periodic variability in heart rate is widely used as a non-invasive
measure of cardiac autonomic control. Variance within high (0.12 -
0.40 Hz) and mid (0.05 - 0.12 Hz) frequency spectra are thought to
reflect parasympathetic and parasympathetic/sympathetic activity,
respectively. A review of the research literature reveals that both
EKG and finger plethysmograph are used to collect data from which
interbeat intervals and heart rate are derived for use in the calculation
of frequency band variance. To examine the adequacy of finger plethysmograph
in providing for accurate heart period determinations we simultaneously
recorded EKG and finger plethysmograph from healthy subjects. Interbeat
interval (IBI) time series were created by sampling cardiotachometer
records triggered by R-spikes for EKG and pulse-peaks for finger plethysmograph.
IBI time series from each collection method were then analyzed using
an algorithm for band-pass filtering of IBI data which applies a moving
polynomial filter to remove aperiodic trends in the data set and then
calculates the natural logarithm of the heart period variance within
the selected frequency bands. A high correlation was found between
EKG- and finger plethysmograph-derived band variance for mid frequency
(r = 0.91), but not high frequency (r = 0.68) bands. These results
suggest that finger plethysmography is not an appropriate method for
data collection when high frequency variability in heart rate is of
interest, but may be acceptable for lower frequency bands.

Anger suppression and fighting spirit: Associations with endocrine and immune function in metastatic breast cancer patients
Janine Giese-Davis, Sandie Sephton, Ron Duran, & David Spiegel
Stanford University School of Medicine
Emotional adjustment may affect the course of cancer progression.
We examined the links between stress as measured by salivary cortisol,
antigen-specific cell-mediated immunity, and two coping strategie
(anger control (Courtauld Emotional Control Scale) and fighting spirit
(Mental Adjustment to Cancer), as a possible mechanism. 49 advanced
breast cancer patients, recruited to study the effects of group psychotherapy
on cancer survival, provided baseline measures of coping, cortisol,
and responses to skin test antigens (Multitest CMI). These were related
to cortisol and CMI follow-ups. Patients reporting more anger control
(AC) at baseline had smaller responses to skin test antigens ( r =
-.40, p<.005). Patients with higher baseline AC had higher mean cortisol
4 months later (high AC baseline vs. 4 months: t = -1.61, p = .06;
high AC vs. low AC at 4 months: t = -2.19, p < .02) and lower responses
to antigens at baseline and after 1 yr (high AC at 1 yr vs. low AC
at baseline: t = 2.76, p = .05; high AC at 1 yr vs. low AC at 4 months:
t = 3.50, p < .005). Patients with low baseline fighting spirit (FS)
had higher mean cortisol 4 months later (low FS baseline vs. 4 months:
t = - 1.73, p < .05; low FS at baseline vs. high FS at 4 months: t
= 2.24, p < .02), but evidenced no effect of FS on responses to antigens,
either at baseline, or 1 yr later. Thus, fighting spirit and anger
control both predicted subsequent cortisol levels, but only anger
control predicted subsequent immune responses.

Effects of personalized emotional activation on laterality and anterior-posterior balance of EEG activity
David G. Gilbert, F. Joseph McClernon, Susan M. Labott, & Shirin Sarkari
Southern Illinois University
It was hypothesized that induced anger and sadness, relative to induced
relaxation would decrease right- relative to left-frontal EEG alpha
power and would increase frontal relative to posterior midline theta
(MT) power. During a single session, eyes closed EEG recordings were
made during three conditions (counterbalanced order): calm/relaxed,
sadness, and anger. Between conditions, participants (N = 17) completed
a series of verbal and visual tasks designed to induce a neutral emotional
state. Emotional inductions were two-minute audio recordings based
on the participant's earlier-written descriptions of situations/events,
thoughts, sensations, and actions. Recordings were obtained from 19
sites using a NeuroScan system, but present analyses are limited to
a priori site comparisons noted below. Consistent with our second
hypothesis, induced anger and sadness, relative to induced relaxation,
increased frontal (Fz) relative to posterior (Pz) MT power (P<0.02);
however, contrary to our first hypothesis there was no significant
effect of Mood on right- relative to left-frontal EEG alpha power
(log right-log left) for F4-F3 or for F8-F7. Failure to observe frontal
EEG asymmetries cannot be attributed to a failure to induce moods
since the inductions produced significant increases (Ps <0.01) in
corresponding moods. These results suggest that future investigations
of EEG correlates of affect would benefit from the analysis of anterior-posterio
gradients in MT. This is especially true given that frontal MT increases
during memory and cognitive tasks, and that such activity may be associated
with cingulate gyrus and dopaminergic activity associated with certain
types of emotional processing.

The covert-to-overt threshold for facial actions: An EMG study
Eric Girard1, Louis G. Tassinary2, Arvid Kappas1, Pierre Gosselin3, & Daniel Bontempo2
1Laval University, 2Texas A&M University, 3University of Ottawa
For almost two decades researchers have used either electromyography
(EMG) or the Facial Action Coding System (FACS) to quantify emotional
reactions and communications emanating from the face. Surprisingly,
these two techniques have been compared directly in only one unpublished
study using two highly skilled participants (Ekman, 1982). EMG and
FACS were found to be highly correlated for clearly visible movements
and EMG alone was found to measure reliably non-visible and barely
visible changes in muscle tension. By implication, the relationship
between the two techniques was portrayed as both monotonic and invariant.
Video records from 12 of 14 participants who participated in an EMG
study of facial muscle control were coded with FACS, and the relationship
between EMG activity in the brow region and FACS coded onset and offset
of Action Unit 4 (brow lower) were examined for 2.5 s poses with either
linear or nonlinear trajectories. Overall, movement onsets were associated
with EMG levels over 3 times higher than those associated with movement
offsets (69.2 vs. 20.4 µVs; F(1,66) = 43.57), and this pattern was
observed in every participant. In addition, there was a nonsignificant
tendency, F(2,66) = 2.38, for higher EMG levels to be associated with
abrupt versus gradual onsets (89.8 vs. 58.8 µVs), as well as with
abrupt versus gradual offsets (21.97 vs. 17.4 µVs). These results
clearly challenge conclusions of invariance and monotonicity, although
additional research is needed to determine whether the observed hysteresis
is in the skin of the encoder, the eye of the decoder, or both.

Validation of the Polar heart rate monitor for measuring heart rate responses to mental stress
J. L. Goodie, S. Schauss, K. T. Larkin, & B. Aragona
West Virginia University
The Polar heart rate monitor (PHRM) provides a method of continuously
monitoring heart rates (HR). The monitor consists of a transmitter
worn around the chest that collects and sends heart rate data to a
specially designed wristwatch. We examined the validity of the PHRM
to measure within-subject changes of HR between tasks and between
subject HR differences. Thirty individuals participated in a 1-hou
session. Heart rates were taken simultaneously by the PHRM and electrocardiography
(ECG) during a hand-grip exercise and a mental arithmetic task, each
preceded by a 4 minute resting period. Within-subject correlations
between the two devices were significant (mean r = .980, t (29) =
20.60, p < .001). All correlations, except for two subjects, exceeded
r = .90. Mean HRs between subjects were virtually identical for baseline
1 (Polar M = 74.08, ECG M = 73.75, r = .998), the hand-grip task (Polar
M = 81.08, ECG M = 81.10, r = .996), baseline 2 (Polar M = 71.92,
ECG M = 72.59, r = .998), and the mental arithmetic task (Polar M
= 80.40, ECG M = 79.74, r = .994). The small differences in observed
mean HRs from the PHRM and ECG, and high correlations of HRs obtained
by both devices, suggest that the PHRM is appropriate for monitoring
HR during stationary laboratory tasks. Future studies should examine
the validity of the PHRM in ambulatory settings.

Smooth pursuit performance in questionnaire-identified schizotypes
Diane C. Gooding, Kirstie K. Danielson, and Thomas R. Kwapil
University of Wisconsin-Madison
The biobehavioral high-risk strategy has not been exploited as a methodology
for understanding how schizophrenia develops. A combined approach
of psychometric and psychophysiological screening has the potential
to improve identification and prediction of subjects at specific risk
for schizophrenia. Prior research (Simons and Katkin, l985, Psychiatry
Research, 275-289) supports the feasibility of identifying a subset
of subjects with high scores on the Chapman psychosis-proneness scales
who also display deviant smooth pursuit. The present experiment was
an attempt to replicate and extend these earlier findings. Undergraduates
were administered the Chapman scales. Subjects identified by Perceptual
Aberration-Magical Ideation (Per-Mag) scores (n=58), Physical Anhedonia
(PhysAnh) scores (n = 19) as well as normal controls (n = 68) underwent
psychophysiological testing and clinical interviewing. Subsequent
asessments were performed by independent investigators who were naive
to group membership. Pursuit during a simple non-monitor tracking
task was measured using root-mean-square (RMS) error scores. Subjects
were administered a clinical interview assessing Axis I symptomatology
as well as psychotic-like experiences. Although the groups did not
differ significantly in terms of their RMS scores (F (2,97) = 0.48,
n.s.), they differed in terms of their variability in performance.
Pairwise F tests on the variances revealed that the Per-Mag subjects
were more variable than the controls, (F = 2.50, df =57,67; p < .01)
and the PhysAnh subjects (F= 2.50, df =57,18; p < .05). Subjects reporting
physical anhedonia and control subjects were equally variable (F =
1.0, df =67,18; n.s.). Examination of a hypothesized association between
pursuit and psychotic-like experiences is underway.

Nailfold capillary plexus visibility in psychosis-prone college students
Diane C. Gooding & Meghan D. Miller
University of Wisconsin-Madison
One strategy for investigating the developmental precursors of schizophrenia
is to longitudinally study individuals believed to be at heightened
risk for psychosis. A two-step screening process, in which the psychometric
high-risk method is combined with a biobehavioral high-risk approach,
may improve the prediction rate of individuals at heightened risk
for the manifestation of schizophrenia. Studies (c.f. Clementz, Iacono,
Ficken, & Beiser, l992, Biological Psychiatry, 378-390) indicate that
patients with schizophrenia and/or schizophreniform disorder show
elevated rates of plexus visibility, though other psychotic patient
groups do not. In this project, we identified psychosis-prone individuals
on the basis of their responses to the Chapman scales. Students who
scored deviantly on the Perceptual Aberration and/or Magical Ideation
scales (n = 47) were rated for the presence of nailfold plexus visibility.
Comparison subjects (n = 42), who were free of personal and family
history of psychosis, were also rated. The visibility scores used
in data analyses were calculated by summing over all ten fingers.
All assessments were made without knowledge of group status. Using
a cutoff score of 10, these subject groups did not differ significantly
in rates of plexus visibility, x2(1) = 0.01. These results suggest
that nailfold plexus visibility as it is currently rated may not be
a suitable biobehavioral screen for heightened vulnerability to psychosis.

Beyond ERP averaging: The benefit of concomitant SCR-ERP measures
Evian Gordon1, Chong Lee Lim1, Christopher Rennie1, Homi Bahramali1, Ilario Lazzaro1, Jim Lagopoulos1, Albert Haig1, Wai Man Li1, Demitri Melkonyan1, & Jim Wright2
1Westmead Hospital-Sydney & The University of Sydney, 2Mental Health Research Institute of Victoria
The average of multiple single-trial Event Related Potentials (ERP)
is a soundly based method for determining the overall ERP. However,
it is possible that there are systematic physiological sub-processes
that are obscured by the average. One such process concerns the "Orienting
Reflex (OR)". Few studies have examined concomitant ERP-OR (skin conductance
responses or SCR) in late component ERP paradigms, with short interstimulus
intervals (ISI). Our group have developed a model to score overlapping
SCRs in such conventional "cognitive" paradigms (Lim et al., 1997
International Journal of Psychophysiology, 25:97-109). We present
a series of interrelated studies in 50 normal subjects, with simultaneous
ERP-SCR measures during a conventional auditory oddball paradigm (IS
1 sec; 40 target stimuli probabilty of 15%; 19 sites). Target ERPs
with associated SCRs showed larger P300 amplitude (F=4.52;p=0.039),
earlier P300 latency (F=10.52;p=0.002) and earlier N200 latency (F=13.81;46;p=0.001)
than target ERPs without associated SCRs. Within-group cluster analysis
of SCRs showed that the first 10 target SCRs clustered differently
from the subsequent 30 SCRs. Sub-averages of the first 10 ERPs (and
pre-stimulus EEG activity) showed significant differences compared
with subsequent ERPs, namely; decreased N200 amplitude (F=24.60;p<0.0005)
and decreased pre-stimulus alpha EEG activity (F=31.99;p<0.0005).
Examination of the dynamics of the CNS-ANS activity across the trial,
showed a systematic decrease in SCL and SCR, increased N200 ERP amplitude
and pre-stimulus alpha EEG, but consistent reaction time responses.
These results are compared with data from patients with ADHD, schizophrenia
and Parkinsons Disease.

Brain activity associated with retrieval from working memory: A comparison of ERPs and EROS data
Gabriele Gratton1, Monica Fabiani1, Jonathan K. Ho2, Marsha R. Goodman-Wood1, & M. Catherine DeSoto1
1University of Missouri-Columbia, 2Columbia University
Brain activity during a variable-set memory-search task was recorded
using event-related brain potentials (ERPs) and the event-related
optical signal (EROS). The memory set comprised two letters flashed
one to the left and one to the right of a fixation cross. After a
2-s interval a test stimulus was flashed centrally. This stimulus
could be the memory-set stimulus presented to the left (OLD LEFT),
the memory-set stimulus presented to the right (OLD RIGHT), or a NEW
stimulus. ERPs were recorded from 7 subjects from 18 scalp locations
covering the whole scalp. EROS was obtained in 4 subjects from 28
locations over the occipital area. To isolate the brain activity associated
with retrieval phenomena, the analysis of both ERPs and EROS focused
on the activity elicited by the OLD test stimuli that was systematically
lateralized as a function of the position of the stimulus at encoding
-- that is, on the comparison between OLD LEFT and OLD RIGHT activity
for right and left recording locations. This analysis showed that
both ERPs and EROS are characterized by a systematic lateralization
in the interval between 250 and 350 ms after stimulation. The EROS
data show that this activity is evident in lateral occipital areas,
indicating an extrastriate origin. These data indicate that: (a) extrastriate
cortex is involved in the retrieval of information from working memory;
(b) visual stimuli presented to one hemifield leave traces that are
stronger on the contralateral hemisphere; and (c) EROS can help understanding
the brain origin of ERP signals.

Low sense of mastery and increased 24-hour ambulatory blood pressure in postmenopausal women
Karen Grewen, Susan S. Girdler, Sheila G. West, Jean Ranc, Jeannie Koo, Alan Hinderliter, Monica Adamian, & Kathleen C. Light
University of North Carolina at Chapel Hill
Past research has shown that several psychosocial variables predict
development of hypertension. The present study examined how cognitive
appraisal of the ability to deal with life stressors might influence
perceptions of stress and impact blood pressure (BP) responses. It
was hypothesized that negative appraisals of coping ability (assessed
by low scores on the Pearlin Mastery Scale) would result in increased
perceptions of stress (Perceived Stress Scale)which would contribute
to increased BP during daily activities. Subjects were 55 naturally
or surgically post-menopausal women, aged 40-68 years, who were not
on hormone replacement therapy. Ambulatory BP was recorded over a
24-hour period as subjects went about their normal daily activities.
Mean SBP and DBP scores w ere calculated for day, evening, and sleep
periods. Subjects were divided into low and high mastery groups (lowest
25% vs. highest 75%). As predicted, low mastery was related to higher
levels of perceived stress (r = -0.68, p < .001). Also, low mastery
subjects had significantly higher ambulatory SBP levels during day,
evening, and sleep periods, and significantly higher DBP levels during
day and evening periods (p<.05). Subjects with low mastery had mean
daytime SBP within the range considered to be borderline hypertensive
compared to subjects with high mastery who had mean daytime SBP within
the normal range (140 vs.124 mm Hg, p<.0001). Thus, in postmenopausal
women, low sense of mastery is related to greater perceived stress
and to increased BP during daily life.

Potentiated startle in children at risk for anxiety disorder
Christian Grillon, Lisa Dierker, & Kathleen R. Merikangas
Yale University School of Medicine
A family history of anxiety disorders is the most consistent and potent
risk factor for the development of this condition. Theoretically,
children at-risk for anxiety disorders are believed to exhibit increased
fear and anxiety to stressful stimuli or situations. The present investigation
used the startle reflex to examine affective reactions to threatening
stimuli in 12-20 years old children and adolescents at high and low
risk for anxiety disorders by virtue of a proband diagnosis of these
disorders. The experiment consisted of an initial startle adaptation
period followed by a potentiated startle procedure. During that procedure,
acoustic startle probes were delivered under threat and safe conditions.
An unpleasant intense airpuff directed to the larynx could be delivere
in the threat but not in the safe conditions. In females, startle
magnitude was significantly elevated throughout the experiment in
the high-risk, compared to the low risk group. However, the magnitude
of potentiated startle to the threat signal did not significantly
differ between the two groups. In contrast, in males, "baseline" startle
did not significantly differ between the two groups, but the magnitude
of startle potentiation to the threat signal was significantly greater
in the high-risk, compared to the low-risk group. These findings suggest
that individual differences in startle potentiation may serve as a
vulnerability marker for the development of anxiety disorders. Results
will be discussed in light of recent findings suggesting that aversive
responses to explicit (threat signal) and contextual (experimental
room) cues are separate processes mediated by different brain structures.

Effects of safety signals on explicit and contextual fear-potentiated startle
Christian Grillon & Brian Harel
Yale University School of Medicine
When subjects are verbally informed to anticipate shocks during threat
but not during safe conditions, startle is increased during the shock
anticipation periods. Startle can also be increased by contextual
stimuli (e.g., the experimental room). Shock electrodes are likely
to be stressful contextual cues. Their presence might increase anxiety.
One of the aims of this study was to examine whether startle was potentiated
by the shock electrodes. If the presence of shock electrodes increases
startle, a critical question is whether the safe signal affects this
potentiation. In other words, is startle different during safe conditions,
compared to signal-free periods? Another aim of the study was to answer
this question. Acoustic startle probes were presented during threat
and safe conditions signaled by different colored lights (12-sec duration),
and during signal-free periods. Subjects were presented with four
separate blocks containing four safe and four threat conditions, and
signal-free periods. The shocks electrodes were attached or removed
on alternative presentations of these blocks. Subjects were told to
anticipate unpleasant shocks during the threat but not during the
safe conditions. Startle was potentiated by the presence of the shock
electrodes (greater startle in the safe conditions with the shock
electrodes present versus absent), suggesting that there is no true
baseline startle during an experiment where shocks are administered.
Startle was also reduced in the safe conditions, compared to the signal-free
periods, suggesting that the safe signal was a general inhibitor of

Bad news for the middle-aged: Your cardiovascular autonomic rhythms respond to stress as if you were eighty
Paul Grossman, Frank H. Wilhelm, & Lana Watkins
Harvard Medical School
Frequency-domain derived parameters of heart period and blood pressure
provide unique insights into cardiovascular autonomic modulation:
Respiratory sinus arrhythmia (RSA) mirrors oscillations of efferent
parasympathetic traffic to the heart. Ten-second rhythms of blood
pressure (Mayer Waves) reflect sympathetic, alpha-adrenergic effects
upon the vasculature. Cardiac vagal baroreflex control, estimated
from cross-spectral analysis, yields knowledge of autonomic short-term
blood-pressure regulation. We examined how age is related to these
autonomic indices during baseline and psychological stress conditions
(mental arithmetic and public speaking). Sixty normotensive men and
women, with no history of cardiovascular disease, were studied (ages,
21-91 years). Subjects were divided into 4 age groups (young, 21-34
years; middle-aged, 48-59 years; old, 60-70 years; and very old, 71-91).
With respect to all autonomic indices, ANOVA's (Age Group X Condition)
revealed main effects for age (p's < .002 - .00001), condition (p's
< .0001) and age X condition interaction (p's < .02 - .0006). Younger
individuals had consistently higher levels of RSA, Mayer-Wave amplitude
and baroreflex control. Only the young group manifested large RSA
and baroreflex changes from baseline to task; however, young subjects
showed no variation in Mayer-Wave amplitude across conditions. Conversely,
although old and very old subjects exhibited almost no change in RSA
or baroreflex activity across conditions, they did show pronounced
alterations in Mayer-Wave amplitude. Notably, the middle-aged pattern
of response across conditions was much more similar to old and very
old groups than to young subjects, i.e. pronounced variation in Mayer-Wave
amplitude across conditions, but little or no change in RSA or baroreflex
control. These findings suggest fundamental cardiovascular autonomic
adjustments may occur in middle age.

Individual differences in cardiac vagal baroreflex control are associated with the extent of cardiovascular responses to mental stress
Paul Grossman, Heather Ristuccia, Frank Wilhelm, & Lana Watkins
Harvard Medical School
The baroreflex comprises the nervous-system mechanism by which short-term
dynamic fluctuations of arterial blood pressure are controlled and
buffered. Therefore, intact baroreflex control should moderate the
abrupt surges of blood pressure evoked by psychological stressors.
Recent advances allow noninvasive assessment of spontaneous cardiac
vagal baroreflex control (CVBARO). We hypothesized that individual
differences in basal CVBARO are negatively associated with extent
of cardiovascular responses to mental stress. Subjects were 74 normotensive
men and women with no history of cardiovascular or psychiatric disorders
(age range, 21- 91 years). Stressors included a computerized menta
arithmetic task (10 min) and simulated public speaking (4 min preparation
and 3 min presentation). Tasks were preceded by baseline periods.
Respitrace respiration, Finapres blood pressure and the EKG were continuously
recorded. CVBARO was assessed from transfer-function analysis of beat-to-beat
blood pressure and heart-period signals (.07-.13 Hz). Both stressors
evoked large changes in heart rate, blood pressure and rate-pressure
product (p's < .001), although speech presentation produced greater
reactions (p's < .01). CVBARO was reduced from baseline only during
speech presentation. After adjustment for age and baseline levels,
heart rate, rate-pressure product and blood pressure reactions were
inversely associated with baseline CVBARO during speech presentation
and speech preparation (r's= -.25 to -.38; p's < .04), but not during
mental arithmetic. These findings suggest that individual differences
in resting CVBARO serve to buffer extreme cardiovascular responses
to mental stress (e.g. public speaking). Moderate cardiovascular stress
responses (e.g. the arithmetic task) may be more independent of the
arterial baroreflex control system.

Differential electrocortical processing of phobic stimuli in phobics and nonphobics
Ingmar Gutberlet, Wolfgang Miltner, & Katrin Vader
Friedrich-Schiller University of Jena
Spider and snake phobics have been shown to respond to the presentation
of their phobic objects with increased sympathetic nervous system
activity even if the stimuli are presented subliminally with backward
masking precluding the subjects from recognizing the stimulus content.
However, the question remains, at which level of central nervous system
control these effects are mediated. Thus, the present study was undertaken
to examine the effect of supraliminally presented slides on central
nervous processing in snake and spider phobics and in control subjects.
Out of a total of 64 subjects participating in the study 19 spider-
and snake phobics as well as 19 control subjects were selected for
data analysis based on their spider or snake anxiety scores (SPQ,
SNAQ). Subjects were shown 240 preevaluated slides of spiders, snakes,
neutral (fish, birds) and positive (young animals) objects. Slides
were presented for 1500ms with a randomized ISI of 8 to 12 seconds.
EEG data were recorded from 11 electrode sites for 2048ms with a Baseline
of 500ms and SAM ratings were obtained after each slide presentation.
The ERP data show significant differences in P3 and LPC amplitudes
for snake and spider phobics compared to control subjects for their
specific phobic objects while no such differences could be shown for
any other stimuli or group. Neither the electrocortical measures nor
the subjective ratings show signs of habituation throughout the experiment.
Thus, these data show stable phobia specific differences in cortical
processing that might possibly mediate the peripheral effects shown
in the literatur

Personality correlates of the P300 during execution of the visual CPT
E. Hafstad & Ch. Klein
University of Freiburg
Assuming a nosological relation between schizophrenia and schizotypy,
the present study investigated whether the CPT and P300 deficits observed
in schizophrenic patients can be found in schizotypal subjects as
well. From about 500 university students who filled in the German
translation of the Schizotypal Personality Questionnaire (Raine; SPQ-G),
34 subjects (61% male, age range 20-29 years) with SPQ-G-scores ranging
between 2 and 58 were tested with the visual CPT, using 7 different
letters as test items. Experimental task variation comprised different
levels of stimulus degradation and working memory load. Under each
condition, 250 items with 20% targets and 80% non-targets were presented
for 100 ms with a SOA of 1.0 s on a PC monitor. The EEG and EOG were
recorded from 32 channels with a DC amplifier (MES, Munich). Scalp
amplitude and current source density maps using spline interpolated
data were generated. Under each target condition, a P300 with peak
at Pz and! a slight left-parietal predominance was observed. Stimulus
degradation was associated with an amplitude reduction and latency
augmentation of the P300 peak, while high working memory load reduced
both peak amplitude and latency. Subjects with low, medium, or high
scores on the SPQ-G did not exhibit differences in the modulation
of that component. While the present results confirm the sensitivity
of the P300 to information processing, they suggest a normal visual
P300 in subjects with schizotypal personality features. Research was
supported by the Deutsche Forschungsgemeinschaft (DFG; Kl 885/4-1)

Hemispheric differences in P300 activity associated with changes in stimulus-onset asynchrony
Barry R. Haimson, Mark Pagliuca, and Marilyn Wasti
University of Massachusetts-Dartmouth
The purpose of the present study was to measure the influence of stimulus-onset
Asynchrony (SOA) on P300 activity at different electrode placements
in the right and left hemispheres during a divided visual field task.
Twenty-eight subjects were asked to match a centrally presented target
with a probe in the left visual field (LVF) or the right visual field
(RVF). Either a short stimulus-onset asynchrony (SOA) of 160 ms or
a long SOA of 2.05 s separated the target and probe. These findings
revealed that varying the SOA did influence both behavioral and ERP
measures of information processing and hemispheric asymmetry. Th
reaction time data indicated that stimuli in the LVF were processed
faster when the SOA was short, while information in the RVF was processed
faster with a long SOA. The ERP data indicated the relationship was
not quite so simple. The peak amplitude data suggested a greater processing
load in the right hemisphere with a long SOA and a left visual field
presentation. However, the peak latency data paralleled the reaction
time data and suggested faster processing in the left hemisphere.
Overall, the data revealed faster processing with a long SOA. These
findings suggest that a long SOA, which makes possible linguistic
encoding, facilitates processing in the right hemisphere. This could
mean that linguistic encoding associated with a long SOA primes perceptual
processing involving the right hemisphere. Such a conclusion is tentative
and requires further analysis of this data.

Voluntary activation of facial muscles alters the frontal asymmetry in EEG activity
Eddie Harmon-Jones & John J. B. Allen
1University of Wisconsin-Madison, 2University of Arizona
Past research has shown that the left and right frontal cortical regions
are involved in the experience and expression of approach and withdrawal-related
affect, respectively. A preliminary test of the idea that contracting
facial muscles used to form a smile as compared to a frown would alter
EEG activity in the left and right frontal regions was performed.
Participants (n = 6) wearing electrodes at left and right zygomatic
and corrugator muscle regions were asked to bring the sensors on their
cheeks toward their ears (or to bring the sensors on their brows together)
so that the equipment could be checked. Participants held each facial
pose for 1 min, then rested, and then held the other facial pose for
1 min. This sequence of poses was then repeated, resulting in 2 min
of zygomatic and 2 min of corrugator muscle region contractions (order
was counterbalanced). EEG activity was recorded at F3, F4, P3, and
P4 (referenced on-line to Cz) while these contractions were held.
Frontal asymmetry indexes were computed using log delta, theta, alpha,
and beta power (F4 minus F3). Results indicated that contractions
of the zygomatic as compared to corrugator muscle region caused greater
right than left frontal delta (p < .10), alpha (p < .10), and beta
power (p < .05). No significant differences between left and right
muscle regions emerged, suggesting that differences in lateralized
facial EMG activity did not account for the EEG effects.

The cardiac defensive response, anxiety, and central cholinergic systems
Sheri Hart, Martin Sarter, & Gary Berntson
The Ohio State Universi
Benzodiazepine receptor (BZR) agonists are prototypic anxiolytics,
whereas BZR inverse agonists appear to be anxiogenic. In accord with
its putative anxiogenic effects, we found that the BZR partial inverse
agonist FG 7142 (FG) augments the defensive-like cardioacceleratory
response to a moderately intense acoustic stimulus. This effect was
similar to one produced by an anxiogenic context, but may be limited
to those stimulus contexts that require higher level processing, as
it did not extend to startle responses nor startle potentiation by
a conditioned-fear CS. In this regard, FG is known to activate the
basal forebrain cortical cholinergic system, and the effects of FG
appear to depend on this cholinergic action. The effect of FG was
mimicked by intracerebroventricular (ICV) administration of the cholinergic
agonist charbachol, and was blocked by ICV atropine. FG effects were
also prevented by cholinergic-specific immunotoxic lesions (192 IgG-saporin)
of the basal forebrain (Berntson, Hart, Ruland, Sarter, 1996, Behavioral
Brain Research, 91-103). One target site of this projection is the
medial prefrontal cortex, which has been implicated in affective processes.
The present study observed an enhancement of the cardioacceleratory
defensive-like response, similar to that seen with FG, after local
administration of carbachol directly into the medial prefrontal cortex.
In the aggregate, these and additional findings suggest that basal
forebrain cholinergic cortical projections may be especially involved
in the cognitive aspects of anxiety.

The cognitive correlates of skilled performance in marksmen: An examination of brain activation profiles
Amy Haufler, Bradley Hatfield, & D. L. Santa Maria
University of Maryland-College Park
Log-transformed averaged alpha power (8-13 Hz) at T3, T4, P3, P4,
O1 and O2 in 15 male (n=13) and female (n=2) national-caliber marksmen
(M +- SD = 26.5 +- 11.1 years), was compared to that observed in 21
male novice shooters (23.1 +- 5.5 years) in the standing position
for the 6 sec prior to trigger pull and averaged over 40 such trials
yielding 6 successive 1-sec epochs. The temporal profiles were contrasted
to those observed during processing of verbal (Word Finding) and spatial
(Dot Localization) tasks, performed in the standing position, to determine
psychological state. The Cz-referenced EEG was amplified by a gain
of 50K using a band pass of 1-100 Hz with a 60-Hz notch filter applied.
Data were edited for EOG and subjected to a Hanning window with 10%
overlap before application of FFT. Marksmen were characterized by
higher and increasing power across epochs at T3 relative to that observed
at T4 which was stable. Novices were characterized by lower but increasing
power at T4 relative to that observed at T4 in the marksmen while
the magnitude and pattern at T3 were similar for the groups. Acros
T3 and T4, the marksmen displayed higher power. Temporal ratios for
the novices obtained during shooting were similar to those observed
during the spatial challenge and were heightened as compared to those
observed in the marksmen. These results are consistent with an interpretation
of less effortful processing in the visuo-spatial domain for the skilled
group (marksmen) as compared to the novice performers.

Validation of respiratory sinus arrhythmia and preejection period as indices of vagal and sympathetic activity in elderly women
Louise C. Hawkley, Mary H. Burleson, Kirsten M. Poehlmann, John M. Ernst, Gary G. Berntson, & John T. Cacioppo
The Ohio State University
Recent research has demonstrated that among young adult women respiratory
sinus arrhythmia (RSA) and pre-ejection period (PEP) reactivity measures
are valid and stable indices of vagal and sympathetic reactivity.
It is not known whether RSA and PEP remain useful indices of autonomic
activity among the elderly in whom autonomic activity is known to
be diminished, and in whom heart rate reactivity has been shown to
be decreased. To assess the validity of RSA and PEP as markers of
vagal and sympathetic activity in the elderly, 50 women ages 49 to
83 performed a speech task and mental arithmetic while cardiovascular
measures were obtained. As in the young, PEP reactivity was correlated
negatively with heart rate (HR) reactivity (r = -.403), RSA reactivity
was correlated negatively with HR reactivity (r = -.511), and PEP
and RSA were not reliably correlated. However, basal activity showed
a different pattern: as in the young, PEP and RSA were nonsignificantly
correlated and RSA and HR were negatively correlated (r = - .453);
contrary to the young, PEP and HR were also uncorrelated (r = -.063).
The inadequacy of PEP in indexing basal sympathetic activity in the
elderly may be attributed to the sensitivity of PEP to the effects
of increased preload and afterload, and to the measurement challenges
posed by morphologically indistinct impedance waveforms among the
elderly. Based on these results, RSA and PEP remain valid indicators
of vagal and sympathetic reactivity among the elderly, while basal
sympathetic activity may not be indexed by basal PEP.

Attentional modulation of startle in the schizophrenia spectrum
Erin A. Hazlett, Larry J. Siever, Monte S. Buchsbaum, Mimi Sevin, & Melissa Biren
Mount Sinai School of Medicine
A pathophysiological relationship between schizotypal personality
disorder (SPD) and chronic schizophrenia is now well-established on
the basis of evidence of shared abnormalities in phenomenology, genetics
biology, course, and treatment response. Thus, in the study of schizophrenia,
a research strategy of increasing popularity is the investigation
of SPD. Previous research has shown that medicated patients with schizophrenia
exhibit deficits in the attentional modulation of startle eyeblink
modification (SEM). This study examined SEM in unmedicated schizophrenia
patients (n=15) and healthy control (n=14) participants. Participants
performed an auditory selective attention task involving the presentation
of to-be- attended, to-be-ignored, and novel tones which served as
prepulses. Acoustic startle probes were presented at short or long
lead-intervals following the onset of tones and occasionally during
the intertone interval. Controls showed significantly greater prepulse
inhibition (PPI) at the 120 ms lead interval and greater prepulse
facilitation at 4500 ms during attended than ignored prepulses, demonstrating
attentional modulation of SEM. The amount of PPI at 120 ms and facilitation
at 4500 ms during the novel prepulse was intermediate between those
for attended and ignored prepulses. In contrast, patients failed to
show differential PPI at 120 ms and tended to show greater facilitation
at 4500 ms during the novel prepulse. Preliminary data for unmedicated
patients with SPD (n=9) and age- and sex-matched healthy controls
(n=9) will also be presented.

Do people mimic what they see or what they know? Facial mimicry revisited
U. Hess, P. Herrera, P. Bourgeois, & S. Blairy
University of Quebec at Montreal
In recent years a number of studies have shown that people show congruent
facial expressions to the emotional facial expressions of others.
It has been proposed that facial mimicry serves to communicate emotion
recognition. Two studies were conducted to assess whether people mimic
what they see or rather show the emotional expressions they recognize.
For the first study, 31 subjects saw expressions of happiness, sadness,
anger, fear, and disgust either in normal or inverted orientation.
Dependent measures were the facial IEMG at the Corrugator Supercilii,
Orbicularis Oculi, and Levator L.A.N. sites as well as emotion recognition.
If people show expressions congruent with the emotion recognized,
mimicry-like effects should be observed for well recognized expressions
regardless of orientation. Results for the Corrugator S. and Levator
L.A.N. sites suggest mimicry effects only for expressions shown in
normal orientation regardless of level of recognition. The findings
suggest that people mimic what they see rather than what they recognize.
However, this notion was not supported by results from a second study
where 27 subjects saw only the mouth region of emotional facial expressions
of happiness, sadness, anger, fear, and disgust. Dependent measures
were the facial IEMG from the upper facial region. Activity at the
Corrugator S. and Orbicularis O. sites is congruent with the notion
that subjects mimic the upper facial components of well recognize
expressions even when these components were not part of the stimulus.
These findings can be reconciled if the social meaning of the stimulus
is considered.

We're # 1: Sports fans and sports pictures
Charles Hillman, Bruce Cuthbert, James Cauraugh, Margaret Bradley, Harald Schupp, & Peter Lang
University of Florida
Pleasant affect, defined on the basis of identification with college
athletic teams, was investigated in the context of picture processing.
Forty undergraduate students (24 male, 16 female) from the University
of Florida (UF) viewed five slides from each of two picture categories
(UF Gator sports and non-Gator sports). Participants were ranked according
to their enthusiasm for UF Gator athletics (high, moderate, low).
During the 6 s picture viewing period, event-related potentials (ERP)
to a startle probe, cortical slow wave potentials to picture onset,
and heart rate were recorded. Following each picture, self report
measures of valence and arousal were also collected. Significant Spectator
Level x Picture Category interactions occurred for the probe-P300
peak of the ERP, with moderate and high fans showing attenuated P300
peaks compared to low fans--consistent with previous results demonstrating
an attenuated P300 component when subjects view highly interesting
pictures. Heart rate and subjective measures of valence and arousal
also demonstrated a similar interaction, as high fans showed greater
heart rate deceleration and increased valence and arousal ratings
to UF Gator pictures, relative to moderate and low fan groups. Significant
overall differences in picture categories were also found for cortical
slow waves following picture onset, with increased positivity for
UF Gator pictures relative to non-Gator sports slides. These results
suggest psychophysiological measures are sensitive to varying levels
of spectator enthusiasm, and are interpretable in the context of previous
results with emotional pictures.

Self-reported and manifest sleepiness during blind withdrawal from Modafinil in patients with narcolepsy
M. Hirshkowitz
Houston VAMC & Baylor College of Medicine
Patients with narcolepsy suffer from pathological sleepiness, commonly
treated with stimulants. This multicenter study examined safety and
efficacy of modafinil in subjects with narcolepsy. 273 drug-free subjects
(age 17-68 years) diagnosed with narcolepsy were randomized into
(placebo), 200, and 400 mg fixed, single daily dose groups. After
a 9-week study period, a two-week withdrawal began. We assessed manifest
sleepiness with the maintenance of wakefulness test (MWT); self-reported
sleepiness was quantified with Epworth Sleepiness Scale (ESS). In
MWT, subjects have four, 20-minute, nap opportunities at two-hour
intervals, beginning two hours after awakening. Subjects were semi-recumbent
in a dimly lit room and instructed to RESIST falling asleep. The rapidity
of sleep onset was determined physiologically with EEG-EOG-EMG recordings,
according to standardized procedure. Sleep latency was the principal
measure. By contrast, ESS entailed self-rating dozing likelihood (never
[0], slight [1], moderate [2], high [3]) in 8 hypothetical situations
(e.g., watching TV). Both modafinil doses significantly (p<.001) improved
MWT and ESS compared to placebo. However, focus here is on double-blind,
placebo-controlled, two-week, abrupt withdrawal. During withdrawal,
the ability to remain alert (MWT) regressed to baseline in the 200
mg group and beyond baseline in the 400 mg group. ESS regressed to
placebo levels for both active-drug groups. The major between groups
adverse event during withdrawal was increased complaints of hypersomnolence
(1, 10, and 6% for placebo, 200, and 400 mg groups, respectively).
Thus, objective and self-reported sleepiness return, but do not necessarily
rebound, within two weeks after subjects with narcolepsy are withdrawn
from modafinil.

Caloric intake, EEG, and ERPs
Lisa D. Hoffman1 & John Polich2
1University of Minnesota, 2The Scripps Research Institute
EEG and ERPs were recorded pre- and post-stimulus in food-deprived
(16 hrs) subjects (N=12, 6M/6F) before and after ingestion of a 500
calorie liquid meal-replacement (food) or a non-caloric (placebo)
drink. Food condition order was counterbalanced across subjects, who
were assessed at the same time and day in two consecutive weeks. Subjects
performed an auditory oddball task at 15 min intervals once before
and five times after eating. P300 amplitude increased in the meal-replacement
condition (caloric) relative to the placebo condition (no calories)
and decreased overall across trial blocks because of habituation effects.
P300 latency demonstrated no reliable effects of caloric intake but
some decrease over trial blocks. EEG data were submitted to spectral
analysis, but few food-related effects were observed for either the
pre- or post-stimulus epochs. However, correlations between EEG power
(post-stimulus) and P300 amplitude were greater in the delta and theta
bands for the food compared with the placebo condition. Similar but
weaker effects were observed for the correlations between alpha-1
and alpha-2 power and P300 amplitude. No reliable outcomes were obtained
for the correlations between EEG mean band frequencies and P300 latency
across trial blocks. These results suggest that caloric intake affect
the P300 ERP but not background EEG when meal-replacement drinks are
consumed. However, caloric intake appears to enhance the relationship
between P300 amplitude and EEG power.

Do the EEG effects of cigarette smoking show tachyphylaxis?
Michael E. Houlihan1, Walter S. Pritchard1,2, John H. Robinson1,2
1Bowman Gray School of Medicine, 2R. J. Reynolds Tobacco Company
Under low mental-workload conditions, cigarette smoking after overnight
abstention typically produces a change in the EEG power spectrum that
has been referred to as 'activation'. This involves a general decrease
in power in lower-frequency (< 12 Hz) power and an increase in higher-frequency
power. In two morning sessions, we recorded EEG from overnight-abstaining
male smokers before and after smoking of the 1st three cigarettes
of the day at 40-min. intervals. The control session involved replacing
the 2nd cigarette with one having a very low nicotine yield (0.05
mg) relative to the other cigarettes (1.1 mg). In both sessions, smoking
the 1.1-mg cigarettes decreased delta power at central and parietal
loci. Smoking the 0.05-mg cigarette did not affect delta. A similar
effect was observed for alpha power at central loci, with the exception
that the decrease following smoking of the 2nd 1.1-mg cigarette was
somewhat attenuated relative to the 1st. and 3rd cigarettes. In both
sessions, smoking the 1.1-mg cigarettes increased beta2 power at central-midline
loci. Smoking the 0.05-mg cigarette did not affect beta2. Overall,
we found little evidence of tachyphylaxis of the effects of cigarette
smoking on the human EEG under resting conditions, at least for a
40-min. inter-cigarette interval.

Short-term memory scanning and smoking
Michael E. Houlihan1, Walter S. Pritchard1,2, John H. Robinson1,2
1Bowman Gray School of Medicine, 2R. J. Reynolds Tobacco Company
Performance of many different cognitive tasks improves after smoking/nicotine.
Several studies have reported that smoking/nicotine speeds reaction
time (RT) in the Sternberg short-term memory (STM) scanning task.
Most of these studies failed to separate specific STM-scanning effects
from effects on total response time. This study investigated the effect
of smoking/nicotine on STM scanning using both RT and event related
potentials (ERPs). Two, three or four consonants were displayed as
a memory set. After a brief interval, a single probe consonant was
displayed. A right-side mouse-button press was required in response
to a probe that was in the memory set and a left-side press to a probe
that was not in the set. The task was performed before smoking, after
smoking a 0.05-mg nicotine-yield cigarette, and after smoking a 1.1-m
nicotine-yield cigarette (FTC method). EEG was recorded from Fz, Cz,
and Pz. Accuracy increased after smoking the 1.1-mg cigarette compared
to both pre-smoking and after the 0.05-mg cigarette. This effect was
mainly due to the improvement with set sizes 2 and 3. RT was fastest
after smoking the 1.1-mg cigarette. In addition, RT after the 0.05-mg
cigarette was faster than the pre-smoking RT. The ERPs showed a great
deal of individual variation in wave shape. Six of the 20 participants
had ERPs with a clearly definable P300 complex. In these subjects,
P300 latency appeared shorter after the 1.1-mg cigarette, but the
effect did not reach significance due to the small N. While there
are significant effects of smoking/nicotine on accuracy and RT, the
lack of a consistent effect on P300 latency indicates that smoking/nicotine
either does not specifically affect the STM-scanning stage of processing,
or may do so only in a subset of participants.

Psychophysiology research and career development: Funding opportunities at NIMH
Lynne C. Huffman
The National Institute of Mental Health (NIMH) 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. In addition, information
will be available on the recent reorganization of NIMH extramural
divisions and the upcoming merger of NIMH and NIH Division of Research
Grants review structures.

An ERP analysis of the influence of temperamental differences on infants' visual attention
Scott J. Hunter1 & Rathe Karrer2
1University of Rochester School of Medicine & Dentistry, 2University of Kansas Medical Center
This study examined the relationship between the development of infants'
self-regulatory capacities, defined as temperament, and visual attention
to discrepancy. Twenty-two healthy six-to-nine month old infants were
recruited in order to investigate how variability in visual attention
may be attributed to temperamentally-based individual differences
in reactivity. Components of the infants' temperament, assessed by
parental report on the Infant Behavior Questionnaire (IBQ), were compared
with neurophysiological (ERPs) and visual fixation indices of task-related
attention. It was hypothesized that more temperamentally reactive
infants would demonstrate greater attention to novelty and stimulus
discrepancy while engaged in a modified visual discrimination task,
as evidenced through components of their ERPs and looking behavior.
Results supported previous findings regarding a differential ERP and
looking behavior response to stimulus discrepancy, as well as revealing
a significant relationship between dimensions of temperament indexed
by the IBQ and the infants' ERP responses related to attention. Specifically,
more temperamentally reactive infants were found to display greater
NSW area and Nc amplitude across the different stimuli, while less
reactive infants were most responsive to Novelty. Looking behavior
did not appear to be similarly influenced by temperament, however.
Results are discussed with regard to the role individual differences
in attention and temperament may play in directing infants' interactions
with their environment and in their cognitive development.

Auditory attentional and sensory gating in Tourette Syndrome and Attention Deficit-Hyperactivity Disorder
Scott J. Hunter, Gubert Tan, Jau-Shin Lou, Donna Palumbo, and Peter Como
University of Rochester School of Medicine and Dentistry
This study examined whether alterations in motor control associated
with Tourette Syndrome (TS) interact with attentional difficulties
inherent in comorbid TS and ADHD, such that separate contributions
of the disordersU pathophysiology to central processing deficits could
be delineated. ERPs associated with sensory gating and preattentive
processes were recorded from four groups of 8-16 year old children
(TS-alone, ADHD-alone, combined TS+ADHD, non-disordered), during two
auditory conditioning tasks: a passive condition requiring participants
to attend-only and an active condition where participants made a motor
response (button press) to an infrequently presented tone. Results
supported the hypothesis that children with ADHD and combined TS+ADHD
demonstrate deficits in stimulus filtering that impacts their allocation
of attentional resources. Results are discussed with regard to the
contribution of a putative basal ganglia pathophysiology to the sensory
gating and attentional disturbances seen in these disorders

Effects of smoking cues on psychophysiological and subjective measures of reactivity
Kent E. Hutchison, Raymond Niaura, & Robert M. Swift
Brown University
Exposure to stimuli previously paired with drug consumption precipitates
conditioned appetitive responses (e.g., craving). The aim of the present
study was to investigate conditioned responses to smoking cues across
several psychophysiological and subjective measures. Sixteen smokers
participated in a 2 (Cue: smoking cues vs control cues) x 2 (Session:
Session 1 vs. Session 2) within subjects design after abstaining from
cigarettes for 12 hours. In the smoking cue condition, subjects were
asked to light and hold a cigarette of their preferred brand for 190
s without smoking. In the control cue condition, subjects held a pencil
for 190 s. Order of conditions was counterbalanced with a 10 minute
break between conditions. Prepulse inhibition of the startle reflex
(measured with 4 control and 4 prepulse trials) and heart rate were
assessed during the 190 s, while subjective measures were collected
immediately thereafter. All subjects returned one week later and repeated
the protocol to determine whether the responses habituate from session
1 to 2. Repeated measures ANOVAs indicated a main effect for Cue such
that exposure to smoking cues resulted in an increase in subjective
craving, F(1,14)=6.01, p<.05, and a decrease in PPI, F(1,14)=8.08,
p<.05 in both session 1 and 2. Changes in heart rate and the magnitude
of pulse alone trials were not significant. There was also a main
effect for session on craving with a decrease in craving from session
1 to session 2. These results suggest that PPI may be a useful measure
of cue reactivity.

The effects of high vs. low nicotine cigarettes on prepulse inhibition
Kent E. Hutchison, Raymond Niaura, & Robert M. Swift
Brown University
The administration of nicotine activates mesolimbic dopamine, and
activation of mesolimbic dopamine is known to attenuate prepulse inhibition
of the startle reflex (PPI). Thus, smoking should attenuate PPI in
humans. However, at least one study has reported that PPI was increased
by smoking (Kumari et al, 1996; Psychopharmacology, p 54-60). The
primary aim of this study was to rigorously test the effects of smoking
on PPI and other measures. Sixteen smokers participated in two experimental
sessions one week apart in a 2 (Dose: High, 1.1 mg vs. Low, .1 mg)
x 2 (Trial: Cigarette 1 vs. Cigarette 2) design. On one session they
received two High Dose cigarettes and on the other session they received
two Low Dose cigarettes. Order of session was counterbalanced across
subjects. In the High Dose condition, subjects received a cigarett
containing 1.1 mg of nicotine, which was divided into twelve 35 cc
puffs with one puff delivered every 25 s. One minute after completing
the cigarette, subjects received five minutes of startle trials (6
pulse alone and 6 prepulse trials) followed by subjective measures.
Subjects smoked a second high dose cigarette 25 minutes after completing
the first. Subjects in the low dose condition received two cigarettes
containing .1 mg of nicotine in the same manner. Repeated measures
ANOVAs revealed a main effect for Dose such that the high dose nicotine
significantly attenuated PPI, F(1,14)=8.08, p<.05 and increased HR,
F(1,14)=18.95, p<.05. These results are consistent with the hypothesis
that nicotine attenuates PPI immediately after smoking.

The effects of amphetamine on psychophysiological and subjective measures of stimulation in humans
Kent E. Hutchison & Robert M. Swift
Brown University
Prepulse inhibition of the startle reflex (PPI) is attenuated in animals
after administration of drugs that stimulate mesolimbic dopamine activity.
In humans, administration of dopamine agonists increases self-reported
stimulation and euphoria. The aim of the present study was to evaluate
the effects of d-amphetamine (20 mg) on PPI, heart rate, and subjective
stimulation in humans. Twenty-eight subjects (14 women) participated
in the double-blind, placebo controlled, repeated measures study.
In one session, subjects received d-amphetamine (20 mg) orally, and
in the other session, subjects received an identical placebo. Sessions
were one week apart and the order of sessions was counterbalanced
across subjects. Subjects were assessed at 60, 90, and 120 minutes
after ingestion. Each assessment consisted of a 5 minute block of
startle trials (6 control trials and 6 prepulse trials) followed by
subjective measures of stimulation and mood. A 2 (Drug: Amph vs. Placebo)
x 3 (Time: 60", 90", 120") repeated measures ANOVA on PPI (i.e., pulse
- prepulse/ pulse) revealed a drug by time interaction with significant
attenuation of PPI at 90 minutes but not 60 or 120 minutes after ingestion
of amphetamine, F(2,50)=3.45, p<.05. Analyses on heart rate and subjective
stimulation revealed main effects with significant increases by d-amphetamine,
F(1,26)=7.78, p<.05 and F(1,27)=57.08, p<.05, respectively. There
were no effects for gender. These results suggest that the effects
of d-amphetamine on PPI occur only within a narrow window of time
(~ 90 minutes post ingestion) while the effects on subjective stimulation
and heart rate are more generalized.

Do measures of novelty seeking predict psychophysiological and subjective stimulation after amphetamine consumption?
Kent E. Hutchison & Robert M. Swift
Brown University
Previous studies have suggested that measures of novelty seeking are
associated with individual differences in mesolimbic dopamine reactivity.
The aim of the present study was to examine the relationship between
two measures of novelty seeking and psychophysiological and subjective
measures of stimulation at 90 min after a d-amphetamine (20 mg) challenge.
Twenty-eight subjects (14 women) completed the Tri-Dimensional Personality
Questionnaire (TPQ) and the Sensation Seeking Scale (SSS) at baseline
and were subsequently given a capsule containing 20 mg of d-amphetamine.
Subjects were assessed at 60, 90, and 120 minutes after ingesting
the capsule with 5 minutes of startle trials (6 pulse and 6 prepulse
trials) and measures of subjective stimulation and elation. Multivariate
regression analyses were performed using a forward selection method
to determine which scales from each of the personality measures significantly
predicted measures of stimulation after amphetamine consumption while
controlling for stimulation at baseline. Only the data for the 90
min assessment were used in the multivariate regression analyses because
the effects of d-amphetamine were most evident at this time. The results
indicated that only the Novelty Seeking scale of the TPQ predicted
PPI, F(1,26)=5.44, p<.05, (R2=.18), and tended to predict stimulation,
F(1,27)=3.11, p=.09, (R2=.11), while none of the TPQ scales predicted
elation. Only the Disinhibition scale of the SSS predicted subjective
stimulation, F(1,27)=7.88, p<.05, R2=.23, and elation, F(1,27)=6.88,
p<.05, (R2=.21), while none of the SSS scales predicted PPI. Thus,
the TPQ was more predictive of psychophysiological stimulation while
the SSS was more predictive of subjective stimulation and elation.

ERP effects of smoking on fixed and varied set memory scanning
Aaron B. Ilan1 & John Polich2
The Scripps Research Institute
Regular cigarette smokers performed different versions of the Sternberg
memory scanning task after abstaining from smoking overnight and performed
the same tasks again after smoking. In a control condition, the same
subjects performed these memory tasks while smoking abstinent and
again after sham smoking. Compliance with abstinence instructions
and levels of carbon monoxide intake after smoking were monitored
with a CO breathalyzer. Electrophysiological data were recorded from
16 scalp electrode sites as subjects performed fixed and varied set
memory search. Smoking and memory set conditions were counterbalanced
across subjects, who were assessed at the same time of day in two
different sessions. P300 amplitude to probe stimuli decreased after
smoking. Consistent with previous results, the magnitude of this amplitude
decrease depended on task difficulty. Comparisons with electrophysiological
and behavioral data from control conditions support the hypothesis
that in relatively difficult tasks, such as memory scanning with larg
memory sets, cigarette smoking may have deleterious effects on cognitive
functioning. In easier tasks, such as memory scanning with small memory
sets, cognitive detriments may be offset by the vigilance and motor
benefits that accompany release from nicotine withdrawal.

A physiological analysis of changes in emotion evoked by tactile stimulation
Jae J. Im1, Jin Wu Park1, Ji-Eun Kim2, Imgap Yi2, & Jin-Hun Sohn2
1Inje University, 2Chungnam National University
Once we know the physical and mental responses to a particular stimulus,
we can apply those findings to improve the quality of human life.
These factors call for an analysis of the physiological signals which
reflect the emotional changes caused by variations in the autonomic
nervous system. The objective of this study was to find the correlation
between emotional changes and the parameters of physiological signals,
ECG and respiration, evoked by tactile stimulation. Twenty-one subjects
were used, and five different types of materials, dry cotton, wet
cotton, vinyl, paper, and sand paper were applied to the left hand
of each subject as a tactile stimulation. HRV(heart rate variability)
was obtained from ECG signals, and the ratio of high to low frequency
amplitude(HF/LF) was calculated from power spectrum of HRV. Also,
five parameters, respiratory rate, inspiration/expiration duration,
and inspiration/expiration peak values, were extracted from respiratory
waveforms. The results showed that HF/LF decreased as the subject
experienced the unpleasant emotion caused by wet cotton and sand paper.
On the other hand, when the subjects experienced the comfortness,
HF/LF increased. Peak values for the inspiration showed a significant
differences among five types of stimulation. It appeared to have higher
peak values, which correspond to deep inspiration, as the subject
experienced more pleasant emotion. Also, when the subject were in
the unpleasant state, peak values for the inspiration decreased. It
was concluded that the proposed methods can be applied to evaluate
the characteristic changes in the autonomic nervous system which regulates
the human emotion in the psychological aspects. (This work was supported
by a grant 17-01-01 from Korean Ministry of Science and Technology)

ERPs and evaluative categorizations: The negativity bias in evaluative space
Tiffany A. Ito & John T. Cacioppo
The Ohio State University
Event-related brain potentials (ERPs) were used to assess the neural
substrates associated with positive and negative evaluative categorizations.
Although positivity and negativity have often been viewed as interchangeabl
and opposite ends of a single continuum, recent research suggests
that positive and negative valent processes are associated with separable
systems with distinguishable activation functions. For instance, the
negative evaluative system can be characterized by a negativity bias:
activation of the negative system appears to result in greater motivational
output than comparable activation of the positive system. To assess
operation of the negativity bias, participants viewed pictures of
affectively neutral, positive, and negative pictures embedded within
sequences of other neutral pictures. Participants performed the evaluative
task of indicating whether each picture was positive, neutral, or
negative. To manipulate comparable amounts of evaluative activation,
positive and negative pictures were equated for self-reported arousal
and affective extremity. Replicating our prior research, the contextually
inconsistent positive and negative pictures evoked a larger late positive-going
potential (LPP) that manifested at approximately 500 ms and was maximal
at midline centroparietal regions. Furthermore, and as predicted by
the negativity bias, negative pictures were associated with larger
LPPs than positive pictures. In a second study, 555 slides from the
IAPS were rated on 5-point unipolar scales to assess the strength
of positive affect and negative affect each evoked. Regression analyses
revealed a steeper slope for the negative than positive reactions
to the slides. Together, these results support distinct activation
functions for positive and negative motivational systems

Anterior brain electrical asymmetries differentiate between pleasant and unpleasant pictures
Daren C. Jackson, Jessica R. Malmstadt, Christine L. Larson, Kenneth C. Rando, & Richard J. Davidson
University of Wisconsin-Madison
Frontal EEG asymmetry has been shown to predict positive and negative
affect: negative affect has been correlated with relative right anterior
activation, while positive affect has been correlated with relative
left anterior activation. In the present study, we experimentally
manipulated emotional states to elucidate the contribution of the
frontal lobes to the initiation and ongoing regulation of emotion.
Twenty-seven channels of EEG were recorded from 12 right-handed undergraduate
subjects while they viewed 126 stimuli (pleasant, unpleasant, or neutral)
from the International Affective Picture Set. Each picture was presented
for six seconds. Artifact-free EEG epochs were re-referenced to the
averaged ears reference. Alpha power (8-13 Hz) was computed and averaged
across trials. Analyses focused on three epochs. During the first
three seconds of picture presentation, there were no significant frontal
asymmetry effects. During the final three seconds of picture presentation,
the F8-F7 asymmetry score showed significantly greater relative left-sided
activation for pleasant than unpleasant pictures. During the three-secon
post-picture offset period, F8-F7, F4-F3, and FC8-FC7 asymmetry showed
significantly greater relative left-sided activation following pleasant
pictures relative to unpleasant pictures. Asymmetrical activation
in posterior regions did not differentiate between conditions. These
data show that frontal EEG asymmetry differentiates between positive
and negative emotional states. However, such differences did not become
apparent until relatively late in the picture presentation, and became
more widespread across the frontal region during the picture offset

Role of the dorsolateral prefrontal cortex in random number generation: A PET activation study
Marjan Jahanshahi, Georg Dirnberger, Rebecca Fuller, & Chris Frith
Institute of Neurology & University of London
Random number generation (RNG) requires suppression of habitual responses
such as counting, and monitoring of responses; and is attention-demanding.
A network modulation model proposes that in RNG, suppression of habitual
counting is achieved through the modulatory (inhibitory) influence
of the left dorsolateral prefrontal cortex (DLPFC) over a number associative
network in the superior temporal cortex (STC). >From this we predicted
(i) increase in rCBF in left DLPFC with RNG relative to a control
counting task (ii) decrease in rCBF in left STC in RNG, but increase
with counting. A 2 (Task) x 6 (Rate) factorial design was used. 6
young, right-handed, males performed 2 tasks: RNG (generated numbers
between 1 and 9 randomly), and COUNT (counted from 1 to 9), at 6 rates
paced by tones. rCBF measurement was with a Siemens CPS ECAT EXACT
HR PET scanner using 15O labelled water. Relative to COUNT, RNG was
associated with significant increases in rCBF in anterior cingulate
(BA 24, 32), left DLPFC (BA 46, 9), superior parietal cortex (BA 7)
bilaterally, right inferior prefrontal cortex (BA 45, 47), cerebellum
bilaterally; and significant decreases in rCBF in right STC (BA 22,
42), right fusiform gyrus (BA 21), left parietal area 40, right medial
occipital cortex (BA 19), left medial temporal cortex (BA 28,38).
The Task x Rate interaction was significant: increase in rCBF in left
STC was associated with decreasing rates of RNG and increasing rates
of COUNT. The results support the network modulation model of RNG.

An in vivo comparison of two penile strain gauges: Introducing a new calibration method
Erick Janssen, Mariette Vissenberg, Sambhavo Visser, & Walter Everaerd
University of Amsterdam
This study assessed the comparability of two types of penile strai
gauges, using both a circular and a newly developed oval calibration
device. A group of 25 sexually functional subjects placed both an
electromechanical and an indium/gallium-in-rubber strain gauge on
the penis and viewed an erotic film. The electromechanical gauge as
calibrated on the circular device resulted in greater penile circumference
changes than the indium/gallium-in-rubber gauge. Mean circumference
changes were not different for the two strain gauges when the oval
calibration device was used. It was concluded that the use of an oval
calibration device improves ecological validity of calibration of
penile strain gauges. Standard inclusion of this method in studies
on male sexual response will increase comparability of research findings.

Physiological preparedness during unpredictable foreperiods: Preparing the heart, eye, and brain
J. Richard Jennings1, Maurits van der Molen2, & Stuart R. Steinhauer1
1University of Pittsburgh, 2University of Amsterdam
Psychophysiological "preparatory" responses may or may not depend
on a focused expectation of when an expected stimulus will occur.
Assuming a unitary preparatory state, positive correlations between
psychophysiological measures, and between these measures and performance,
would be predicted when variable strategies for preparation are induced
by a non-aging foreperiod design. Changes in heart rate, pupillary
diameter, and the CNV were examined during trials in which the foreperiod
of a simple reaction time task was either fixed or unpredictable.
In addition, separate trials were recorded in which the imperative
stimulus was triggered by psychophysiological changes occurring spontaneously
during the foreperiod. 32 male and female college- aged volunteers
participated. When a non-aging foreperiod was used to reduce expectancy,
transient pre-stimulus psychophysiological responses were eliminated,
but slow changes over the foreperiod -- slowing of heart rate, dilation
of the pupil, and cortical surface negativity -- were not eliminated.
Triggering of the imperative stimulus by physiological change (either
heart rate slowing or decrease in pupil diameter) did not influence
RT. Correlations between psychophysiological changes in the foreperiod
and between these changes and RT were generally low; heart rate deceleration,
pupillary dilation, and CNV exhibited independent preparatory changes.
The results were more consistent with a multi-process than a unitary
process view of preparation.

ERP evidence for separate context memory systems for source and temporal sequence information
Ray Johnson, Jr., Kurt Kreiter, Britt Russo, & John Zhu
Queens College/CU
While much has been learned about the neural substrates of long-term
memory, the majority of studies have been concerned with retrieval
of "event" information. However, a distinguishing characteristic of
episodic (i.e., autobiographical) memories is that they include information
about the particular spatio-temporal context in which the event occurred.
Thus, the purpose of the preset study was to use ERPs to provide a
more detailed characterization of the processes and neural substrates
underlying the retrieval of contextual information from long-term
memory. Recognition-related ERPs from 32 sites were compared with
those elicited during retrieval of source (which list the item belonged
to) and temporal sequence (which half of the list the item was in)
information. Thirteen subjects learned 80 words, in order, prior to
the experimental session (Home list). A second list of 80 words, was
learned, in order, while ERPs were recorded (Lab list). Recognition
elicited the previously described positive ERP activity, maximal over
centro-parietal scalp in the 500-800 ms epoch. In contrast, context
retrieval elicited frontal maximal negative ERP slow waves with durations
matching those of retrieval operations (i.e., 300-2000 ms), as measured
by reaction time. Importantly, significant differences in the scalp
distributions of the ERP activity elicited during retrieval of source
and temporal sequence information were also found. The data suggest
that there are separate frontal-based memory systems for different
types of context information, as well as different neural systems
for retrieving item and context information. The temporal aspects
of the data will be discussed.

Memory systems underlying recall of items from semantic and episodic memory: An ERP analysis
Ray Johnson, Jr., Kurt Kreiter, & John Zhu
Queens College/CUNY
The purpose of the present study was to specify further the brain
mechanisms underlying recall of information from semantic and episodic
memory. ERPs were recorded from 32 scalp sites (bandpass = .01-35
Hz) from twelve subjects (9 females). Subjects performed three cued-recall
tasks in which they were presented with different three-letter word
stems on each trial. In the Semantic task, subjects recalled the first
word that came to mind that would complete the stem. In the Episodic-Easy
condition, subjects completed the stem with a word from a list of
words learned to a 90% recall criterion prior to the recording session
(Home list). In the Episodic-Difficult condition, subjects completed
the stem with a word from a list learned during the session (Lab list).
To obtain RT data, subjects pressed a button as soon as they recalled
an appropriate word. At the end of the 4-sec recording epoch, the
stem appeared in red and subjects said the recalled word out loud.
The ERP data revealed a complex pattern in which central (Cz) scal
became active at 200 ms, left lateral frontal scalp (F7, F3, Fc1)
at 300ms and left and right frontal pole (Fp1, Fp2) and cerebellar
areas (Cb1, Cb2) at 400 ms in all three conditions. Both Episodic
recall conditions (Difficult and Easy) showed significantly more prolonged
activation (by 200-300 ms) at frontal (F3, Fz, Fc1, Fc2) and frontal
pole (Fp1, Fp2) sites than was seen in the Semantic condition. These
data support and extend (in the temporal domain) results from recent
PET studies.

Am I too late? The neural consequences of missing a deadline
Tara M. Johnson1, Leun J. Otten1, Kirsten Boeck2, & Michael G. H. Coles1
1University of Illinois at Urbana-Champaign, 2Humboldt University
The Error-Related Negativity (ERN) occurs when subjects make errors
in choice reaction time tasks. Previous research has shown that ERN
amplitude depends on task instructions. In particular, when accuracy
is stressed, an incorrect response is associated with a larger ERN
than when speed is stressed. In the present experiment, we determined
(a) whether the ERN processes can be engaged by errors defined in
terms of response latency in addition to accuracy; (b) whether the
amplitude of the ERN varies with the emphasis placed on each error
type. Subjects performed two-choice reaction time tasks under speed
and accuracy instructions. Stimuli were large and small versions of
the letters S and T. The speed task (assigned to one letter) was performed
with middle and index fingers of one hand and the accuracy task (assigned
to the other letter) with the same fingers of the other hand. ERNs
were observed for both error types under both instructional conditions.
ERN amplitude was larger for inaccurate button presses when accuracy
was stressed, but larger for late responses when speed was stressed.
These data support the idea that the processes manifested by the ERN
use response representations that may include temporal information.
Furthermore, what constitutes an error depends on task goals. The
relative importance of a particular error feature varies with the
degree to which that feature is stressed in the specification of the

The measurement and psychological correlates of the cardiovascular effects of a real life stressor
D. W. Johnston & M. Tuomisto
University of St. Andrews
Laboratory studies of the cardiovascular (CV) effects of stress may
sacrifice ecological validity for experimental control while real
life stressors can be too uncontrolled to be interpretable. Howeve
advances in ambulatory physiological measurement can improve control
in real life situations markedly. The immediate and more distal psychological
correlates of the CV effects of the real life stressor of speaking
to a seminar group were investigated in 60 male undergraduates. The
CV measures were heart rate (HR) and pulse transit time (RPI), while
activity and posture were measured for control purposes. All physiological
measures were recorded continuously on an RM10 ambulatory recorder
for the duration of the students' working day (approximately 8 hours).
The psychological measures were ratings of anxiety at the time of
recording, a measure of the stressfulness of seminars in general and
trait anxiety. The seminars were matched with control periods for
duration, activity level and posture. HR was reliably higher (91.8
bpm) and RPI lower (128.1 ms) during the seminar than the control
periods (80.7 bpm & 142.3 ms). The subjects were also more anxious
during the seminar. Activity and posture did not differ between the
seminar and control periods so some of the usual confounds in real
life studies of stress can be discounted. The only reliable psychological
predictor of CV arousal during public speaking was the anxiety level
during the seminar, the higher the anxiety the greater the HR (r =
.49) and the shorter the RPI (r = -.36).

Does the cardiovascular response to active coping tasks predict cardiovascular reactivity in real life best?
D. W. Johnston, M. Tuomisto, & T. Kramer
University of St. Andrews
We have maintained that the elusive correlation between the cardiovascular
(CV) response to laboratory stressors and CV responses in real life
is most likely to be found when the laboratory stressor involves active
coping and when metabolic effects on CV response in the field are
minimized or controlled statistically. This was tested in 60 male
students in which the CV response to two active coping tasks, a choice
reaction time task and a version of Ravens Matrices, and two passive
tasks, the cold pressor and viewing a film, were related to HR and
pulse transit time (RPI) response in real life. HR was measured during
all the laboratory stressors. SBP was measured during the Ravens Matrices
and the passive coping tasks. "Additional HR", calculated from oxygen
consumption, was derived during the RT task. The ambulatory measures
of reactivity were HR variability over the day and the change in HR
and RPI associated with speaking at the seminar. HR, RPI, activity
and posture were measured continuously. Contrary to expectation the
only consistent correlate of HR variability during the day was the
HR response to the cold pressor (r = .46). The HR response to the
seminar correlated modestly with the responses to examples of active
and passive tasks. RPI change during the seminar related to the effect
of the cold pressor on SBP (r = .54). The results do not confirm that
the CV response to laboratory tasks thought to involve active copin
are the best predictors of CV reactivity in real life.

Age-related changes in child and adolescent auditory ERPs for two ADHD subtype groups and controls.
Stuart J. Johnstone1, Robert J. Barry1, and John W. Anderson2
1University of Wollongong, 2Private Psychological/Electrophysiological Practice, Sydney
Previous studies examining auditory ERPs in ADHD have (a)
not addressed the subtypes of the disorder as defined by DSM-IV, (b)
not fully considered responses to the standard stimuli of the ìoddballî
paradigm, and (c) examined effects over a narrow age range. This study
investigated the effects of diagnostic group and age on ERP components
elicited to target and standard stimuli in children and adolescents
(aged 8-18) diagnosed with AD/HD Combined Type, AD/HD Predominantly
Inattentive Type and normal controls (each N = 50). EEG was recorded
during an active auditory "oddball" task from 17 sites which were
clustered into nine topographic regions prior to analysis. Regression
analyses (for each diagnostic group separately) and multiway repeated
measures ANOVAs were used for data analysis. Across regions, regression
results indicated that several component amplitudes (target: N1, P2, N2;
standard: P2, N2) and latencies (target: N1; standard: N1, N2) changed
linearly with age for control subjects. The AD/HD Predominantly
Inattentive Type group showed weaker linear changes in amplitude
(target: N1, N2; standard: N2) and latency (target: N2, P3) with age. The AD/HD
Combined Type group did not show any significant linear changes with age
for component amplitude, but one significant latency change (standard:
N2). The control and each of the clinical groups also exhibited
differences in the topography of several components. These data are
conceptualised as indicating differences between (a) the clinical and
control groups and (b) the two AD/HD subtypes in their development
fundamental aspects of cognitive processing.

Effects of diaphragmatic breathing on motion sickness susceptibility
Michael D. Jokerst, Melinda Gatto, & Robert M. Stern
The Pennsylvania State University
The electrogastrogram (EGG) permits the noninvasive study of gastric
dysrhythmias associated with the experience of nausea caused by motion
sickness. Previous studies in our laboratory have found that increased
parasympathetic activity is associated with decreased motion sickness
susceptibility. Finding an effective non-pharmacological treatment
that would increase parasympathetic activity would be beneficial as
far as cost and the elimination of undesirable drug side effects.
The goal of this study was to see if deep diaphragmatic breathing
would decrease motion sickness susceptibility since it increases parasympathetic
activity. Method: Sixty participants, pre-screened for susceptibility
to motion sickness, were placed in one of three conditions: control,
pseudo-control, or experimental. All of the participants were exposed
to a rotating optokinetic drum to stimulate vection-induced motion
sickness. The pseudo-control group counted the number of breaths taken
during the rotation period. The experimental group did deep diaphragmatic
breathing throughout the rotation period. The drum period consisted
of a 6-minute baseline and a 16-minute rotation period. Electrogastrograms
were recorded for each participant. Results: Experimental condition
participants reported significantly fewer symptoms than did the control
or pseudo-control groups. The mean Subjective Symptoms of Motion Sickness
scores were 7, 9.5 and 10 respectively. Conclusion: Participants who
did deep diaphragmatic breathing were less susceptible to vection
induced motion sickness than those who did not. This is consistent
with theory that parasympathetic activity plays a major role in susceptibility
to motion sickness.

Reliability of percent distribution of power in the electrogastrogram
Michael D. Jokerst, Melinda Gatto, & Robert M. Stern
The Pennsylvania State University
There have been numerous studies exploring the validity of the electrogastrogram
(EGG) by comparing cutaneous and serosal recordings of gastric myoelectric
activity which, indeed, confirmed the validity of the EGG. However,
studies concerning EGG reliability are lacking. The purpose of this
study was to investigate the reliability of the EGG signal across
three bandwidths: 1) bradygastria (0 to 2.5 cpm), 2) normal (2.5 t
3.75 cpm), and 3) tachygastria (3.75 to 10 cpm). Nineteen healthy
volunteers participated. EGG data were obtained 5 and 15 hours after
a meal, and immediately after a meal on two separate occasions. Reliability
was assessed using generalizability theory. In the normal band (3
cpm), the percent of variance accounted for by occasions was 0.79%
while conditions accounted for only 2.4% of the variance. In the tachygastria
band, the percent of variance accounted for by occasions was 3.5%
while conditions accounted for only 0.07% of the variance. In the
bradygastria band, occasions did not account for any of the variance
while conditions accounted for 7.43% of the variance. A one-way analysis
of variance between percent normal power in the fasted and fed state
was not significant. In conclusion, results suggest that the percent
distribution of power of the EGG is very reliable. However, we failed
to find a difference in normal power in the fasted and fed states.
It is unknown why this occurred, but some possibilities include anticipatory
influences on normal activity during the preprandial condition, aversion
to the meal, and/or insufficient volume of the meal.

Females not more susceptible to motion sickness than males
Michael D. Jokerst, Melinda Gatto, & Robert M. Stern
The Pennsylvania State University
There are several anecdotal reports that females are more susceptible
to motion sickness than males, but these reports have failed to take
into account the possible effects of the gender of the experimenter
and the subjective nature of reports of symptoms of motion sickness.
To deal with the first possible confound, we used two male and two
female experimenters. To deal with the second issue, we examined electro-gastrographically
(EGG) recorded gastric tachyarrhythmia, an objective measure that
we have shown previously correlates highly with severity of symptoms.
Thirty-four male and thirty-four female participants were assigned
to either a male or female experimenter. Symptoms of motion sickness
were induced by placing participants in an optokinetic rotating drum
for an 8 minute baseline period followed by a 16 minute rotation period.
EGG was measured continuously with cutaneous electrodes, and reports
of symptoms were obtained from the participants every three minutes
during rotation. A factorial analysis of variance on the symptom scores
did not reveal a significant effect of the gender of the experimenter
or the subject, or their interaction. However, female subjects reported
more symptoms (10.3) than males (8.4). A factorial analysis of variance
on the difference in percent tachyarrhythmia from baseline to rotation
yielded no significant differences as a function of experimenter and
subject gender, or their interaction. However, male subjects did show
a larger increase in percent tachyarrhythmia (6.5) from baseline to
rotation than females (2.3). The results suggest that female subject
are not more susceptible to vection-induced motion sickness than male
subjects, although males may have a tendency to underreport symptoms.

Modified sham feeding of a pleasant and disgusting food: Cephalic-vagal influences on gastric myoelectric activity
Michael D. Jokerst, Max E. Levine, & Robert M. Stern
The Pennsylvania State University
The influence of the cephalic phase of digestion on gastric myoelectric
activity using electrogastrography (EGG) has previously been studied
in our laboratory. Using a modified sham feeding procedure with a
pleasant food, 3 cpm power increased significantly during sham feeding.
The aim of this investigation was to study systematically the sham
feeding effects of a pleasant and disgusting food on gastric myoelectric
activity. Thirty-eight healthy human participants were placed in one
of two conditions: 1) a pleasant sham feeding condition in which participants
chewed and expectorated two cooked hot dogs (N=19), and 2) a disgusting
sham feeding condition in which participants chewed and expectorated
two cold tofu dogs (N=19). Vegetarians and others who normally eat
tofu were excluded from the study. All participants were asked to
chew each mouth-full 6-7 times and to be very careful not to swallow
any of the food. EGG was recorded for 10 minutes prior to sham feeding,
during sham feeding, and for 15 minutes post sham feeding. A questionnaire
was given to each participant after the procedure to assess appetite
and food palatability. Results from the questionnaire showed that
the pleasant group found the food to be significantly more appetizing
than the disgust group, t(35)=8.41, p<.01, while appetite significantly
decreased following sham feeding in the disgust group only, t(18)=4.16,
p<.01. In the pleasant sham feeding group, as expected, 3 cpm power
was significantly increased during sham feeding as compared to 3 cpm
power in the disgust group, F(1,36)=4.06, p<.05. In conclusion, we
have demonstrated that the cephalic-vagal reflex, as measured by power
in the 3 cpm frequency region, depends upon the subjective palatability
of the food.

Differences in the startle response: Moderation by sex, distractibility, and arousal
Lyla Kaplan & Helen J. Crawford
Virginia Polytechnic Institute and State University
Excessive startle response (SR) in some individuals and immediate
habituation in others has been reported (Graham, 1979; Landis & Hunt,
1939). Ethnographic research has suggested that men and women differ
in the startle response (Simons, 1985, 1996). Mediators of the S
include the direction of attention (Anthony, 1985) and individual
differences of attention (Zelson & Simons, 1986). The present study
addresses how the acoustic eyeblink SR is moderated by sex differences,
by individual differences in distractibility (Cognitive Failures Questionnaire;
Broadbent, Cooper, Fitzgerald & Parkes, 1982) and arousability (Arousal
Predisposition Scale; Coren, 1990). Participants, (n=48), either low
or high in distractibility and arousability, were presented six 96dB
probes during two counterbalanced conditions in which participants
were instructed to attend to the probes or ignore the probes by mentally
counting backwards by threes. Subsequently, a Stroop Color-Word interference
task was administered. Results indicated that those who were easily
distractible exhibited more SR amplitude when instructed to ignore
the probes as compared to those who were not easily distractible.
Self reports obtained in post-experimental interviews corroborrated
this finding. Within both the highly distractible participants and
the highly arousable participants, men exhibited faster SR latency.
Contrary to previous findings (Tipper & Baylis, 1987), those who were
easily distractible did not exhibit a larger Stroop effect than who
were less distractible. Results suggest that controlled top-down processing
influences the startle response when manipulations of the direction
of attention are given. The theory of ironic processes of mental control
(Wegner, 1994) and developmental histories (Simons, 1996) are addressed.

ERP components related to anticipatory and stimulus-triggered processes of task switching
Frini Karayanidis1, Karen Murphy1, Max Coltheart1 & Patricia Michie2
1Macquarie University, 2University of Western Australia
Using a predictable task-set switching paradigm, Rogers and Monsell
(1995, J.Exp.Psychol:Gen, 124:207-231) found that reaction time (RT)
increased for switch compared to non-switch trials and that the effect
of task-set switching on RT declined with increasing response-stimulus
interval (RSI). However, even with long RSI, there remained a residual
effect of switching task-set. Rogers and Monsell (1995) proposed the
existence of two processes associated with task switching under pre-cued
conditions. An initial 'anticipatory' process is associated with partial
cognitive reconfiguration in anticipation of a change in task-set
and contributes to the decline in switch cost with increasing RSI.
A second 'stimulus-triggered' process is initiated only after stimulus
onset and accounts for the residual switch cost, even at long RSI.
The present study aimed at identifying electrophysiological markers
of these processes. ERPs were recorded from 30 scalp electrodes while
subjects (n=24) alternated predictably between two simple cognitive
tasks requiring binary classification of letters and digits, respectively.
RSI was varied across 150, 300, 600 and 1200 ms in different blocks.
RT data replicated Rogers and Monsell`s findings and supported th
existence of two processes. Stimulus-locked ERP difference waveforms
(switch - no switch) showed an early positivity which was primarily
evident at short RSI, followed by a negativity which declined in amplitude
with increasing RSI. Response-locked ERPs showed no difference between
switch and no-switch trials prior to stimulus onset for short RSI,
whereas, with long RSI, CNV amplitude was reduced for switch trials.
These results are discussed in relation to 'anticipatory' and 'stimulus-triggered'

ERPs indicate differences in visual attention between attention-deficit hyperactivity disorder and control boys in the absence of performance differences
Frini Karayanidis1, Philippe Robaey1, Michelle Bourassa1, Guy Geoffroy2, & Gilles Pelletier1
Saint Justine Hospital
The cognitive processes underlying the pattern of behavioral deficits
in attention-deficit hyperactivity disorder (ADHD) still remain unclear.
The difficulty in defining the cognitive profile of ADHD has resulted,
at least partly, from the changing behavioral profile of the disorder,
substantial heterogeneity of symptom profiles and varied pattern of
comorbidity. This study investigated visual attention in ADHD and
control boys using strict selection criteria. Inclusion into the ADHD
group was based on diagnosis of ADHD on both DSM III-R and DSM IV,
and a score exceeding 2 sd of normative mean on hyperactivity and/or
impulsivity indices of Conner`s Parent or Teacher Rating Scales. ADHD
subjects (n=12) were further selected so as to match the control group
(n=12) for age and IQ. Exclusion criteria included conduct or anxiety
disorders and tics. EEG was recorded from 30 scalp electrodes using
a visual oddball paradigm. No significant group differences were obtained
in number of correct responses, misses or reaction time (RT), but
the ADHD group showed greater trial-by-trial variability in RT. Anterior
N1 latency was delayed in ADHD for both frequent and target stimuli.
ADHD showed enhanced anterior negativity (300-600 ms) most evident
for deviants over left scalp. P300 was greater over the left hemisphere
in controls but not ADHD. The late posterior slow wave was also reduced
in ADHD for both stimulus types, the effect being lateralised over
the right hemisphere for deviants. The significance of psychophysiological
indices of attentional differences in the absence of overt behavioral
differences will be discussed.

Orit Karni and Emanuel Donchin
University of Illinois at Urbana-Champaign
Squires et al (1977, JEP-HP&P(3),299-315) presented subjects wit
concurrent auditory and visual oddball sequences. Of interest were
the ERPs elicited by rare events (5% of the trials) in either modality
when the subject focused exclusively on one modality. Only the rare
events in the attended modality elicited a P300 component. However,
when rare events appeared in both modalities this "double rare" event
elicited a P300 larger than that elicited by the attended rare. These
data suggest that unattended events are not entirely ignored. We used
a similar concurrent bimodal oddball task to further investigate how
the amplitude of the P300 is affected by an unattended sequence. With
respect to P300 our results were quite similar to those reported earlier.
However, a frontal negative deflection was observed when an unattended
rare was concurrent with a frequent event. The amplitude of the frontal
negative peak was equal in both modalities and larger when the unattended
rare was coupled with a frequent event, then when elicited by the
double rare event. Peak latencies were about 320 ms when subjects
attended the auditory sequence and 250 ms when the attended sequence
was visual. These data imply that (a) the instruction that a sequence
be ignored does not necessarily lead the subject to ignore the sequence
and (b) the processing of the "to be ignored" is affected by the processing
of stimuli in the focus of attention.

A new correlate of visual recognition memory in early infancy, Nc2: Event-related brain potentials (ERPs) from six-month-old infants with and without Down Syndrome (DS)
Jennifer Hill Karrer, Rathe Karrer, Lisa Chaney, Denis Fitzpatrick, & Keith Gora
University of Kansas, University of Kansas Medical Center
A negative component, Nc, is a prominent feature of the ERP from infants
experiencing a visual recognition paradigm. The present study doubled
the number of trials from our previous studies to determine the effects
of repeated stimuli experience on ERP components. Fz, Cz, Pz, C3,
C4, Oz and EOG were recorded from 20 infants (10 DS, 28.2 weeks &
10 n-DS, 27.5 weeks; 6 male & 4 female/group). The visual recognition
task involved 160 presentations of two female faces within an infant-controlled,
standard 80/20 oddball paradigm. (Stimulus duration=1s, ITI=2s.) Rare
stimuli resulted in longer fixations. Within ERPs, Nc peak amplitudes
were larger from DS infants for both frequent and rare trials at all
leads except C4. Nc peak latencies for frequent trials were similar
for both study groups (DS=560ms, n-DS=550ms). However, Nc latencies
for rare trials were faster among DS infants (560ms; n-DS=610ms).
While both DS and n-DS infants clearly held probability effects for
Nc, ERP morphology was somewhat different. NSW was most prominent
with Pb frequently absent in DS waveforms. A new component, "Nc2"
(n-DS=800ms), held later and more varied latencies in ERPs from DS
infants. The effects of repeated experience also depicted marked differences
between DS and n-DS infants. Nc and Nc2 decreased across blocks much
more quickly among the n-DS. Nc2 was present at C3 early, at C4 late,
but only for n-DS infants. Nc2 may provide evidence of memory formation
during cognitive processing of visual stimuli among infants. (Supported
by NICHD#30868
Maps of inhibitory functions within primary motor cortex (MI) in awake performing rhesus macaques: MRI and stimulus-triggered averaging (STTA) homology
Jennifer Hill Karrer1,2 & Paul D. Cheney2
1Univ of Kansas, 2Univ of Kansas Medical Center
Employing STTA techniques, we previously reported that 22 muscles
in the forelimb of rhesus macaques are consistently represented in
MI output maps. These maps depicted proximal muscles of the shoulder
and elbow surrounding the distal muscles of the wrist, fingers and
intrinsics which were represented deep in the precentral sulcus of
MI. Macaques were operantly trained to perform two tasks: (1) reaching
and prehension and (2) push and pull movements. The tasks involved
reciprocal activation of both agonists and antagonists implanted with
EMG across five joints of the forelimb. Systematic penetrations were
made at .5 mm intervals over contralateral forelimb MI, and every
.5 mm in depth. Single-pulse microstimulation at 15 uA was delivered
until a minimum of 2000 triggers were applied at each site. Many sites
within the forelimb representation yielded inhibitory effects at multiple
joints. The STTAs further revealed that the inhibitory representations
of functional muscle groups at each joint were primarily located medially
to representations for facilitation of the same muscle groups. This
was true even for compact, dense representations of the intrinsic
hand muscles. Based upon stereotaxic coordinates of MI, three glass
pipettes, filled with vitamin E, were inserted in the brain for MRI.
Structural aspects of our MI model generated from STTA maps are supported
by MRI. Furthermore, partitioning of MI topography confirmed interrelationships
of antagonist to agonist representations. The results of this study
depict functional anatomy of MI as organized for the coordination
of complex movements. Supported by NS25646, HD02528 & HD31772.

EMG-triggered movement-related brain potentials (MRPs) in 4 to 18-month-old infants: A pilot study
Jennifer Hill Karrer1,2, Rathe S. Karrer1,2, Denis Fitzpatrick1, and Keith Gora2
1University of Kansas, 2University of Kansas Medical Center
Early motor development within the first year of life is difficult
to study with traditional psychophysiological methodologies due to
an inability of the infant to respond within standardized behavioral
paradigms. The use of MRPs in the study of motor development within
infancy has not been previously reported. Employing electromyogram
(EMG)-triggered data collection, we have established methodology to
record MRPs as early as 4 to 6 months of age. Bipolar EMG electrodes
were placed on the anterior deltoid, biceps brachii and brachioradiali
to monitor reaching movements. To monitor kicking movements, bipolar
EMGs were placed on the gastrocnemius, soleus and tibialis anterior.
Beyond the subjects' natural predisposition to reach and kick, discrete
movements were also prompted with toys. FZ, CZ, PZ, C3, C4 and EOG
were recorded continuously (Neuroscan Synamps, DC-30hz, Scan 3.0 Acquire
software). Trials were determined by visual inspection of EMG and
EEG records. Criteria for averaging included: EMG inactivity at least
1 sec prior to EMG onset and artifact-free EEG. Averages were composed
of data 2 seconds prior to and 1 second after EMG onset. Similar to
our previous research on young children 5 to 6-years-of-age, the prominent
feature of the MRP was a positive slow wave starting at 750ms prior
to EMG onset. Analogous to the N2 in adult MRPs, a negative deflection
was observed 100ms prior to EMG onset. These data support that MRP
indices can be applied to study early motor development within infancy.
(Supported by HD31172 and HD30868.)

ERP indicators of associative learning in a Go-NoGo paradigm
Norbert Kathmann & Tom Nolde
University of Munich
To study the brain processes during associative learning we used a
differentially forewarned visual Go-NoGo paradigm. Each imperative
stimulus was preceded by one of three warning stimuli. One was consistently
followed by the Go stimulus, another one by the NoGo stimulus and
the third one randomly by either the Go or the NoGo stimulus. ERPs
elicited by cue and imperative stimuli were recorded from 19 scalp
positions. Response times showed that subjects learned the association.
RT improved by 20 ms after the cue signaling the Go stimulus, no change
was observed after the noninformative cue. Learning of the Go-contingency
was reflected by larger temporo-occipital N150-waves to the cue. The
P300 component was less positive after the cue predicting the NoGo
stimulus. Learning of the NoGo-contingency was also reflected in a
reduced late CNV. ERPs to the imperative stimuli showed a prominent
negative shift 200-300 ms after NoGo stimuli. This NoGo-N2 was unsensitive
to the prediction. The amplitude of the late positive component (LPC)
was reduced after predicted as compared to unpredicted imperative
stimuli. The results show that ERPs allow for a more detailed analysis
of associative learning processes. Go-contingencies probably change
attentional processing of the cue, whereas NoGo-contingencies specifically
alter response preparation. The amplitude of the LPC elicited by imperative
stimuli seems to be an objective measure of the information transmitted
by the cue. The paradigm used here might be useful in investigating
subtle changes of associative learning, e.g. the latent inhibition

Is the P300 related to the Stroop effect?
Voyko Kavcic & Jeffrey M. Clarke
University of North Texas
Previous ERP findings suggest that the locus of the Stroop effect
(SE) occurs after the P300 component (Duncan- Johnson & Kopell, 1981;
Science, 938-940). We investigated this further using four lateralized
Stroop-like experiments. Experiments 1 and 2 used classical Stroop
stimuli (i.e., color patches and color words), while the remaining
two experiments used face/word Stroop analog stimuli (i.e., prototypical
man's, woman's, and baby's face and the corresponding words MAN, WOMAN,
and BABY). Separate words and colors or faces were presented as congruent
or incongruent word-color/face pairs either to the same visual field
(i.e., unilateral presentations) or to separate visual fields (i.e.,
bilateral presentations). Separate groups of participants manually
identified the color patch in Experiment 1, performed a delay color
matching response in Experiment 2, identified the face in Experiment
3, and identified the word in Experiment 4. Reaction time results
revealed a SE for all four experiments, and across both unilateral
and bilateral conditions. For all four experiments, P300 amplitudes
were larger for congruent than for incongruent stimuli, while P300
peak latencies were unaffected. Contrary to Duncan-Johnson and Kopell
(1981), our findings indicate that P300 amplitudes are related to
the SE, presumably in terms of stimulus evaluation. Larger P300 amplitudes
for congruent than incongruent stimuli appear to reflect greater ease
of processing. The presence of a SE for reaction times, but not for
P300 latency (as was also found by Duncan-Johnson & Kopell), suggests
that there are multi-staged components to the SE, whereby response-processing
factors occur independent of the P300.

Brain event-related potentials (ERPs) in schizophrenia during a word recognition memory task
J. Kayser, G. Bruder, D. Friedman, C. Tenke, X. Amador, S. Clark, D. Malaspina, & J. Gorman
Columbia University
Impairments of recognition memory for words and attenuation of the
ERP "old-new" effect have been found in patients with left medial
temporal lobe damage (Rugg et al., 1991; Smith & Halgren, 1989). If
left temporal lobe dysfunction in schizophrenia involves medial structures
(e.g., hippocampus), then schizophrenic patients might show similar
abnormalities of word recognition memory. This study recorded ERPs
from 30 electrode sites while subjects were engaged in a continuous
word recognition memory task (Friedman, 1990). Preliminary findings
are reported for 18 patients having a diagnosis of schizophrenia (n=12)
or schizoaffective disorder (n=6) and 19 age and sex matched healthy
controls. Accuracy of word recognition was poorer in patients tha
controls (p<.001), and patients showed markedly smaller N2 amplitude
(p<.001). Moreover, N2 amplitude was associated with better performance,
with highest correlations over the left inferior parietal region (r=-.54).
N2 was greater over left than right hemisphere sites, with the largest
asymmetry over inferior parietotemporal sites (p<.001). This N2 asymmetry
was, however, less evident for patients than controls, and patients,
unlike controls, failed to reveal significant left-greater-than-right
N2 amplitudes at temporal-parietal sites. Both patients and controls
showed the expected "old-new" effect, with greater positivity to correctly
recognized "old" words at posterior sites (p<.001). There was no significant
difference between groups in P3 amplitude or in the "old-new" effect.
These findings suggest that poorer word recognition performance in
schizophrenia may arise from a left lateralized deficit at an early
stage of processing, 200-300 ms after word onset.

Phase transitions of hand movements and ambiguous figures in schizophrenia
Andreas Keil & Thomas Elbert
University of Konstanz
Studies of EEG and MEG dynamics indicate that schizophrenic patients
produce increased rates of alterations in their CNS dynamical states
(Rockstroh et al.,1997). This has been related to formal thought disorders.
In the present study we examined the performance of 16 schizophrenic
patients and 16 age and sex matched controls in a bimanual coordination
task as proposed by Kelso (1981), including rhythmical finger oscillations
(alternating activity of homologue muscle groups) at increasing speed
levels, using accelerosensors fixed at the index finger. Preferred
oscillation frequencies, and critical frequencies, i.e. frequencies
where phase transitions occurred (change towards simultaneous activity
of homologue muscle groups), were evaluated. In addition, a measure
of local dimensional complexity (pointwise D2) was determined for
the acceleration profiles of the subjects' movements. In the second
part of the study, the number of reversals in the perception of two
ambiguous figures, the Schroeder stairs and the Rubin vase, was monitored
as a measure of perceptual stability. It was found that schizophrenics
have less stable movement dynamics in horizontal finger cycling indicated
by a lower ratio critical/preferred frequency (p=.01) and by higher
means and standard deviations of the pointwise D2 (p<.005). Vertical
cycling did not differentiate between groups. While there were significantly
higher reversal rates in the schizophrenic group for the Rubin-Vase
(p<.05), n o differences were found for the reversal rate of Schroeder
stairs. Schizophrenic patients, in contrast to controls reported to
see the stairs for a longer time from below than from above (p<.0002).
Correlational analyses showed no evidence for a relation between perceptual
and motor stability.

Effects of prior exposure and evaluative observation on cardiovascular adaptation to stress
Robert M. Kelsey, Tamera R. Schneider, and Stefan Wiens
State University of New York at Stony Brook
The impact of different amounts of exposure to stress on subsequent
cardiovascular reactivity and adaptation to a change in stimulation
was evaluated in 96 undergraduate men and women. Subjects performed
three, two, one, or zero 4-min vocal mental arithmetic (MA) tasks
before performing a 4-min test task, which involved evaluative observation
during MA. All subjects then performed an additional familiar 4-min
MA task without evaluative observation. Each task period was preceded
by a 4-min baseline rest period. Changes from baseline in pre-ejection
period (PEP) and heart period (HP) were analyzed simultaneously in
multivariate trend analyses, and changes from baseline in total peripheral
resistance were analyzed in corresponding univariate trend analyses.
Within group comparisons showed that evaluative observation during
the test task disrupted cardiac but not vascular adaptation. Cardiovascular
adaptation was not disrupted upon re-exposure to the familiar MA task.
PEP and HP reactivity during the test task varied as a negative linear
function of the amount of prior exposure to stress (F(1, 88) = 4.44
and 4.08, respectively, p <.05). Stepdown F-tests indicated that PEP
contributed uniquely to this effect. Declines in PEP and HP reactivity
during the test task also varied linearly with the amount of prior
exposure to stress (F(1, 88) = 6.04 and 8.26, respectively, p < .02),
with more exposure leading to faster adaptation. These cardiovascular
effects were not attributable to variations in arithmetic performance
or cognitive appraisals of stress. The implications of these results
for competing theories of cardiovascular adaptation to recurrent stress
are discussed.

Reactivity and resiliency in older adults revisited: A five-year follow-up
Arlene R. King & Robert W. Levenson
University of California-Berkeley
The life stress-disease linkage is well-established, but individual
differences abound. We conceptualized these individual differences
as indicators of "resiliency", which we operationalized as the likelihood
of remaining in good health at a given level of life stress. We hypothesized
that people who show such resiliency over the long term, are those
who are physiologically less reactive to more acute stressors. We
chose an elderly sample (age 60-70) because of their increased health
vulnerability. Resiliency scores were determined by regressing physical
and psychological health data on recent (past 12-months) life stress
data. Physiological reactivity was determined by measuring cardiovascular
electrodermal, and somatic responses of 58 elderly married couples
during a 15-minute naturalistic discussion of a conflictive issue
in their marriage. Reactivity scores were computed by subtracting
averaged physiological levels during a five-minute silent pre-interaction
period from averaged levels during the 15-minute discussion. Five
years later, we recontacted couples and determined their current health
and recent life stress. We then determined whether reactivity during
the original marital interactions would predict resiliency to stress
five years later. Results indicated that individuals who had lower
levels of cardiovascular reactivity and experienced less negative
affect during the original marital interactions were more physically
and psychologically healthy five years later than would be expected
given their recent life stress. Thus, low levels of physiological
reactivity to acute interpersonal stress might be a marker of more
persistent resilience to the health impact of chronic life stress
in the elderly.

Using the lifted wavelet transform to denoise ERP data
Jonathan W. King & Marta Kutas
University of California-San Diego
Signal-to-noise ratio is of extreme importance to many psychophysiological
measures. Boosting signal strength can be difficult in practice, so
much work has gone into "denoising" data. The efficacy of classical,
Fourier-based methods depends on a possibly unlikely separation of
signal and noise spectra. The signal processing community has more
recently developed new, highly effective methods using wavelet transforms
to denoise data. Essentially, these methods transform the signal into
a set of wavelet coefficients that differ in "local frequency" and
scale, eliminate or attenuate those coefficients that are close to
zero, and then invert the transform to yield a smoother, less noisy
signal. These techniques are nearly optimal under a wide variety of
circumstances (see, e.g., Donoho & Johnstone, 1995, also http://stat.stanford.edu/~donoho/).
Very recently, Sweldens (1996) has published work on the so-called
"lifted" wavelet transform, (see also http://cm.bell-labs.com/who/jelena/wavelet.html).
This "second generation" approach to wavelet analysis is easier to
understand and to implement than the first generation and is computationally
more efficient (i.e., it can be computed "in-place" on large data
sets without any additional memory). More importantly, it is easily
extended to the analysis of data on (irregular) spatial arrays of
scalp electrodes. I will demonstrate the advantages of lifted wavelets
in the analysis and denoising of (1) single-trial EOG, (2) single-channel
ERP data with very modest number of trials, and (3) multichannel ERP
data obtained from geodesic electrode nets, where we can both spatial
and temporal information to denoise data

Spontaneous skin conductance activity as a marker of task engagement: Refining the relations
Leslie D. Kirby & Craig A. Smith
Vanderbilt University
Spontaneous skin conductance (SC) activity has been shown to be associated
with task engagement (e.g., Pecchinenda & Smith, 1996, Cognition &
Emotion, 481-503). However, task engagement involves both physical
effort and concentration (which, itself, can be either internally
or externally directed), and it is unclear to which of these SC activity
is most closely related. To examine this issue, three facets of task
engagement were factorially manipulated in a within-subjects design:
For physical effort, participants applied a high (100 kilopascals
[kps]) or low (25 kps) level of pressure to a recording device with
their index finger; for external vigilance, they tried to maintain
the pressure within a narrow (+/-2 kps) or wide (+/-10 kps) bandwidth;
and for internal attention, they attempted to remember a long (7)
or short (3) series of words. All three tasks were performed simultaneously
in a series of 30 s trials. Increased effort and increased memory
load were associated with increased skin conductance responses (SCRs)
within a trial; however, increased vigilance (i.e., a more narrow
bandwidth) was associated with fewer SCRs. The average magnitude of
the SCRs increased with effort, especially under conditions of high
vigilance. In addition, within-subjects, and across trials, the number
of SCRs was negatively correlated with how calm, and positively correlated
with how interested, participants reported being during the trial.
These results suggest SC activity is related to multiple facets of
task engagement, but that there appears to be specificity in these
relations across both SC parameters and facets of task engagement.

Topography of the post-imperative negative variation (PINV) in schizotypal personality
Ch. Klein1, P. Berg2, E. Hafstad1, & W. Mueller1
1University of Freiburg, 2University of Konstanz
Previous studies have revealed that subjects with high scores on the
German translation of the Schizotypal Personality Questionnaire (Raine;
SPQ-G) and schizophrenic patients do not exhibit the right-frontal
predominance of the PINV that is seen in (healthy) control groups
but, instead, a less asymmetric distribution. The present study aimed
at replicating the former result, using the same hardware and software
as in the original study, but a different laboratory and a new sample.
Thirty subjects were selected from a sample of about 500 university
students screened with the SPQ-G. Low- and high-scoring groups wer
defined by a median split of the sample distribution of SPQ-G scores.
The EEG and EOG was recorded with a 32 channel DC amplifier (MES,
Munich) and a digitization rat e of 100 Hz (DC to 30 Hz bandpass).
The PINV was elicited with a modified visual delayed matching-to-sample
task, which comprised randomly interspersed trials with clear (66%
of the trials) or ! ambiguous (34%) matching of the S2 with the S1
presented 4 s earlier. Grand average curves and maps reveal a bilateral
frontocentral PINV under both conditions, with larger amplitudes following
ambiguous as compared to clear S2. The groups exhibited condition-dependent
differences in the PINV topography which differed from those obtained
in the original study. Research was supported by the Deutsche Forschungsgemeinschaft
(DFG; Kl 885/4-1)

Schizotypal personality and the antisaccade task
Ch. Klein, G. Bruegner, F. Foerster, W. Mueller, & A. Schweickhardt
University of Freiburg
While disturbed smooth pursuit eye movements have repeatedly been
found in subjects with schizotypal personality features, confirming
the putative nosological relatedness to schizophrenic disorders, so
far only one study indicates antisaccade deficits in this group as
well (Holzman et al, 1995). The present study investigated saccadic
eye movements in schizotypal subjects selected by the German adaptation
of the Schizotypal Personality Questionnaire (Raine, 1991; SPQ-G).
Forty-two university students participated in the experiment which
comprised the pro- and antisaccade tasks under standard overlap, 100
ms gap, and 200 ms gap conditions of fixation and target stimulus
presentation. The sample was divided into two groups by a median split
of the SPQ-G score distribution. Seventy-five leftward and 75 rightward
saccades were elicited under each of the 6 experimental conditions.
Antisaccades were significantly slower than prosaccades, and were
associated with about 15! % direction errors. Under the gap conditions
saccadic reaction times were significantly reduced when compared to
the overlap conditions, and this effect was stronger for pro- as compared
to antisaccades. The saccadic latency reduction following 100 ms gaps
as compared to the overlap condition was significantly larger in high-
than low-scoring subjects for prosaccades, with the reverse group
differences found for antisaccades. These effects were, however, of
small absolute magnitude only. Substantial overlap in the proportions
of antisaccade direction errors and no overall saccadic reaction time
differences between the groups were found. These results suggest only
minor interindividual differences in oculomotor control along the
personality continuum defined by the SPQ-G. Research was supported
by the Deutsche Forschungsgemeinschaft (DFG; Kl 885/4-1)

Situational factors and task difficulty influence ANS reactions during laboratory arithmetic tasks
Kriss Kline, Kathi Heffner, Ben Hayes, Gerald P. Ginsburg, & Melanie E. Harrington
University of Nevada, Reno
Arithmetic tasks are frequently employed in studies of autonomic reactivity,
but task difficulty is rarely manipulated explicitly. The present
pilot study examined the effects of task difficulty and order of presentation
on multiple measures of autonomic reactivity. 14 college students
were randomly assigned to one of two conditions: 7 subjects saw increasingly
difficult problems and 7 saw increasingly easy problems. All subjects
had to indicate whether the answer presented with the problem was
correct. Heart rate [HR], finger pulse volume [FPV], and skin conductance
level [SCL]) were recorded during a 10 minute baseline period and
five 20-sec trials of arithmetic. Change scores for each ANS measure
were calculated as deviations from baseline for the initial and final
math trials. Analyses revealed that task difficulty and order had
differential effects for each ANS measure. Change in SCL was greatest
for difficult problems, when encountered first [Order X Difficulty,
F(1,12)=8.6, p<.05]. For HR change, there was a trend for an interactive
effect [Order X Difficulty, F(1,12)=3.4, p=.09]. Unlike SCL, where
difficult problems accentuated order effects, HR change du ring difficult
problems was similar for the two groups. However, easy problems led
to higher HR increases in Ss who encountered those problems first.
The present results suggest that task difficulty and its interaction
with other experimental factors may be important predictors of autonomic
reactions during arithmetic tasks, and SCL, HR and FPV responses are
differentially influenced by task difficulty and suggest the need
for physiologically-specific hypotheses in experiments using math
problems as stressors.

Relative EEG alpha diminishes to low concentrations of Iso Amyl Acetate (IAA): Sensation without perception
John P. Kline1, Gary E. Schwartz2, Lynn A. Myers2, Douglas J. Streibich2, Ziya V. Dikman2, & Ernest H. Polak2
1Eastern Washington University, 2University of Arizona
We have previously demonstrated EEG alpha decreases to odors when
behavioral detection was at near chance (55%) levels in a forced choice
task (See Schwartz et al., 1994). Here we extend these findings to
include dose/response information. 34 participants received eight
double-blind trials each of a water control, 0.0001, 0.001, 0.01,
and 1 ppm IAA in water paired with water only. Order was randomized
within blocks, and odorant vs. water order was counterbalanced. Detection
accuracies were significantly above chance for .001 and above (44
49, 62, 85, and 98% respectively). Mean confidences (0-10 scale) for
each concentration were 2.5, 2.6, 3.5, 5.3, and 9.0 respectively.
EEG was recorded from 19 channels referenced to linked ears, and digitized
at 256 Hz (band pass 2 to 64 Hz). Impedances were below 5 Kohm. Relative
alpha (8-12 Hz) magnitude significantly decreased over the entire
scalp in response to .001 ppm concentrations and above, suggesting
that the EEG was only as sensitive as the behavioral measures. However,
when the EEG recorded on trials for which participants were correct
was compared with those where participants were incorrect for water-water,
.0001, and .001 ppm, alpha decreased on correct trials for the .0001
ppm, but not for the water-water. EEG alpha decreased for IAA relative
to water for .001 ppm regardless of accuracy. The EEG appeared to
be more sensitive to registering low concentrations of IAA than were
accuracy, confidence, and intensity ratings. The data suggest that
odors may be sensed without being perceived.

Event related potentials during Multi-Attribute Task Battery performance
Timothy F. Knebel
P300 amplitude has been reported to reflect resource allocation to
perceptual tasks. The sensitivity of P300 to motoric tasks, including
compensatory tracking, is not clearly established. Little has been reported
about the effects of workload and perceptual versus motoric task effects on
other ERP components. The current study reports on significant task-related
amplitude differences for ERP components: N100, P200, and P300. Twenty-one
adult volunteers performed the System Monitoring, Tracking, and Resource
Management Subtasks of the Multi-Attribute Task Battery. The tasks were
performed in single, dual, and triple-task conditions. The visual warning
signals of the System Monitoring task were used to drive the generation of
auditory warning tones in a pseudorandom odd-ball sequence including
standard tones as well. For both tone stimuli, P300 and P200 amplitudes were
greater in the single-task than in either of two dual-tasks, or the
triple-task. N100 amplitude was greater in the single-task than in t
Tracking dual-task or triple task. For the warning stimuli only, P300
amplitude was greater for the Tracking dual-task than for the Resource
dual-task or triple-task.

Operant self-regulation of the brain without motor mediation
Andrea Kuebler, Boris Kotchoubey, Hans-Peter Salzmann, Hans Schleichert, & Niels Birbaumer
University of Tuebingen
The results presented here answer two questions: (1) is it possible
to learn self-regulation of the brain with the motor system being
separated from neural centres and (2) can brain self-regulation be
used for communication in a brain-computer interface ( BCI)? For both
questions the answer is yes. In order to construe a BCI for communication
it is necessary to generate at least two brain responses rather quickly.
Therefore we developed a biofeedback technique, where subjects learn
to control their slow co rtical potentials (SCP) at the vertex in
a 2-seconds-rhythm producing either cortical positivity or negativity
according to the task. This paradigm was used to train two patients
with severe motor paralysis. One patient was locked-in following a
brain ste m infarct, and the other suffered from amyotrophic lateral
sclerosis (ALS) with only two small facial muscles left to move (M.
levator labii superioris alaeque nasi and M. depressor anguli oris).
After 4-6 weeks training both patients differentiated significantly
between negativity and positivity. Thus an intact motor system is
not necessary for self-regulation of the brain. Since Neal Miller's
failure to replicate his data on visceral learning in the acute curarized
rat preparation (Dworkin et al., 1986, Behavioural-Neuroscience 100
(3), 299-314) we are the first to present data indicating the existence
of an operant learning for brain self-regulation independent of the
motor system. With the ALS patient training sessions have been continued
for subseque nt 4 months. This prolonged training resulted in perfect
self-control, i.e. the patient is now able to consistently produce
negativity or positivity on demand with more than 90 % accuracy. This
enables the patient to use this skill for communication. Supported
by the German Research Society (DFG).

Physiological predictors of adjustment to retirement in long-term married couples
Cenita S. Kupperbusch, Rachel Ebling, & Robert W. Levenson
University of California-Berkel
Among the major late-life marital transitions, retirement stands out
as requiring spouses to make significant adjustments, including adjusting
to the greater amount of time spent together. In previous research
with young couples, physiology and affect during marital interaction
have been found to predict subsequent changes in marital satisfaction
(e.g., Levenson & Gottman, 1985). In this study, we asked whether
these kinds of variables would also predict how well older couples
will fare in retirement. In 1989, we measured autonomic and somatic
physiology, self-reported affect, and affective behavior during three
15-minute conversations (events of the day, conflict, pleasant topic)
between spouses in long-term marriages (mean duration = 40 years).
Five years later, we contacted those couples who had retired and assessed
how satisfied they were with retirement (N = 59 couples). Low levels
of somatic activity and high levels of vasodilation in husbands during
conversations in 1989 predicted greater retirement satisfaction five
years later. Additionally, husbands' reports of positive feelings
and husbands' positive affective behaviors during conflict conversations
in 1989 predicted their greater retirement satisfaction five years
later. Physiological and affective data obtained from wives in 1989
did not predict their retirement satisfaction. We interpret these
findings as indicating that husbands who were physiologically relaxed
and affectively positive when interacting with their wives prior to
retiring are most likely to be satisfied with their retirement, a
time in which interaction between spouses increases dramatically.

MMN and auditory event-related potentials during hypnotically suggested deafness: Effect of hypnotizability level
Juan R. Lamas1 and Helen J. Crawford2
1University of La Coruna, 2Virginia Polytechnic Institute & State University
It is has been hypothesized that high hypnotizable ("highs") persons
are better than low hypnotizables ("lows") at ignoring distracting
stimuli. Prior research has shown that hypnotically suggested alterations
of perception may have an impact on auditory event-related potentials
(ERP) of highs but not lows. The present study examined the effects
of hypnotizability level and suggestion of deafness in the processing
of physical changes in unattended auditory stimuli. Special attention
was paid to the Mismatch Negativity (MMN), thought to be the first
ERP marker of stimulus change detection (Naatanen, 1992). Using an
odd-ball paradigm [standard 982 Hz (80%) and randomly dispersed 1040
Hz (20%) 60 dba tones, constant 500 ms ISI], auditory ERPs were recorded
at nineteen standard locations from 17 lows and 12 highs while performing
the Revised Minnesotta Paper Board Test in two counterbalanced conditions:
without specific instructions and posthypnotic deafness suggestion
1000 tones were presented in each condition. Highs reported less perceived
loudness than lows in the normal condition, and almost complete or
complete deafness during the suggestion. In both conditions, difference
waves (deviant minus standard) revealed highs had greater positivity
between 200 and 300 ms, located in the right fronto-central regions
(FP2, F4, F8, CZ, C4, T4) and bilateral posterior regions (P3, PZ,
P4, T5, T6, O1, O2). In the deafness condition, both groups showed
similar decreases in positivity around 300 ms post-stimulus in central
and posterior sites (C3, CZ, P3, P4, T5, O1, O2). Analyses revealed
that early detection of stimulus change (MMN) is independent of hypnotizability
level and not affected by hypnotically suggested deafness.

Topographical ERP study of long-term brain effects of cranial irradiation in girls treated for leukemia
I. Lamothe1,2, P. Robaey1, S. PrÈcourt1,2, S. KabËne1, & A. Moghrabi1
1Ste-Justine Hospital, 2University of Montreal
In children treated for acute lymphoblastique leukemia, cranial irradiation
(CI) and intrathecal chemotherapy (IT-C) decrease the risk of meningeal
relapse but lead to cognitive and attentional sequelae, especially
in irradiated girls. Attentional processes were studied in 29 girls
aged 7-11:10 treated with IT-C, 9 with IT-C+CI and 10 normal controls.
While no difference was found on age, irradiated group showed lower
mean IQ than controls. ERPs (P300 and Slow wave) were recorded by
using a visual oddball task with 30 electrodes equally spread on the
scalp. In contrast with results described in previous studies using
more heterogeneous groups, the P300 recorded over the posterior electrodes
was not significantly delayed neither decreased in amplitude. However,
the IT-C+CI group showed a delayed parietal negative Slow Wave (1246ms±172)
as compared to the IT-C (1125ms±94) and the controls (1093ms±119)
(K-W(c2=7,15);p<0,05). In addition, the posterior P300 and the parietal
SW latencies were negatively correlated with the IQ (Spearman's rho>-0,4;p<0,05).
With regard to the ERP waves recorded on the prefrontal leads, the
IT-C+CI group tended to show an earlier P600 (593ms±128) than the
IT-C (743ms±145) and the control (679ms±128) groups (K-W(c2=4,73);p=0,09).
The amplitude of this prefrontal positivity was found to be significantly
correlated with reaction times, but only in the control group (Spearman's
rho=0,8;p<0,05). These results suggest that delayed late ERP components
are related to the lower IQs observed in irradiated girls. Moreover,
a deficit of frontal inhibitory processes may be associated with anti-cancer
treatment, especially with brain irradiation.

Attention, arousal, and television viewing
Annie Lang, Paul D. Bolls, & Robert F. Potter
Indiana University
Arousing television messages attract larger audiences and are remembered
better than calm messages. Thus, when the content of a message is
not intrinsically arousing (say Japanese cooking shows or how to quilt)
producers often try to add arousal by jazzing up the production --
frequently through the use of fast paced cutting and editing. Previous
research suggests that viewers are aroused by both content and pacing
but that the arousal elicited by content may be reflective of an emotional
dimension of arousal while that elicited by pacing may reflect an
attention dimension of arousal. This research explores whether (a)
viewers perceive both message content and pacing as arousing, (b)
if pacing elicits physiological responses indicative of attention
(decreased HR), while (c) arousing content elicits physiological responses
indicative of emotional arousal (increased SCL). 30 adult subjects
viewed thirty, 30-second segments of television programming taped
off air. Heart rate and SCL were collected time locked to message
presentation. Between messages, subjects rated how aroused they felt.
Results showed that both fast paced production and arousing content
increased viewers' self-reported arousal. Messages with arousing contents
resulted in lower heart rate levels compared to messages with calm
contents (suggesting increased attention to the arousing contents).
Content arousal resulted in increased SCL only when the messages were
slow paced. On the other hand, fast paced messages increased SCL (for
both calm and arousing messages) but had no effect on HR. Thus, the
arousal elicited by content and pacing appears to be psychologically
similar, but physiologically different.

Limitations in daily activities and physiological arousal in patients with multiple unexplained symptoms
Wolf Langewitz & Hartmut Schaechinger
University of Basel
Current concepts on the origin of multiple unexplained symptoms (MUS)
focus on the heightened sensitivity to bodily sensations and their
'catastrophic' interpretation leading to health seeking behavior.
A recent publication found that women with MUS show larger physiological
arousal in response to a cold pressor task than normal controls suggesting
a biological basis for the increased awareness of physiological changes
(Cameron et al. 1995, Psychosom Med 57:37-47). We investigated whether
physiological reactions to mental arithmetic (m.a.) and to a reaction
time task (rtt) are related to the amount of MUS (List of Functional
Complaints (LFC)) and/or to the intensity of Limitations in Daily
Activities (LDA) within a cohort of 178 men with MUS severe enough
to justify an in-hospital rehabilitation treatment. Whereas MUS alone
were not related to physiological arousal, LDA were: After dividin
the cohort into quartiles (Q1-Q4), higher limitations in daily activities
were related to lower physiological arousal during the rtt (test-induced
systolic/diastolic blood pressure increase in Q1: 17±8/10±5; Q2: 15±10/9±5;
Q3: 14±9/8±5; Q4: 11±6/6±4 mmHg; F-tests: p<0.05); results from the
m.a. were similar. Conclusion: There is no heightened reactivity to
mental tasks in men with multiple unexplained symyptoms. There is
rather a tendency for men with more severe MUS to display lower physiological
reactions. We believe that this result supports the importance of
cognitive and perceptional factors in the interplay of bodily sensations
and health seeking behaviour.

Resting EEG alpha power asymmetry: A comparison of asymmetry metrics and references
Christine L. Larson, Jeffrey B. Henriques, Bonny D. Donzella, Isa V. Dolski, Steven K. Sutton, & Richard J. Davidson
University of Wisconsin-Madison
Using a large sample, we examined three methodological issues concerning
resting EEG asymmetry. First, a common index of hemispheric activation
asymmetry is the right - left log alpha power difference score. It
has been suggested that residualized scores are more reliable than
difference scores. Therefore, we examined the stability of residualized
scores (regressing left hemisphere alpha power on alpha power in the
homologous right hemisphere lead) compared to difference scores. Second,
we compared test-retest stability of asymmetry scores computed using
three different references: derived linked ears, average, and Cz.
Finally, we compared internal consistency of asymmetry scores from
the three different references. 175 undergraduates completed two sessions,
6 weeks apart, during which 29 channels (13 homologous pairs) of resting
EEG (eight 1-minute trials) were collected. Artifact-free data were
re-referenced to derived linked ears, average, and Cz references.
Mean alpha (8-13 Hz) power across the eight trials in each session
was computed. Correlations between difference scores and residualized
scores were high (e.g., r > .89 for all pairs using derived ears reference).
In addition, correlations between the first and second EEG assessments
were virtually identical for difference and residualized scores. Test-retest
correlations were higher for the average reference than for derived
linked ears or Cz references. Internal consistency was nearly identical
for the three references. These data demonstrate that when using resting
EEG asymmetry measures, 1) right - left difference scores are essentially
equivalent to residualized scores and show equal test-retest reliability,
and 2) the average reference is more stable across time.

Additional cardiac output reactivity: Examination of a novel active coping task
Mark R. Larson & Alan W. Lang
Syracuse University
The primary intent of this study was to examine the effectiveness
of a newly-designed active coping task at eliciting cardiovascular
reactivity. The task, a novel combination of mental arithmetic and
shock-avoidance paradigms, was evaluated in terms of additional cardiac
output (ACO) using impedance cardiography. Subjects were 48 young
male volunteers who each completed two experimental sessions. One
session included a 20-minute baseline period followed by a progressive,
graded exercise task involving three levels (25, 50 and 75W). The
second session also included a 20-minute baseline, followed by the
active coping task. Sessions were counterbalanced across subjects
to eliminate order effects. Steady-state cardiac output and VO2 values
assessed during the exercise task were used to determine ACO during
all baseline and task periods. ACO reactivity values were then calculated
by subtracting pre-task baseline ACO from task ACO. Results indicated
that the active coping task elicited significant amounts of ACO reactivity
(df = 47, t(one-tailed) = 3.173, p < .002). In fact, ACO was shown
to increase by an average of .548 liters per minute in response to
the task. Studies utilizing active coping stressors often attribute
inconsistent or inconclusive findings to the inability of the task
to elicit sufficient amounts of sympathetically-mediated cardiovascular
reactivity. The present active coping task may help to eliminate such
hindrances, as it was found to elicit a significant amount of reactivity
even after metabolic factors were taken into account.

The location and time-course of the brain regions active during task switching: PET and EEG evidence
Erick J. Lauber1, David E. Meyer2, & Jeffrey E. Evans2
1University of Georgia, 2University of Michigan
Many recent investigators are turning their efforts toward understanding
the executive-type processing or strategic control that the mind appears
to exert over its lower-level processes. We have attempted to capture
some subset of these operations by using the 'task switching' paradigm
in our behavioral and neuroimaging research. Our recent PET results
have localized the brain regions involved in shifting one's mental
set between two simple cognitive tasks to prefrontal and superior
parietal regions. However, limitations in experimental design, and
the poor temporal resolution of PET, have limited our interpretations
of the PET data. Our most recent studies, employing high-density EEG
scalp recordings, have allowed us to further infer that the superior
parietal regions are active in the midst of task performance (approx.
300-600 ms post-stimulus) and that the prefrontal regions are active
during the intertrial interval. Supplemented by over a dozen behavioral
studies, this has allowed us to conclude that task switching causes
response selection and rule activation interference from the other
recently performed task, and that the executive operations of goa
shifting and attention shifting occur during the intertrial interval
and are supported by frontal lobe structures. Our EEG recordings have
also allowed us to conclude that the predominantly left hemisphere
PET activations we observed were not caused by the right-hand-only
task requirement used in the PET scanner. This research program demonstrates
the utility of multi-method, converging sources of data in cognitive

Cardiovascular and electrodermal reactivity to physical and social stress in menopausal women with and without hot flashes
M. Leal1, F. A. Garcia-Sanchez2, J. M. Martinez-Selva2, J. Abellan 1, M. V. Candela1, R. Meseguer1, & L. Carbonell1.
1Centro de Salud de Molina de Segura, 2University of Murcia
Menopausal status has been known to influence cardiovascular reactivity
to stress. This is important since both ovarian changes and stress
are risk factors in developing cardiovascular diseases. We studied
differences in cardiovascular reactivity (heart rate, blood pressure,
finger pulse volume) and electrodermal activity (skin conductance
levels and skin conductance responses) in 37 healthy, menopausal women
aged between 45 and 55, with and without hot flashes, n=22 and n=15
respectively, and a control group of 8 pre-menopausal healthy women
of 35 and 40 years of age. None of them was receiving estrogen replacement
therapy. Physiological activity was recorded at rest, a structured
interview, the cold pressor test, and a recovery period. Women performed
the cold pressor test and the interview task in a counterbalanced
order. All groups showed increased physiological responses to the
stressful tasks. Mixed multivariate analyses of variance showed a
significant Group x Task interaction in skin conductance levels (F6,126=2.92,
p=.011) and skin conductance responses (F6,126=9.78, p<.001). No significant
differences between groups were found in cardiovascular responses.
Menopausal women showed higher reactivity to the stressful situations,
especially during the structured interview. Menopausal women reporting
hot flashes showed higher electrodermal activity than those not experiencing
them. These results give support to the existence of higher electrodermal
reactivity in menopausal women reporting hot flashes. It suggests
higher arousal in this group, although cardiovascular activity did
not show this differential effect. The occurrence of hot flashes is
associated to an increased heart rate, however menopausal women reporting
hot flashes do not show increased cardiovascular activity during stressful

Cardiac slow (Traub-Herring-Mayer) waves among panic disorder and asthma patients
Paul Lehrer1, Nicholas Giardino2, Stuart Hochron1, Patricia Sonsalla1, & Jonathan Feldman1
1UMDNJ-Robert Wood Johnson Medical School, 2Rutgers University
We studied cardiac Traub-Hering-Mayer (slow) waves (THM, .06-.12 Hz)
in 21 young adults (age 18-40) with panic disorder, 83 with asthma,
8 with both disorders, and 33 with neither. Subjects had participated
in a study involving a rest period and four psychological tasks (mental
arithmetic, reaction time, and watching two stressful movies). During
baseline, patients with either disorder showed an elevation in the
amplitude of this wave compared to healthy participants, and the combination
of both disorders appeared to contribute additively. Diagnostic groups
did not differ during the psychological tasks. THM wave amplitudes,
but not respiratory sinus arrhythmia (RSA, .13-.4 Hz), were positively
correlated with urinary epinephrine but not urinary norepinephrine.
RSA and THM waves shared approximately 25% of common variance. Both
were positively correlated with cardiac interbeat interval and negatively
correlated with respiration rate. The results are consistent with
the interpretation that the THM wave is augmented by both sympathetic
and parasympathetic function, and that baseline levels of sympathetic
reactivity are elevated in panic disorder and parasympathetic activity
in asthma. Our data suggest that behavioral factors can override these
baseline tendencies. The data are consistent with the notion that
behavioral interventions can affect autonomic reactivity in both asthma
and panic disorder, and that the THM wave may be a useful measure
for studying autonomic balance noninvasively in psychophysiological

Orthographic and phonological contributions to semantic processing
Alistair M. Lethbridge & Allison M. Fox
University of Wollongong
Event-related potential (ERP) investigations of language have indicated
that the N400 component is sensitive to aspects of semantic processing.
Specifically, it has been suggested that the N400 indexes processes
in which word representations are activated or integrated into the
preceding semantic context. Studies which have manipulated semantic
relatedness, for example, have reported reduced N400 amplitude when
target words are preceded or primed by semantically related words.
The current study sought to further explore the sensitivity of the
N400 to priming by manipulating orthographical and phonological aspects
of the prime. Pseudohomophones differ in orthography but have identical
phonology to real words and therefore offer the possibility of exploring
the interaction of orthography and phonology in semantic priming.
ERPs were recorded from 10 right-handed, native English speaking adults
during a semantic categorization word list task in which target words
were preceded by related words, related pseudohomophones, unrelated
words or nonwords (eg. target - "animal" preceded by primes of "beast"
"beest", "badge" or "bedge"). In line with previous literature, a
significant N400 semantic priming effect was observed, with targets
preceded by an unrelated prime eliciting greater N400 amplitude than
targets preceded by a related prime. Although smaller in magnitude,
a priming effect was also observed in the pseudohomophone condition.
The reduced priming effect to pseudohomophones suggests that both
orthographic and phonological attributes of words contribute to the
amplitude of the N400 component.

Topographic differences in CNV amplitude as a result of preparing for cognitive versus motor tasks
P. Andrew Leynes, Joseph D. Allen, & Richard L. Marsh
The University of Georgia
Two experiments found topographic differences in Contingent Negative
Variation (CNV) while people were preparing for cognitive versus motor
tasks. Experiment 1 found that subtraction procedures can assist in
the process of isolating two types of preparation (i.e., cognitive
and motoric). Experiment 2 found similar results when the motor requirements
of the cognitive tasks were eliminated. Topographic analysis from
both experiments revealed that CNV was more frontally distributed
prior to cognitive tasks and more centrally distributed when the tasks
were predominantly motoric. Experiment 2 also demonstrated that a
recognition memory paradigm can be useful in the investigation of
the psychological correlates of CNV.

A new skin conductance response model: Application to short interstimulus interval paradigms
C. L. Lim1, E. Gordon1, J. J. Wright2 , H. Bahramali1, W. M. Li1, C. Rennie1, P. Clouston1 and G. L. Morris1
1University of Sydney & Westmead Hospital-Sydney, 2Mental Health Research Institute of Victoria
A series of interrelated studies are presented, that demonstrate the
feasibility of scoring skin conductance response (SCR) and skin conductance
level (SCL) in conventional short ISI paradigms. We employed our SCR
sigmoid-exponential model method to score overlapping SCRs (Lim et
el., 1997, Int. J. Psychophysiol., 25:97-109). Electrodermal activity
(EDA) was studied in 50 normal subjects, during an habituation (30
stimuli, ISI 5 s) and auditory oddball paradigm (40 targets, ISI 1
s). Fewer subjects responded in the habituation than oddball paradigm
(88% vs 100%). A total of 1421 SCRs in the oddball (71% of targets)
and 371 SCRs in the habituation paradigm (25% of tones) were identified
and scored. More than 60% of subjects in the oddball paradigm showed
an exponential decline in SCR amplitude across the trial (far more
than in the habituation paradigm). This data was compared with 3
medicated schizophrenics, who showed significantly reduced SCR size
(F=4.99; p=0.028) and prolonged decay time constant (F=17.8; p=0.000).
There were no between group differences in SCL or response onset time,
suggesting a cortical dysfunction. Concurrent EEG alpha and beta activity
in normals, exhibited an exponential increase across the trial and
were negatively correlated with SCR activity (r = -0.61 and -0.64;
p<0.001). These findings are contrasted with models of failed inhibition
in schizophrenia, and interpreted with respect to combined CNS-ANS
measures in conventional "cognitive" short ISI paradigms.

The extinction of conditioned fear-potentiated startle inhumans
Ottmar V. Lipp, Simone M. Baker-Tweney, & David A.T. Siddle
The University of Queensland
Theories of evaluative conditioning, the learning of likes and dislikes,
hold that evaluative learning is qualitatively different from other
forms of conditioning. For example, it is said not to extinguish if
conditioned stimuli are presented alone. Studies in humans that have
used fear potentiated startle as an index of evaluative conditioning
seem to confirm this prediction. In the present study (N=48), three
groups of subjects were asked to rate 10 pictures on the dimensions
of valence, arousal, and dominance. All subjects were trained in a
differential conditioning procedure in which one of the 10 pictures
(CS+) was paired with an electric shock US whereas a second (CS-)
was presented alone. One group of subjects (Delayed) rated the 10
pictures before extinction commenced whereas extinction followed without
interruption for the others. Two groups (Short, Delayed) received
16 extinction trials for each CS whereas Group Long received 32 trials.
All subjects rated the 10 pictures after extinction. During acquisition,
CS+ elicited larger electrodermal responses than did CS- and blink
startle potentiation was larger during CS+. The rated valence of CS+
and CS- decreased across the experiment. Differential electrodermal
responses and differential startle potentiation were not found during
extinction, a result that is not consistent with evaluative conditioning
theory. However, post-hoc analyses revealed that subjects who were
not aware of the stimulus contingencies did not display differential
electrodermal responding, but did display differential startle potentiation.
This finding supports the notion of evaluative conditioning as a qualitatively
different form of learning.

Odor and tone effects on continuous performance tasks
Tyler S. Lorig, David G. Elmes, & Viginia L. Yoerg
Washington and Lee Universi
Considerable controversy surrounds the effects of odors in indoor
air environments. Little research has addressed the question of odor
effects on performance and these studies have led to mixed findings.
The present study attempted to investigate odor effects on task performance
and brain electrical activity. Event-related potentials were recorded
from 30 scalp sites in 12 subjects in a CNV-like paradigm. The subjects
completed a continuous performance task in which they judged whether
the previous S2 was larger than the current S2. Subjects made judgments
on the size of the referents of nouns (car, dog, house, etc.), circles,
and numbers in three distinct and counterbalanced phases.The S1 stimuli
lasted 0.2 s and preceded the S2 by 2 s. These stimuli consisted of
tones (500 or 1000Hz), or odors (butanol or peppermint), or blank
trials. ERPs were collected for the entire S1 -S2 interval and reaction
times were collected to the S2. Analysis of the integrated amplitude
of the portion of the S2 corresponding to the P300 component indicated
an interaction of the modality of the S1 and the type of task being
completed. Amplitude was greatest when odors preceded words and slightly
less when odors preceded numbers. Similar findings were made for a
negative component of the S2 which occurred at approximately 450 ms
post-stimulus.The behavioral findings and other results will be discussed
as to the unique effects of odors on verbal and symbolic tasks. Supported
by NIH grant number ES08838-01.

Hemodynamics during rest and behavioral stress in normotensive men at high risk for hypertension
William R. Lovallo and Mustafa al'Absi
Veterans Affairs Medical Center & University of Oklahoma Health Sciences Center
The nature of prehypertensive hemodynamic patterns is critical to
evaluating competing models of hypertension and the potential for
behavioral factors to contribute to the development of the disease.
Persons at risk for hypertension may show elevated blood pressure
(BP) at rest and during mental stress, however the cardiac outputs
and vascular resistances underlying BPs in high risk (HiRsk) persons
are not well characterized. Studies finding increased cardiac output
and normal vascular resistance in prehypertensives have used invasive
methods or have done testing in a novel place and/or under conditions
of threat, such as threat of shock. We chose 21 HiRsk and 21 low risk
(LoRsk) men using both parental hypertension history and resting systolic
BPs (GE or LT 125 mmHg) from two screenings. We then examined whether
the HiRsk's raised BPs were due to greater vascular resistance or
cardiac output as measured noninvasively on a day of extended rest
and a day of rest with prolonged mental arithmetic and reaction time
tasks using monetary bonuses as motivators. HiRsk men had higher systolic/diastolic
BPs (F's = 74/15, p's < .0001/.0001) at all times, accompanied b
higher vascular resistance at rest and during mental stress (F = 6.6,
p < .02) with negligible differences in heart rate and cardiac output.
The results implicate vascular resistance as the altered hemodynamic
element in BP control in these HiRsk young men when tested in a familiar
environment with minimal task-related threat. These data are consistent
with models of hypertension development which assume vascular factors
to play a role early in the etiology of hypertension.

Children's reading ability and ERPs during spatial and rhyme classifications of letters
Deborah Lovrich1,2, Jeff C. Cheng3, & Drew M. Velting3
1SUNY Empire State College, 2Long Island Jewish Medical Center, 3New York State Psychiatric Institute
During visual letter discrimination tasks, event- related potentials
(ERPs) were obtained from 15 electrode sites in six average and six
impaired reading children matched for IQ and age (respective mean
age of 12.44, SD=1.02, and 12.06, SD=1.28, years). Reading ability
was assessed with standardized tests of passage comprehension and
nonsense word decoding. During the ERP tasks, subjects responded to
target letters (50% of trials) with an enclosed area in the form task
and to letters that rhymed with "e" in the rhyme task. Response accuracy
was similar between the groups. Reaction time was significantly longer
for the impaired group than the average group during the rhyme task.
Analyses of the non-target ERPs yielded condition differences, with
the rhyme P450 more negative in amplitude and distribution than the
form at anterior sites. The non-target ERPs also differed between
the groups, with relatively larger P270 amplitudes for the impaired
group. For the target ERPs, the impaired readers' P450 was more positive
in amplitude and distribution at central sites than the average group.
Generally, the average readers' ERPs were more negative than those
of the impaired group at anterior sites in the 300-800 ms latency
range. Group differences in the ERP did not vary as a function of
condition, suggesting that differences were not due to task demands
of rhyme and spatial classification but to processing common to both
tasks, possibly including automatic letter name retrieval.

Motor programming and learning of finger movements in Parkinson's disease
Kathy A. Low1, Steven A. Hackley1, Fernando Valle-Inclan2, & Mark Stacy1
1University of Missouri-Columbia, 2University of La Coruna
This study utilized the lateralized readiness potential (LRP) and
the contingent negative variation (CNV) as indicators of possible
motor planning differences between patients with Parkinson's disease
(PD) and healthy age-matched controls. Participants performed a warne
choice reaction time task with response requirements varying on complexity.
Each trial began with a warning signal, an arrow pointing right or
left, indicating the hand of response. After two seconds a GO (90%)
or NO-GO (10%) signal was presented indicating whether the specified
response should be performed or withheld. In the sequential condition,
the response consisted of a pattern of three keypresses (e.g., middle-index-ring
finger) while in the single condition, the requirement was to press
only one key. Overall, both response- and stimulus-locked LRPs were
reduced in PD patients compared to controls. However, this difference
was only reliable under the more difficult sequential response condition.
Similar to findings with young adults (Hackley & Miller, Psychophysiology,
230-241), elderly adults exhibited a greater degree of lateralization
prior to sequential compared to single responses. This trend was also
seen in the PD group, but failed to reach significance in the preliminary
analyses. Finally, behavioral measures and learning effects will also
be discussed, but these measures have not yet been analyzed. Given
that PD is a result of neuronal damage in the substantia nigra, the
LRP findings support the contributive role of the basal ganglia in
motor preparation and suggest that this role may be greater with increasing
response complexity.

Baseline corrections on autonomous measures and their influence on the variance components associated with subjects, treatments, interaction, and error
Marc Luxen
Tilburg University
Correction procedures using parameters derived from measurements during
pre, inter, or post treatment rest periods are an integral part of
pyschophysiological research (Stemmler & Fahrenberg, 1989). Choice
of parameter(s) and restperiod(s) can influence specific parts of
the variance in the reactivity. This becomes visible in an ANOVA,
with "treatments" and "subjects" as factors, as differences in the
size of variance components, a derivative of the Mean Square (Shavelson
& Webb, 1991). Heart Rate, Skin Conductance Level, and Blood Pressure
(Systolic and Diastolic) of fourteen young police officers were continously
measured during two films and three restperiods: one before, one after
and one between the films. Means and the standard deviations of combinations
of the restperiods as well as logarithmic and linear regression coefficents
on the scores and standard deviations of all rests together were used
as correction parameters and compared. Using correction parameters
based on the means and standard deviations of the restperiods just
before the treatment resulted in the finding that treatments explained
most of the variance in the reactionscores. Using parameters based
on all rests (the regression parameters and the combined means)result
in more variance explained by subjects. Finally, correction with standard
deviations during rest reduced the variance attributed to the interactio
combined with error. This implies that choice of correction procedure
should be, besides more substantive considerations, based on the specific
hypotheses that are tested.

HTML Coding by Chris Lovelace, 970819