Hemispheric differences in emotional psychophysiology

HEMISPHERIC DIFFERENCES IN EMOTIONAL PSYCHOPHYSIOLOGY

By

BETH S. SLOMINE

A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

1995

ACKNOWLEDGEMENTS

First, I want to thank my Chairperson, Dr. Dawn Bowers,

for teaching me the skills needed to become a competent

researcher and writer. I also want to thank Dr. Russell

Bauer for explaining psychophysiological methodology to me

in a way that I could easily understand. I am also grateful

to Dr. Heilman for being available to answer my questions

about neurology, help me to choose patients, and map out the

CT/MRI scans. I would like to thank my other committee

members, Drs. Bradley, Rao, and Fennell for contributing

their time and expertise to this project.

Additionally, I would like to thank the many people who

provided technical support for this project. Samel Celebi

wrote all the computer porgrams and set up the interface

between hardware and software. Barbara Haas taught me

appropriate electrode placement and forced me to use an

impedence meter. I am also grateful to those individuals at

the West Roxbury VAMC who helped me to finish this project.

Bill Milberg provided me with the time and computer

facilities needed to conduct data reduction and analyses.

Patrick Kilduff patiently helped me to reduce the tremendous

amount of data I had collected. Also, Gina McGlinchy

assisted me with my statistics and never got angry as she

showed me the same steps over and over again.

I would also like to thank the research assistants who

helped me with this project. Hillary Webb, Kim Roberts, and

Brian Howland all helped in heartrate reduction. I would

especially like to thank Scott Lebowitz who worked

diligently on many aspects of the project from subject

recruitment to data management.

I would also like to thank those individuals and

organizations who helped me to find participants for this

study, including Beth McCauley; Anne Rottman, M.D.; Laura

Hodges, P.T.; Orlando Stroke Club, Golden Gators; and the

other seniors groups from local churches who allowed me to

recruit subjects through their organization. And, of

course, I would like to thank all of those individuals who

spent the many hours required to participate in this

project.

Lastly, I would like to thank my family and friends who

supported me and attempted to calm me down through all of my

catastrophizing over the last three years.

iii

TABLE OF CONTENTS

ACKNOWLEDGEMENTS.........................

ABSTRACT .................................

CHAPTERS

1 REVIEW OF THE LITERATURE...........

Theories of Emotion..................
Hemispheric Assymetry of Emotion....
Emotional Psychophysiology..........
Critical Issues .....................

2 STATEMENT OF THE PROBLEM...........

Overview of Experimental Design.....
Hypotheses and Predicitions .........

3 METHODS.............................

Subjects.............................
Baseline Evaluation.................
Experiment 1.........................
Experiment 2 ........................
Design Issues .......................

4 RESULTS......... ....................

Group Data ..........................
Subgroup Data .......................
Individual Case Studies..............

5 DISCUSSION .........................

Differential Responding in Normal
Subjects.........................
Group Differences in Emotional
Responding ......................
Global versus Bivalent Models of
Emotion..........................
Neuroanatomic Correlates.............
Limitations of the Study..............
Future Directions.....................

iv

page

..... ii

..... vi

...1

...2
..15
..38
..51

..56

..59
S.60

..69

S.69
S.73
S.74
..83
..85

..88

..88
.134
.137

.145

..147

..155

..159
..161
..163
..165

APPENDICES

A PSYCHOLOGICAL MEASURES.................167

Self-Assessment Manikin.................167
Positive and Negative Affect Schedule...167

B DEMOGRAPHIC INFORMATION................. 169

C STATISTICAL INFORMATION................185

REFERENCES.................................... 246

BIOGRAPHICAL SKETCH ............................263

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

HEMISPHERIC DIFFERENCES IN EMOTIONAL PSYCHOPHYSIOLOGY

BY

Beth S. Slomine

August 1995

Chairperson: Dawn Bowers
Cochairperson: Russell M. Bauer
Major Department: Clinical & Health Psychology

Two theories have been proposed to explain the

organization of emotions within the cortical hemispheres.

According to the global right hemisphere model, the right

hemisphere takes a predominant role in modulating emotions.

Based on the global theory, patients with right hemisphere

damage (RHD) have a deficit in emotional processing of all

emotions. According to the other hemispheric theory of

emotion, the bivalent model, the right hemisphere modulates

negative emotions and the left hemisphere modulates positive

emotions. This model predicts that RHD patients would be

deficient in emotional processing of negative emotions,

whereas patients with left hemisphere damage (LHD) would be

impaired in processing positive emotions.

In this study, emotional experience as measured by

autonomic responding, facial muscle activity, and verbal

report was examined in 12 patients with RHD, 12 patients

with LHD, and 24 normal control subjects (NCS) during

anticipation of shock and reward. Results revealed that

during the shock condition, RHD subjects displayed a deficit

in skin conductance responding compared with the NCS, but

not compared with the LHD subjects. None of autonomic or

facial muscle variables differentiated the reward from the

control condition during the reward task. These results are

discussed in light of the global and bivalent theories of

emotion as well as neuroanatomic correlates of electrodermal

activity.

vii

CHAPTER 1
REVIEW OF THE LITERATURE

Introduction

Patients with unilateral brain damage have been used to

investigate hemispheric contribution to emotional

perception, voluntary expression, and to a lesser extent

"experience" as indirectly assessed through physiologic

arousal, overt behavior, and verbal report. Although some

studies have suggested that differences in post-stroke mood

occur following right hemisphere damage (RHD) and left

hemisphere damage (LHD), few studies have assessed brief

emotional experience while measuring psychophysiological and

behavioral indices of emotion in these patients. Moreover,

when emotional experience has been studied using

physiological indices of emotion, patients needed to decode

emotional stimuli, which may be problematic for some RHD

patients. Additionally, no study to date has employed

facial EMG when examining emotional experience in

unilaterally damaged patients. In the current project,

stroke patients with left or right hemisphere lesions

participated in two experiments designed to examine specific

deficits in pleasant and unpleasant emotional experience as

a function of unilateral brain damage. Both physiological

2

responding, facial behavior, and verbal report were measured

during "in vivo" affective situations.

Before discussing the experiments further, a brief

overview of the literature is provided. The review includes

prominent theories of emotion which have stemmed from the

works of James (1884/1922), Lange (1922), and Cannon (1927).

Moreover, theories of hemispheric specialization of emotion

are provided. Specifically, two predominant

neuropsychological theories of emotion are explored: (1)

the global right hemisphere theory which states that the

right hemisphere is responsible for affective processing;

and (2) the bivalent view which conceptualizes the right

hemisphere as predominant for negative emotions and the left

hemisphere as predominant for positive emotions. In

addition, studies of hemispheric differences in emotional

evaluation, expression, arousal, and mood are discussed.

Lastly, an overview of emotional psychophysiology is

presented.

Theories of Emotion

The quest to understand emotion has stimulated the

development of many theories and much empirical data over

the past century. According to Kleinginna and Kleinginna

(1981) the numerous definitions of emotion complicate

research in emotion. After an extensive review of emotional

definitions, they classified psychological definitions of

emotions into 11 non-mutually exclusive categories on the

3

basis of emotional phenomenon and theoretical issues. They

concluded that a definition of emotion should be broad

enough to include significant aspects of emotion, but still

be able to distinguish emotion from other psychological

phenomenon. They suggested the following definition:

Emotion is a complex set of interactions among
subjective and objective factors, mediated by
neural/hormonal systems, which can (a) give rise to
affective experiences such as feelings of arousal,
pleasure/displeasure; (b) generate cognitive processes
such as emotionally relevant perceptual effects,
appraisals, labeling processes; (c) activate widespread
physiological adjustments to the arousing conditions;
and (d) lead to behavior that is often, but not always,
expressive, goal directed, and adaptive. (p. 355)

Like the numerous definitions of emotions, there are many

theories of emotion. These differ in their

conceptualization of emotional experience and the role of

cognition in emotional experience. A few prominent emotion

theories are described below.

James-Lanqe versus Cannon Debate

James (1884/1922) and Lange (1922) were the first to

challenge the common sense view that perception of an event

was followed by the experience of emotion. James stated

that "...the bodily changes follow the perception of the

exciting fact, and that our feelings of the same changes as

they occur is the emotion" (p.13). James proposed that, in

order to experience emotion, one must simultaneously exhibit

physiological and expressive changes, such as tensed muscles

and quickened heart rate during fear. Specifically, the

James-Lange theory states that perception occurs when an

4

object stimulates one or more sense organs relaying afferent

impulses into the cortex. Next, cortical efferents send

information to skeletal and visceral musculature producing

complex changes. Lastly, sensory information from the

affected musculature is projected back to the cortex.

Perception of this sensory information produces the

experience of emotion. In the early 20th century, the

James-Lange theory predominated the study of emotion (Izard,

1977).

In 1927, Cannon presented five criticisms of James-

Lange's hypotheses that perception of autonomic/visceral

changes are responsible for the experience of emotion.

First, Cannon cited evidence that spinal cord transactions

in dogs, in which the sensations of the viscera were

separated from the CNS, did not alter emotional experience.

Additionally, he stated that cats who had their entire

sympathetic division of the autonomic nervous system removed

showed all the manifestations of rage when presented with a

dog (i.e., hissing, growling, and retraction of the ears)

except the cats did not raise the hairs on their backs.

Second, Cannon pointed out that the same visceral changes

occur during sympathetic arousal even though different

emotion states may be experienced. Additionally,

sympathetic arousal produces similar changes in non-

emotional states such as fever or exposure to cold. Third,

Cannon argued that the viscera are relatively insensitive

structures and changes are often not experienced

consciously. Fourth, he stated that visceral changes are

slow and thus, cannot be a source of emotion. Fifth, he

claimed that producing artificial visceral changes does not

produce affect. He used adrenalin as an example stating

that adrenalin produces bodily changes that are not

accompanied by affective states. He concluded that the

sensation of visceral responses cannot produce affect.

Cannon hypothesized that "emotional expression results

from action of subcortical centers" (p.115). Cannon cited

studies in which various types of decorticate animals

displayed abnormal affective responses, whereas animals with

hypothalamotomies failed to display affective behavior.

Consequently, Cannon concluded that the cerebral cortex

normally inhibits thalamic activation. He purported that

during normal emotional experience sensory information

arrives at the cortex and is projected to the hypothalamus

releasing it from cortical control. Cannon proposed that

hypothalamic activation relays information to somatic

musculature and smooth musculature of the viscera to produce

characteristic manifestations of emotion. Simultaneously,

the hypothalamus projects to cortex which produces the

conscious awareness of emotion. According to Cannon

muscular changes, visceral changes, and conscious experience

of emotion all occur simultaneously. The result is intense

emotional experience accompanied by behavior and

physiological indices of emotion.

Later scientists elaborated on Cannon's theory. Papez

(1937) postulated that a circuit of emotion exists that

relays information to the hypothalamus from the anterior

thalamus, cingulate cortex, and hippocampus. He posited

that emotion originates in the hippocampal formation and is

relayed through the above circuit to the cortex. He

described the cingulate gyrus as the receptive cortical

region for emotion. About a decade later, MacLean (1949,

1952) described the limbic system as a group of

phylogenetically old cortical structures that are involved

in emotion.

More recently, LeDoux (1989) has argued that emotion

and cognition are mediated by separate though interacting

neural systems. According to LeDoux, the amygdala is the

major component of the brain's affective processing system,

whereas the hippocampus is critically involved in cognitive

processing. Both affective and cognitive computations can

occur without conscious awareness. According to LeDoux,

affective computations occur via thalamo-amygdala

projections which process the affective significance of

simple sensory cues, whereas the cortico-amygdala pathway

processes complex affective stimuli. The thalamo-amygdala

projections are adaptive because this pathway often leads

directly to motor responses with brief processing time,

i.e., fleeing from a dangerous snake. LeDoux proposed that

the amygdala receives exteroceptive sensory, interoceptive

sensory, and neural input. In addition, LeDoux (1984)

explains that sensory information from the peripheral

nervous system feeds back to the amygdala to intensify

amygdala excitation and increase the duration and intensity

of the experience of emotion.

LeDoux suggested that the amygdala performs the

functions that Cannon (1927) and Papez (1937) thought

belonged to the hypothalamus. Together, Cannon, Papez, and

LeDoux challenged the James-Lange Theory in hypothesizing

that emotional experience can be generalized in the brain

without the participation of the peripheral nervous system.

However, none of these theories discuss the differing roles

that the right and left cerebral hemispheres may play in

modulating emotional behavior.

Appraisal Theories

Other theorists have attempted to address Cannon's

criticism of autonomic feedback proposed by James and Lange.

Russell (1927/1961) stated that cognition as well as

physiological feedback compose the experience of emotion.

Within the past few decades, some theorists have viewed

emotion as a phenomenon developing from cognitive appraisal

of an event, situation, or condition. Arnold (1960)

described emotion as the nonrational judgement of an object

which follows perception and appraisal. Schacter and Singer

(1962) proposed that physiological arousal along with

cognitive appraisal are both essential for emotion to

result. They suggested that some event or condition creates

physiological arousal which is combined with evaluation of

the event or condition (cognitive appraisal) to lead to the

experience of emotion.

Central to appraisal theories is the view that the

experience of any emotion (i.e., joy, anger, fear) involves

the same physiological arousal, but different cognitive

appraisals. Lazarus and Averill (1972) explained that

emotion results from appraisal of stimuli and the

formulation of a response. In their view, appraisal reduces

and organizes stimulus input to a specific concept, (e.g., a

threat). Lazarus and Averill also asserted that personal

psychological structure and social norms also influence

appraisal. Most importantly, they concluded that appraisal

determines the specific emotional experience. For example,

anger has been associated with the perception of goal

obstacles, whereas fear is associated with perceived

uncertainty about and unpleasant situation (Ellsworth &

Smith, 1988). However, these theorists place little or no

emphasis on neural hardware which might underlie or

contribute to appraisal.

Differential Emotion Theory

The Differential Emotion Theory was developed by

Tomkins (1962, 1963) who proposed that awareness of

proprioceptive feedback from facial muscles constitutes the

experience of emotion. According to Tomkins, emotion-

specific innate programs for groups of facial expressions

are stored in subcortical centers. Tomkins hypothesized

that once an emotion has been activated, facial feedback is

provided to the cortex. Additionally, Tomkins argued that

it is the facial feedback that initiates visceral

activation.

Differing slightly from Tomkins, Izard (1977) argued

that emotion involves three components; neural activity or

the density of neural firing per unit time, striate muscle

feedback to the brain, and subjective experience. Izard

posited that each component can be dissociated from the

others, but that the three are normally interdependent.

Specifically, according to Izard, internal or external

stimuli affect the gradient of neural stimulation in the

limbic system and sensory cortex. Information from these

areas are relayed to the hypothalamus which plays a role in

determining the facial expression to be effected. From the

hypothalamus, impulses are relayed to the basal ganglia

where the neural message for facial expression is mediated

by motor cortex. Impulses from motor cortex, via cranial

nerve VII lead to the specific facial expression. Cranial

nerve V receives sensory input from the face and projects,

via the posterior hypothalamus, to sensory cortex. It is

10

the cortical integration of facial expression feedback that

generates subjective experience of emotion.

Proponents of the Differential Emotion Theory have

conceptualized a certain number of fundamental emotion

categories which are comprised of specific phenomenological

characteristics, expressive responses, and physiological

patterns. Darwin (1872) was one of the first to discuss his

observations of the expression of discrete emotions. He

described many emotions which he viewed as having

corresponding facial expressions which are universally

displayed and recognized by humans cross culturally.

According to Izard (1977) there are 10 fundamental emotions

such as happiness, sadness, anger, fear, and disgust.

The concept of discrete emotions developed mostly from

direct observation and study of facial expressions.

Fridlund, Ekman, and Oster (1987) reviewed the literature on

facial expressions including phylogenetic, cross-cultural,

and developmental research. They determined that there is

much support for discrete emotions. Their conclusions,

based on the literature, are as follows: (1) phylogenetic

studies have shown that many nonhuman primates show a

variety of differentiated facial patterns and similar facial

patterns have been observed among human and nonhuman

primates; (2) cross-cultural studies have revealed that

members of different cultures display the same facial

expressions and use analogous emotion labels when

identifying the underlying emotions of posed expressions;

and (3) developmental research has indicated that facial

musculature is fully formed and functional at birth and

infants display many facial expressions similar to adult

expressions. Also, infants demonstrate differential

responses to facial expressions by 3 months of age and have

the capacity to imitate facial movements within the first

few days of life.

One problem not addressed by the differential emotions

theorists is whether spontaneous experience of these

emotions is accompanied by the occurrence of the predicted

facial expression (Davidson, in press). For instance,

Davidson stated that little is known about the incidence of

different facial expressions depending on context or type of

emotion elicitor (i.e., imagery, emotional film clip). For

example, Tomarken and Davidson (1992) found very few overt

expressions of fear in response to fear film clips. Also,

Davidson (in press) raised questions concerning the facial

expressions of positive emotion. Specifically, he indicated

that while there are multiple forms of positive affect as

evidenced using behavioral, subjective, and physiological

indices, there is only one facial expression indicative of

the experience of positive emotion.

Dimensional Approaches

In an attempt to explain the polarity of emotion,

dimensional theorists have conceptualized emotion as varying

on two or three polar dimensions. Wundt (1896) suggested

that emotions can be conceptualized in terms of three

different dimensions: pleasantness-unpleasantness,

relaxation-tension, and calm-excitement. In addition, the

dimensional views of emotion were supported by Cannon's

(1927) claim that the same visceral changes occur in

different emotional states. Consequently, theorists such as

Duffy (1957) conceptualized emotions as varying along a

general state of activation or arousal. Other contemporary

investigators have used dimensions to characterize facial

expressions (e.g., Scholsberg, 1941; Osgood, 1966) and

verbal report (e.g., Russell & Mehrabian, 1977). Lang

(1985) stated that most variance within factor analytic

studies of the verbal report of emotional experience was

accounted for by two dimensions, activation (arousal-

quiescence) and valence (pleasure-displeasure). Because the

bidimensional view seems to neglect a certain amount of

variance, Lang proposed that the dimension termed dominance-

submission by Russell and Mehrabian (1977) may account for

the residual variance.

Similar to the view of the discrete emotions theorists,

Lang (1985) suggested that emotional behavior has developed

phylogenetically for basic survival tasks (e.g., searching

for food or fighting for territory). Further, Lang

hypothesized that the combination of valence (approach vs.

avoid), arousal (energy mobilization), and dominance

(postural stance) are critically important for smooth

execution of behaviors necessary for success in survival

tasks. Lang asserted that it is essential to determine how

emotion is represented in memory in order to ascertain how

emotion drives cognitive processing. Lang proposed that

emotion information is coded within memory in the form of

propositions which are organized into associative networks.

The associative networks are comprised of three tiers;

semantic codes, stimulus representation, and response

programs.

According to Lang's Bioinformational Theory (1979,

1984), emotions are associated with action. Access of

emotional propositions are associated with efferent outflow,

and thus emotion can be measured in terms of three response

systems; verbal report, overt behavior (i.e, facial

expression, body posturing, and emotional prosody), and

peripheral and central physiological measures. However,

only stable networks which are called emotion prototypes,

such as those found in phobics, demonstrate a reliable

behavioral output in all three response systems.

Consequently, emotional experience is an epiphenomenon of

the 3 response systems which reflect an underlying centrally

represented propositional network.

Taken together, theories of emotion differ quite

dramatically in their emphasis on and definition of

emotional experience. James and Lange view emotional

14

experience as the awareness of bodily sensations associated

with emotion. Cannon, on the other hand, views conscious

awareness of emotion as arising from neurological activation

which may be accompanied by visceral and muscular changes.

Papez, MacLean, and LeDoux support this view. Appraisal

theorists emphasize the importance of cognition combined

with physiological arousal in the awareness of emotion.

Discrete emotion theories view the experience of categorical

emotions which corresponded to specific facial expressions.

Lastly, most dimensional theorists emphasize the experience

of emotion based on two or three polar emotional dimensions,

whereas Lang views emotional experience as an epiphenomenon

of overt behavior, physiological activity, and verbal

report.

For purposes of the present study, emotional experience

is defined as a psychological phenomenon or subjective

experience which can be measured indirectly through

physiological measures, verbal report, and overt behaviors

(e.g., facial muscle responses). Because emotional

experience is not directly observable, problems are inherent

in any definition of and attempt to measure it. In terms of

the present definition of emotional experience, it is

unclear what the impact of decreased responding in any of

the three response systems means in terms of emotional

experience. For instance, if an individual reports

experiencing anxiety, but displays no physiological or overt

responses, it is unknown whether there is a disconnection

between experience and motor output or whether the person is

not experiencing the emotion as completely as someone who

reacted with all three response systems.

Hemispheric Asymmetry of Emotion

Along with general psychological theories of emotion,

investigators have examined the organization of emotion in

the brain. Historically, emotion has been associated with

the limbic system (Papez, 1937; MacLean, 1952). More

recently, neuropsychologists have examined the role of the

cerebral hemispheres in modulating emotional behavior.

Research involving neurologically impaired patients has

aided in developing an understanding of how various domains

of emotional behavior (i.e., evaluation, expression,

arousal) are disrupted by focal lesions of the left and

right hemispheres. Based on some clinical studies, along

with findings from normal individuals (see review, Heilman,

Bowers, & Valenstein in press), inferences have been made

regarding the neural networks that might underlie different

aspects of emotional behavior including evaluation,

expression, arousal, and experience.

Early observations of individuals following hemispheric

damage revealed differences in mood reactions depending on

whether the left or right hemisphere was involved.

Babinski (1914) was one of the first to note that patients

with right hemisphere damage (RHD) appeared indifferent or

16

euphoric. Others have reported similar observations (Denny-

Brown, Meyer, & Horenstein, 1952). Denny-Brown et al.

described a 55 year old woman with a right parietal infarct,

who appeared "indifferent" towards her illness as well as

apathetic towards her family's affairs. By contrast,

individuals with left hemisphere dysfunction (LHD) have been

observed to appear depressed, which was termed "catastrophic

reaction" by Goldstein (1948). Terzian (1964) noted that

injection of sodium amytal into the left carotid artery,

which inactivated the left hemisphere, was associated with a

depressive reaction, whereas injection of sodium amytal into

the right carotid artery was associated with an euphoric

reaction. More systematic large-scale studies of RHD and

LHD patients have been consistent with the early clinical

reports. Gainotti (1972) investigated the verbal

expressions and behavior of 160 patients with left and right

hemisphere lesions. Behaviors indicative of catastrophic

reactions or anxious-depressive mood were more frequent

among LHD patients, while indifference reactions were more

prevalent among RHD patients. Observations of post-stroke

mood changes has generated a large body of research over the

past 20 years in an attempt to understand the contributions

of the left and right hemispheres to emotion.

Two prominent theories of hemispheric differences in

emotion have arisen from the clinical studies reported

above. According to the global right hemisphere view, the

right hemisphere is involved in interpreting emotional

stimuli and has a unique relationship to subcortical

structures which mediate cerebral arousal and activation

(e.g., Heilman, Watson, & Bowers, 1983). Consequently,

damage in the right hemisphere interferes with processing

emotional stimuli, programs of expressive behavior, and

cerebral arousal and activation. In contrast, the bivalent

view of emotion posits that the anterior portion of the

right hemisphere is dominant for negative/avoidance emotions

and the anterior region of the left hemisphere is dominant

for positive/approach emotions (e.g., Fox & Davidson, 1984).

According to the bivalent view, right hemisphere damage

causes positive/approach affect and left hemisphere damage

evokes negative/avoidance affect. Both models and the

empirical research in support of each are discussed below.

Global Theory of EI-tion

According to the global right hemisphere model,

observations of emotional indifference in RHD patients can

be explained by the right hemisphere's specialization for

coding nonverbal affective signals and mediating arousal and

activation (Heilman et al., 1983). The global right

hemisphere theory is supported by research exploring

emotional evaluation, expression, and arousal/activation,

which has revealed that RHD patients are deficient in

interpretation of emotional stimuli, are emotionally

flattened, and physiologically hypoaroused. The relevant

research is discussed below.

E'.'alu(L ti of emt: t icn

Most of the research in support of the global right

hemisphere view of emotion has arisen from investigations of

evaluation and perception of affective stimuli (i.e., facial

expression and emotional prosody). Many patients with RHD

have impairments in identifying and discriminating facial

expressions. This research was initially conducted by

DeKosky, Heilman, Bowers, and Valenstein (1980) and has been

consistently replicated across other laboratories (Cicone,

Wapner, & Gardner, 1980; Etcoff, 1984; Bowers, Bauer,

Coslett, & Heilman, 1985). From an historical perspective,

one critical issue was whether the RH superiority in

identifying facial expressions was secondary to the role of

the RH in mediating complex visual configurational stimuli.

Evidence against this view point comes from covariance

studies, individual case reports, and studies which find RHD

patients impaired in identifying facial affect when it has

been verbally described.

First, in covariance studies, visuoperceptual ability

has been controlled for and equated statistically. In these

studies, deficits in RHD patients in recognition of

affective facial expressions have been observed above and

beyond deficits in visuoperceptual ability (Ley and Byrden,

1979; Bowers et al., 1985). Second, case descriptions have

documented dissociations between performance on

visuoperceptual facial recognition and performance on

affective facial expression recognition (Dekosky et al.,

1980). Third, Blonder et al. (1992) found that RHD patients

were impaired relative to LHD patients and NHD controls in

identifying emotion associated with a verbal description of

a non-verbal signal, i.e., he scowled. Similar results

were found in RHD patients compared to LHD patients and

normal controls when asked to imagine facial expressions

(Bowers, Blonder, Feinberg, & Heilman, 1991). Because these

nonverbal affect signals were verbally described, poor

performance of the RHD group could not be attributed to

perceptual impairment.

Taken together, these studies suggest that there are

specific subsystems for processing affective facial stimuli.

This evidence is comparable to findings in the animal

literature. Using single cell recordings, neuroscientists

have identified visual neurons in the temporal cortex and

amygdala of monkeys that responded selectively to faces and

to facial expressions (Perret et al., 1984; Leonard, Rolls,

& Wilson, 1985).

In addition -o deficits in comprehension of emotional

faces, many patients with RHD also have impairments in

understanding emc:ional prosody. For example, many patients

with RHD have difficulty identifying emotional prosody,

which includes the pitch, tempo and rhythm of speech.

20

Discrimination of effectively intoned speech was found to be

worse in patients with RHD in the temporoparietal regions

compared to patients with LHD (Tucker, Watson, & Heilman,

1977; Heilman, Scholes, & Watson, 1975; Ross, 1981).

In addition, there is evidence to suggest that RHD

patients are impaired in understanding nonemotional as well

as emotional prosody (Weintraub, Mesulam, & Kramer, 1981).

Both RHD and LHD patients were impaired compared to NHD

controls in nonemotional prosody, while RHD were more

impaired than the LHD patients in emotional prosody

(Heilman, Bowers, Speedie, & Coslett, 1984). Consequently,

these authors conclude that both hemispheres are important

in comprehension of nonemotional prosody, but the right

hemisphere plays a more vital role in the comprehension of

emotional prosody.

Not all studies find hemispheric specific prosody

dysfunction. Schlanger, Schlanger, and Gerstmann (1976)

found no differences between RHD and LHD patients in

comprehension of emotional prosody; however, only 3 of 20

RHD patients in this study had temporoparietal lesions.

More recently, Van Lancker and Sidtis (1992) found equally

poor affective prosodic recognition in RHD and LHD patients.

Moreover, they determined that LHD and RHD patients use

different cues in attempting to recognize affective prosody.

Specifically, RHD patients tended to use timing cues,

whereas LHD patients used information about pitch. These

authors concluded that affective prosody is a multifaceted

process which cannot simply be explained by differences in

hemispheric specialization.

Studies of normals using dichotic listening tasks have

also been employed to explore hemispheric differences in

processing emotional prosody. In dichotic listening, two

different messages are simultaneously presented to the right

and left ears. Words were recalled best from the right ear

indicative of left hemisphere superiority (Kimura, 1967),

while mood of the speaker was recalled better from the left

ear, suggestive of right hemisphere superiority in

processing emotional prosody (Haggard & Parkinson, 1971; Ley

& Bryden, 1982).

In contrast to the tasks involving nonverbal signals,

evidence for a unique role of the right hemisphere in

mediating emotional understanding of messages that are

conveyed through propositional language is equivocal.

Recognition of emotional words has been found to be better

when presented tachistcscopically to the right hemisphere

(Graves, Landis, & Goodglass, 1981). However, RHD and LHD

patients did not differ in the ability to comprehend the

meaning of emotional and nonemotional words (Morris et al.,

1992), the ability to identify emotionality of short

propositional sentences (Heilman et al., 1984; Cicone,

Wapner, & Gardner, 1980; Blonder, Bowers, & Heilman, 1991;,

or the ability to judge similarity between two emotional

words (Etcoff, 1984).

However, recent evidence contradicts these findings.

Borod et al. (1992) found that, when compared to LHD and NHD

patients, RHD patients were more impaired in identifying and

discriminating emotional words and sentences. In addition,

RHD patients were impaired in their understanding of

emotionality in complex narratives (Gardner, Brownell,

Wapner, & Michelon, 1983; Gardner, Ling, Flam, & Silverman,

1975; Brownell, Michelon, Powelson, & Gardner, 1983). The

deficits of RHD patients in understanding complex narratives

may not be related to emotion, but to difficulties of RHD

patients in drawing inferences, reasoning, and interpreting

figures of speech (Heilman, Bowers, & Valenstein, in press).

However, this explanation does not explain the results of

Borod et al. (1992) who found that RHD were impaired in

identifying and discriminating words and short sentences.

Taken together, the above studies indicate that

patients with RHD have more difficulty than LHD patients and

NHD controls in evaluating nonverbal signals of emotion,

including facial expressions, emotional prosody, and verbal

messages of emotions. Moreover, RHD patients are equally

impaired for both positive and negative emotional signals.

Although some deficits in recognition of facial expressions

in RHD patients are related to general dysfunction in

visuospatial ability, others are apparently independent of

visuospatial ability. In part, some deficits in affective

prosody may be due to more elemental dysfunction in complex

auditory analysis. In contrast to nonverbal affective

signals, the role of the right hemisphere in processing

verbal emotional signals remains unclear. At present, some

argue that an emotional semantic network is widely

distributed between the hemispheres whereas other argue that

the RH may be dominant for emotional semantics.

Expression of emotion

The global right hemisphere view of emotion has also

been supported by investigations of deficits in expression

of emotion. Overall facial expressivity of emotions has

been evaluated in RHD, LHD, and NH controls. Some authors

have reported that RHD patients were less spontaneously

expressive than LHD and NH controls (Blonder, et al., 1991;

Borod, Koff, Lorch, & Nicholas, 1985; Borod, Koff, Perlman-

Lorch, & Nicholas, 1988; Buck & Duffy, 1980). However,

Weddell, Miller, and Trevarth-n (1990) found LHD and RHD

patients who had tumors were equally impaired and less

expressive than NHD controls. When excisions occurred or

tumor and CVA patients were combined, RHD and LHD patients

did not significantly differ from controls (Kolb & Milner,

1981; Mammacuri, et al., 1988). Additionally, re nt

evidence exists from studies using a carefully delineated

facial scoring system which contradicts the findings that

RHD patients are less facially expressive. For example, no

24

differences in facial expressiveness has been found between

LHD and RHD patients when Ekman's facial action scoring

system (FACS) has been used (Mammacuri, et al., 1988;

Caltagirone, et al., 1989).

Other studies have examined the ability of RHD and LHD

patients to voluntarily pose specific facial expression.

Some investigators have found that RHD patients were more

impaired than LHD patients and NHD controls in their

voluntary expression of facial affect (Borod, Koff, Perlman-

Lorch, & Nicholas, 1986; Borod, & Koff, 1990; Kent, et al.,

1988; Richardson, Bowers, Eyeler, & Heilman, 1992). Other

investigators (Kolb and Taylor, 1990) found that RHD and LHD

patients are equally impaired relative to NHD controls,

whereas others found no differences in expressivity among

these three groups (Caltagirone et al., 1988; Heilman et

al., 1983; Weddell, et al., 1990).

Borod (submitted) reviewed the literature on facial

expressiveness in unilateral damaged patients. She

concluded that the patients in those studies finding RHD

patients to be more impaired than LHD and normal controls

differed from those in which differences were not found.

Specifically, she noted that the first group was more likely

to be older, male, with cerebrovascular pathology, and a

longer time since disease onset. The second group was more

likely to have tumor pathology. Additionally, subjective

ratings were used in the first group, while FACS and

concealed videotapes were used in the second group. One

problem with these differences is that stroke patients may

have more severe cognitive deficits than comparable tumor

patients (Anderson, Damasio & Tranel, 1990). Secondly,

acute pathology is associated with more pervasive deficits

(Borod, in press). Thirdly, FACS may be insufficiently

sensitive to facial expressive communication (Buck, 1990).

Asymmetries in facial expressiveness have also been

examined in normal adults. In a recent review of 23 studies

of spontaneous expression and 24 studies of posed

expression, Borod (in press) concluded that the left

hemiface is more intense and moves more than the right

hemiface. According to Borod, these results were stronger

for negative than positive emotions. There have been

fewer studies of prosodic emotion than facial expression of

emotion in patients with unilateral damage. Studies of

spontaneous prosodic expression have revealed deficits in

RHD patients compared to LHD patients and NHD controls (Ross

& Mesulam, 1979; Borod et al., 1985; Gorelick & Ross, 1987;

Ross, 1981). Similar results were found in investigations

of voluntary affective prosody, such that RHD patients

showed impairment relative to LHD and NHD controls (Borod et

al., 1990; Gorelick & Ross, 1987; Tucker, et al., 1977).

However, Cancelliere and Kertesz (1990) found no impairments

in either RHD and LHD patients relative to NHD controls.

Emotional arousal/activation

Few studies have examined affective psychophysiological

reactivity in brain-lesioned individuals. In the most

commonly used procedure, emotional slides have been used to

evoke affective responses while skin conductance response

(SCR) is measured. Findings indicate that normals and

patients with LHD have significantly higher SCRs to

emotional than neutral slides. In contrast, RHD patients do

not differentially respond to emotional and neutral slides

(Morrow, Vrtunski, Kim, & Boller, 1981; Zoccolatti, Scabini,

& Violani, 1982).

Similar results were obtained by Meadows and Kaplan

(1992) using slides depicting neutral and negative content

(i.e., mutilations). Relative to NHD controls, RHD patients

had smaller SCRs to both emotional and neutral slides, LHD

patients had high SCRs to both types of slides. Contrary

to the above findings, Schrandt, Tranel, and Damasio (1989)

found that left hemisphere lesions and many right hemisphere

lesions did not interfere with SCR during presentation of

emotional slides. In this study, patients with focal

lesions in left or right frontal, parietal, or temporal

lobes were examined. Only patients with right hemisphere

lesions involving the supramarginal gyrus displayed abnormal

SCRs.

In another study, Heilman, Schwartz, andWatson (1978)

investigated SCR while a mildly noxious electrical stimulus

was delivered to the forearm ipsilateral to the lesion in

RHD, LHD, and NH patients. The RHD group had smaller SCRs

than either the LHD or NH groups. Also, the LHD group had a

higher SCR than the normal group.

Cardiovascular activity has also been examined in

patients with LHD and RHD. Yokoyama, Jennings, Ackles,

Hood, and Boller (1987) examined RHD, LHD, and NC patients

using a reaction time task, while HR interbeat intervals

were obtained. The controls and LHD subjects displayed

anticipatory deceleration, followed by postresponse

acceleration. The HR responding of the RHD patients varied

little during anticipation and postresponse.

To sum, emotional slides evoke smaller SCR or less

arousal, in right hemisphere damaged patients compared to

NHD and LHD patients. Moreover, one study only found this

distinction in patients with right parietal lesions.

Additionally, in some studies, LHD patients responded with

accentuated SCRs, (i.e., greater arousal), in response to

emotional slides. Similar findings of decreased SCRs in RHD

patients and increased SCR in LHD patients have been

obtained in response to mildly noxious stimuli. Also,

patients with RHD have attenuated HR reactivity in response

to a reaction time task. Taken together, it appears that

RHD patients are hypoaroused and LHD patients may be

hyperaroused in response to emotional, painful, or

attention-demanding stimuli.

Bivalent Model of Emotion

In its simplest form, the bivalent model posits that

the right hemisphere is specialized for negative/avoidance

emotions, whereas the left hemisphere is specialized for

positive/approach emotions. According to the bivalent

model, the catastrophic reaction noted in left hemisphere

patients results from the predominance of the right

hemisphere's negative emotion. On the other hand, the

observations that right hemisphere damaged patients are

euphoric or cheerful can be explained by the overcontrol of

the left hemisphere's mediation of positive emotions

(Davidson & Fox, 1982; Kinsbourne & Bemporad, 1984; Reuter-

Lorenz & Davidson, 1981). This model was first based on

observation of emotional behavior during inactivation of the

left and right hemispheres with injection of sodium amytal

(Terzian, 1964).

Evaluation of emotion

Research investigating the hemispheric differences

during evaluation of nonverbal signals of emotion has

yielded conflicting results. Although tachistoscopic

studies in normals generally support the view that the right

hemisphere is superior for processing emotional faces (i.e.,

Suberi & McKeever, 1977), closer examination reveals some

support for the bivalent view of emotion. For example, the

finding of right hemisphere superiority was attenuated with

happy and angry facial expressions, which can be

conceptualized as approach emotions (Suberi & McKeever,

1977). Additionally, Reuter-Lorenz and Davidson (1981)

presented subjects with an emotional face and a neutral face

of the same individual simultaneously to each visual field.

Reaction times for identifying happy expressions were faster

during presentation to the right visual field (left

hemisphere) and faster for sad expressions when presented to

the left visual field (right hemisphere). However results

have not been consistently replicated (Duda & Brown, 1984;

McLaren & Bryson, 1987), and the vast majority of studies of

affect perception in normals or focal lesion patients failed

to demonstrate hemisphere-specific valence asymmetries.

Expression of emotion

Many studies of facial expressiveness have found that

the left side of the face is more expressive than the right.

These studies have been interpreted as reflecting a dominant

role of the right hemisphere in emotional expression

(Sackeim & Gur, 1978; Borod, Koff, & White, 1983; Campbell,

1978; Heller & Levy, 1981; Moreno, Borod, Welkowitz, &

Alpert, 1990). However, Schwartz, Ahern, and Brown (1979)

recorded bilateral corrugator and zygomatic EMG during a

mood induction task. They found that subjects expressed

positive emotions more intensely on the right side of the

face and negative emotions on the left side of the face.

However, the majority of research investigating emotional

expressivity in normals and patients with focal lesions

30

supports the global rather than the bivalent model (Blonder,

et al., 1991; Borod et al., 1985, 1988; Buck & Duffy, 1980).

Emotional arousal/activation

Hemispheric activation during emotional responding in

normal subjects have been investigated using measures such

as electroencephalography (EEG) and lateral eye movements

(LEM). Using EEG, it has been found that in the frontal

zones, positive emotions produced more left than right

hemisphere EEG activation, while negative emotions produced

more right than left EEG activation (Ahern & Schwartz, 1985;

Tucker, Stenslie, Roth, & Shearer, 1981; Davidson et al.,

1979; Davidson, et al., 1990). In addition, Ahern and

Schwartz (1985) found that the right parietal zone was

related to emotional intensity, whereas Bennett, Davidson

and Saron (1980) as well as Davidson and colleagues (1990)

found no differences in parietal activation related to

emotion.

Lateral eye movements (LEM) have also been used as a

measure of hemispheric activation. LEM towards the right

have been interpreted as reflecting left hemisphere

activation, while LEM to the left is suggestive of right

hemisphere activation. Initial findings revealed more LEMs

to the left during emotional experience (Davidson &

Schwartz, 1976; Schwartz, Davidson, & Maer, 1975; Tucker,

Roth, Arneson, & Buckingham, 1977). Ahern and Schwartz

(1979) investigated lateral eye movement in response to

reflective questions in normal subjects. They found that

positive emotional questions evoked more LEMs to the left.

They interpreted this as left hemisphere specialization for

positive emotions and right hemisphere specialization for

negative emotions. However, the lateral eye movement

methodology has been criticized (Erlichman & Weinberger,

1978).

Research on mood

Observation of mood after hemispheric damage has also

been viewed as supporting the bivalent model. Sackheim et

al. (1982) reported that pathological laughing was more

likely to be associated with RHD and pathological crying was

associated with LHD. Additionally, they found that patients

with right hemispherectomies were judged to be euphoric in

mood, while patients with left hemispherectomies were not.

Also, they examined published case reports of gelastic

epileptics, typified by laughing outbursts during ictal

experience, with either left or right lateralized ictal

foci. They found that ictal foci in gelastic epileptics was

predominately left-sided. Based on previous literature, the

authors suggested that the laughing outburst which occurred

during ictal experience were caused by hyperactivity in the

focal area. These authors concluded that both disinhibition

and excitation cause different manifestations in mood in the

right and left hemispheres.

Robinson and his colleagues have investigated

depressive symptoms following stroke in both right and left

hemisphere patients. In two studies, Robinson and Price

(1982) and Robinson et al. (1984) found that patients with

left hemisphere strokes were more depressed than patients

with right hemisphere strokes. Starkstein, Robinson, and

Price (1987) also noted that right hemisphere patients were

indifferent and sometimes euphoric immediately following

stroke. Additionally, Robinson and Szetela (1981) reported

that patients with traumatic brain injury, while equally as

impaired cognitively and physically, were not as depressed

as stroke patients. Consequently, frequency and severity of

depression is not solely related to amount of physical and

cognitive impairment.

Differences in mood depending on caudality (anterior

versus posterior location) of the lesions were also observed

(Robinson et al., 1984). The left anterior group showed

significantly more overall depression than the left

posterior group, whereas the right posterior group were more

depressed than right anterior group. Similarly, Starkstein

et al. (1987) reported that when depression was present in

RHD patients, it was associated with parietal lesions.

Additionally, depression was found to be correlated with

closeness of the lesion to the frontal pole (Robinson &

Szetela, 1981; Starkstein, Robinson, and Price, 1987).

In a subsequent study, Sinyor, et al. (1986) assessed

both cognitive and vegetative signs of depression using a

variety of verbal report measures in unilateral stroke

patients. Contrary to the above findings, no overall

differences in depression were found between groups.

However, consistent with the above findings, severity of

depression in LHD patients was positively related to

proximity of the lesion to the frontal pole. In addition, a

curvilinear relationship was found for RHD patients such

that both anterior and posterior lesions were associated

with depression. Moreover, House et al. (1990) reported

that RHD patients may be depressed more than is believed,

but due to their deficits in emotional communication, their

depression goes undetected.

Taken together, the results are equivocal. There is

evidence in support of differential moods in left and right

hemisphere damaged patients. Some investigators suggested

that RHD patients express enhanced cheerfulness (e.g.,

Terzian, 1964), and LHD patients express or report

experiencing more depression than RHD patients (e.g.,

Robinson et al., 1984). However, other investigators found

no differences in depressed mood between LHD and RHD

patients. Additionally, some studies revealed that during

negative emotion, greater EEG activation was associated with

anterior right activation. In contrast, during positive

emotion, greater EEG activation was associated with anterior

left activation. However, EEG activation of right frontal

and right parietal regions was associated with emotion

intensity. Also, inferring hemispheric activation using

LEM, findings supported greater right hemisphere activation

during negative emotion experience and left hemisphere

activation during positive emotion experience, but LEM

methodology has also been criticized.

Specific bivalent models

In general, the bivalent model posits that the left

hemisphere is specialized for positive/approach emotions and

the right hemisphere is specialized for negative/avoidance

emotions. However, there are many variations of the general

bivalent model. Kinsbourne and Bemporad (1984) suggested

that the left fronto-temporal cortex exerts action control,

defined as manipulating external stimuli. They argued that

left posterior parietal cortex sends exteroceptive input to

the left fronto-temporal cortex. The right fronto-temporal

cortex, on the other hand, controls emotional, internal

arousal, while the right posterior cortex relays

interoceptive information to the emotional control system.

Consequently, in patients with right focal lesions,

meaningfulness of environmental stimuli is deficient. Thus,

RHD patients experience inappropriate emotionality.

Additionally, Kinsbourne and Bemporad explained that the RH

is specialized for monitoring both positive and negative

emotional valence, but positive states enhance motivation

and readiness to act which are left hemisphere attributes.

Specifically, passivity and involvement in perceptual

judgement relates to RH activation, whereas overt responses

or covert response planning is associated with left

hemisphere activation.

Davidson and his colleagues (Fox and Davidson,1984;

Davidson, 1985; Davidson et al., 1990) proposed a similar

theory. They purported that the behavioral dimension of

approach-withdrawal is the organizing dimension for

hemispheric specialization in that the right hemisphere is

specialized for withdrawal emotions such as disgust, whereas

the left hemisphere is specialized for approach emotions

such as interest. In addition, Davidson (1985) postulated

there are reciprocal relations between the frontal and

parietal lobes. Specifically, left frontal activation is

balanced by right parietal activation and vice versa. For

example, he stated that spatial cognition (right parietal)

and positive affect (left frontal) are more likely to occur

concurrently than verbal cognition (left parietal) and

positive affect.

Heller (1990) posited a similar view. She asserted

that the right hemisphere may be specialized for

interpretation of emotion, but not specialized for the

regulation of mood. Heller also emphasized the importance

of distinguishing between the functions of the anterior and

posterior regions of the brain, citing evidence that the

right temporal parietal regions are involved in

interpretation of emotional information and evidence that

implicates the frontal regions of both hemispheres in the

experience of mood. Heller (1990) stated that the right

parietal cortex mediates both cortical and autonomic

arousal, while bilateral frontal regions mediate valence.

She purported that experience of emotion is associated with

patterns of activation in frontal and parietal brain

regions.

Summary

As reviewed in the preceding sections, most evidence

supportive of the bivalent model has been derived from two

lines of research. These include: (a) findings of different

mood reactions following right versus left hemisphere

lesions, particularly those involving the anterior regions;

and (b) findings in normals of hemispheric EEG activation

asymmetries during induction of positive versus negative

mood. In contrast, data from neuropsychological studies of

affect perception are more in line with the view that the

right hemisphere is critically involved in appraising

nonverbal emotional signals, regardless of their valence.

The discrepancy between such studies corresponds to the

distinction raised by Heller (1990) between interpretation

of emotion (viewed to be right hemisphere dependent) versus

the regulation of mood (which is not viewed to be right

hemisphere specific).

Observations that RHD patients are autonomically

hypoaroused in response to affective scenes (relative to NHD

and LHD patients) have been interpreted as support for a

dominant role of the right hemisphere in emotional arousal.

However, this interpretation is not without question given

that such studies have generally measured autonomic

responsivity only in response to neutral and unpleasant

scenes (Meadows & Kaplan, 1992; Zoccolatti et al., 1982) or

situations (Heilman et al., 1978). Pleasant scenes or

stimulus materials have not been used in such studies and it

remains unknown whether stroke of the right hemisphere

equally attenuate autonomic reactivity to pleasant scenes.

In and of itself, the current existing data that RHD stroke

patients are hypoaroused to negative-affective scenes are

equally consistent with the bivalent as well as the global

right hemisphere model. Of relevance, Morris et al. (1991)

recently reported valence-specific hypoarousal in a patient

following a right temporal lobectomy. Skin conductance

responses were obtained to unpleasant (mutilations),

pleasant (attractive nudes), and neutral breadbasketss)

slides. This patient showed abnormally reduced SCR to

unpleasant but normal SCR to pleasant and neutral slides, a

pattern of findings that is consistent with a bivalent

model. Had only unpleasant scenes been used in this study

one would not be able to logically distinguish between the

bivalent and global right hemisphere model. For this

reason, it is crucial to include both pleasant and

unpleasant scenes or situations when studying

psychophysiological responses in neuropsychological

investigations of emotion. Such was employed in this study.

Before discussing the proposed study more fully, a

brief overview of relevant literature on emotional

psychophysiology will be presented. This is being done

since the current study will include several

psychophysiological indices (i.e., skin conductance, heart

rate, facial electromyography) for assessing emotional

responsivity in patients with right or left hemisphere

lesions.

Emotional Psychophysioloqy

Autonomic Responding

At the psychophysiological level, the relationship

between autonomic activity and emotion has been recognized

for centuries. Recent technological advances have made the

prospect of online physiological measurement more feasible.

Theorists have attempted to understand the factors which

influence skin conductance and heart rate. Sokolov (1963)

described two types of responses which occur during

conditioning: orienting and defensive reactions. He

purported that the purpose of the orienting response (OR) is

to increase sensitivity to incoming stimuli and that it

includes both a transient increase in skin conductance. The

defensive response (DR), on the other hand, is evoked in

39

response to high intensity or aversive stimuli and helps the

organism to limit activity with the stimulus. This response

includes increases in sympathetic activity such as cephalic

vasoconstriction and increase in skin conductance.

Lacey and Lacey (1970) extended Sokolov's views of

autonomic responding. They suggested that heart rate

acceleration (tachycardia) during acute affective states is

not a index of arousal per se, but reflects instead the

organism's attempt to limit or terminate bodily turmoil

produced by some stimulus. By contrast, heart rate

deceleration (bradycardia) is induced with intention to

respond to a task, attention to stimuli, and during

vicariously experienced stress. Thus, Lacey and Lacey

argued that the cardiovascular system is not a nonspecific

index of arousal, but a highly specialized response

mechanism which is integrated with affect and cognition and

which also reveals individual differences in the way people

deal with the environment.

Graham and Clifton (1966) pointed out that Sokolov

(1963) and the Laceys (1958) agreed on the existence of an

orienting and defensive response. However, Graham and

Clifton indicated that they did not agree on the

relationship between orienting and defensive responses and

heart rate. Sokolov inferred that heart rate (HR)

acceleration was related to increased sensitivity of

incoming stimuli, whereas HR deceleration was related to

decreased sensitivity of incoming stimuli. The Laceys

hypothesized the reverse pattern. In their thorough review

of the literature, Graham and Clifton concluded that, in

fact, the Laceys hypotheses have been supported in that HR

deceleration is associated with orienting and HR

acceleration is associated with defensive responding.

A large body of research exists in which the autonomic

correlates of affective states have been investigated.

Throughout the second half of this century, researchers have

systematically explored the relationship between emotion and

psychophysiological measures including skin conductance and

heart rate. Early studies of systematic desensitization in

phobic patients revealed that as the subjects imagined more

fearful images, HR and skin conductance responses (SCR)

increased (Lang, Melamad, & Hart, 1970).

In the late 1960s and early 1970s, a series of studies

by Hare and colleagues indicated that slides of mutilated

bodies evoked HR deceleration, an orienting response (OR).

These results were initially confusing because it had been

hypothesized that the slides would evoke fear and HR

acceleration, a defense response (DR). Upon reanalyzing his

data (Hare, 1972), it was found that some subjects had

consistently reacted with HR acceleration, some with marked

deceleration, and some with moderate deceleration.

Subsequently, researchers explored the differing

reactions of phobics and nonphobics in response to affective

slides. The findings indicated that presentation of a

feared object resulted in initial HR acceleration, e.g.,

(DR), while presentation of a nonfeared object results in HR

deceleration, e.g., (OR) (Hare, 1973; Klorman, Weissberg, &

Wiesenfeld, 1977; Klorman, Wiesenfeld & Austin, 1975).

Additionally, SCR was elevated with the presentation of

fearful stimuli (e.g., Klorman, Weissberg, & Wiesenfeld,

1977) and, in some studies, the amount of elevation was

higher for phobics (Klorman, Wiesenfeld & Austin, 1975).

Imagery has also been used to evoke emotional states.

It is important to note that during imagery, autonomic

responsivity (i.e., HR and SCR) is influenced not only by

the affective state, but also by other factors such as

imagery instructions and the subjects' ability to image

(Lang, Kozak, Miller, & Levin, 1980; Miller, Levin, Kozak,

Cook, McLean, & Lang, 1987. Vrana, Cuthbert, and Lang

(1986) found that normal subjects verbally reported

experiencing more arousal, more unpleasantness, and less

control during fear imagery than during neutral imagery.

Fear images also evoked HR acceleration which lasted over a

10 second period. In contrast, neutral images produced

acceleration followed by deceleration. Thus, HR and

subjective report distinguished fearful from neutral

imagery.

Taken together, the results of these studies are

consistent with the views of Graham and Clifton (1966) and

Lacey and Lacey (1970). Heart rate typically increases in

response to feared stimuli when presented visually or

imagined. On the other hand, HR deceleration follows the

visual presentation of a novel or interesting stimulus,

whereas imaging of a novel or interesting stimuli produces

HR acceleration followed by deceleration.

Facial Electromyography (EMG)

Before describing the facial electromyography research,

the neuroanatomical pathways involved in facial muscle

movements will be briefly reviewed. Motor neurons send

information from the brain to innervate muscle and can be

distinguished from sensory neurons which bring information

to the brain. There are two types of motor neurons: upper

motor neurons (.T) and lower motor neurons (LMN) Upper

motor neurons carry impulses from motor centers in the brain

to the brain stem and spinal cord. Lower motor neurons

carry information from brain stem and spinal cord to

muscles. At the UMN level, fibers from either the

contralateral or both hemispheres supply impulses to the LMN

nucleus, the motor nucleus of the facial nerve, which

innervates muscles of facial expression. The voluntary and

involuntary motor pathways mediating facial expression are

distinct from one another. Voluntary movement is mediated

by the corticobulbar tract, originating in the precentral

gyrus of the motor cortex of the frontal lobe. The

involuntary pathway includes the basal ganglia, red nucleus,

and midbrain reticular formation. (Rinn, 1984). Although

the pathways of voluntary and involuntary emotions are

different, the measurement of facial expressions are the

same regardless of the volitional quality of the expression.

Detailed facial coding systems, such as Ekman's FACS

(Ekman & Friesen, 1978) and Izard's MAX (1978) have been

used to measure minute muscle movements of the face.

Because these rating systems are quite time intensive and

because spontaneous facial muscle activity is often brief

and too small to be observed overtly, facial

electromyography (EMG) has sometimes been used to measure

subtle changes in muscle movements. The most common facial

muscle regions measured using EMG are the corrugator

supercilli (brow) and zygomatic major (cheek) muscles

regions. Various methods have been used to induce emotional

states while EMG of the corrugator and zygomatic muscles

have been measured. These emotion eliciting procedures have

included imagery, viewing affective slides, self-referential

statements, and self-disclosing interview. Consequently,

the facial expressions that accompany these emotion

induction procedures involve involuntary/spontaneous facial

movements. The UMN innervation of the corrugator muscle is

bilateral, whereas the UMN innervation of the zygomatic is

contralateral (Rinn, 1984). Thus, muscle activity in the

left and right corrugator regions cannot be activated

independently, but muscle activity of the left and right

zygomatic regions can be stimulated separately.

During affective imagery, positive emotional states

have been associated with decreased corrugator and increased

zygomatic activity. Conversely, negative emotional states

have been associated with increased corrugator activity and

decreased zygomatic activity (Schwartz et al., 1976a,

1976b). Also, when verbal report of emotions has been

obtained, corrugator activity positively correlates with

unpleasant emotions and negatively correlates with pleasant

emotions. The opposite pattern has been found for zygomatic

activity (Brown & Schwartz, 1980; McCanne & Anderson, 1987;

Slomine and Greene, 1993). Similar results have been

reported from other investigators using self-referent

statements designed to induce either elation or depression

(Sirota, Schwartz, & Kristeller, 1987), and affective slides

(Cacioppo, Petty, Lasch, and Kim, 1986). Additionally, an

interview technique was employed to elicit and investigate

naturally occurring emotional states (Cacioppo, Martzke,

Petty and Tassinary, 1988). Replicating previous findings,

elevations in corrugator EMG were related to lower positive

emotion ratings and higher negative emotional ratings.

In sum, the above studies attest to the importance of

the covert activity of the corrugator supercilli and

zygomatic major muscles as indexes of emotion.

Specifically, EMG activity of the corrugator supercilli has

been consistently found to increase during exposure to

stimuli rated as unpleasant or during the reported

experience of unpleasant affect. Conversely, activity of

the zygomatic major has been found to increase during the

report of positive emotional states.

Facial and Autonomic Studies

Few studies have included measures of both facial and

autonomic responding. In one study of affective slides

viewing, Greenwald, Cook, and Lang (1989) examined emotional

ratings, HR, SCR, zygomatic and corrugator EMG. Zygomatic

activity was positively related to pleasure ratings and

corrugator activity was negatively related. Zygomatic EMG,

however, also increased during unpleasant slides viewing.

Neither muscle site was related to arousal ratings. Phasic

HR acceleration was positively related to valence ratings,

but not arousal. This relationship was weaker than the

valence/EMG relationship. Skin conductance responses were

significantly related to increased arousal ratings, but not

valence ratings. Quite similar results were found when

autonomic and facial responding were measured during imagery

(York, 1991; Bradley, Lang, & Cuthbert (1991) in that HR

acceleration and SCR were larger for pleasant and unpleasant

compared to neutral imagery, and corrugator EMG was higher

for the unpleasant compared to pleasant and neutral imagery.

Ekman and colleagues have found that giving subjects

either muscle-by-muscle instructions to contract voluntary

sets of facial expression and asking subjects to relive a

past emotional experience produced similar autonomic

changes, i.e., increases in HR and SCR (Ekman, Levenson, &

Friesen, 1983; Levenson, Ekman, and Friesen, 1990). These

authors concluded that there are biologically innate affect

programs which, when activated, provide instructions to

multiple response systems including skeletal muscles,

facial muscles, and the autonomic nervous system.

Taken together, the above research suggests that

zygomatic EMG increases with reported pleasantness, and

somewhat with extreme unpleasantness. Corrugator EMG

increases with reported unpleasantness. Skin conductance

responses are positively related to reported experience of

arousal, which can be induced through pleasant or unpleasant

emotional states. Heart rate, however, is variable and

depends on many factors such as reported affect, type of

evoking stimuli, and individual differences in responding.

However, during the presentation of emotional slides, HR

acceleration is positively related to valence, but

acceleration may be associated with aversive rather than

pleasant stimuli when phobics are presented with their fear

object. During imagery, HR typically accelerates during

both pleasant and unpleasant scenes. Additionally,

voluntary facial expressions produce changes in the

autonomic nervous system consistent with other tasks used to

induce emotional experience.

Anticipation of Affective Stimuli

Anticipation of affective stimuli has also been used to

elicit emotion. Lang, Ohman, and Simons (1978) described

the triphasic response of cardiac activity during a 4-8

second anticipation period. They reported that the onset of

the preparatory period is characterized by a brief

deceleration (D1). The initial deceleration is followed by

an acceleratory peak (Al). Lastly, a deceleration occurs

which lasts until the end of the preparatory interval (D2).

D1 is observed when subjects are presented with single pure

tones which are not followed by other stimuli and is thought

to be an index of orientation. The acceleratory phase is

seen in response to an abrupt stimulus or single stimulus

with an uncomfortable intensity level. Al has been

interpreted as an index of a defensive reflex. It has also

been evoked in the absence of noxious stimuli and during

problem solving or mentation.

According to Lang et. al (1978), most investigators

interpret the second deceleration, D2, as an index of

anticipation of an overt response. D2, however, has been

conditioned in classical conditioning paradigm even though

no motor response is required. Consequently, D2 has also

been viewed as an index of an attentive set. Similar HR

patterns have been found by Simons, Ohman, and Lang (1979)

in response to anticipation of slides (Simons, Ohman, &

Lang, 1979; Klorman & Ryan, 1980).

There is a large body of literature based on

anticipation of aversive stimuli. In one study, cluster

analysis was used to identify different patterns of HR

responses during anticipation of aversive noise (Hodes,

Cook, & Lang, 1985). Results indicated that there were

three types of responders; accelerators, decelerators, and

moderate decelerators similar to the groups obtained by Hare

(1972). The authors concluded that both the accelerators

and decelerators developed the expectancy that the CS+ would

precede the presentation of UCS. Accelerators, however,

associated fear with the CS+, while the decelerators did

not. The authors suggested that because the classical

aversive conditioning paradigm specifies no overt response

set, the subjects spontaneously assumed a response

disposition. Specifically, some responded with an

anticipatory, attentive set demonstrated by decelerators,

whereas others displayed an implicit avoidance characterized

by a defensive response. Moderate decelerazors showed

discordance between verbal and physiological behavior.

Thus, these subjects maintained an attentive set, but

evaluated the stimuli as aversive instead of interesting.

The authors concluded that "It is conceivable that the

tactile assault of shock is necessary to consistently elicit

DR's to such potentially skin mordant stimuli as snakes and

spiders," (p.555).

Psychophysiological responses of HR and skin

conductance have been measured during anticipation of

electric shock. Deane (1961) found that during anticipation

of shock, HR accelerated over the baseline level.

Additionally, in the groups who expected to receive shock

when a 'target' number was presented, there was HR

deceleration immediately preceding that number, even though,

in one of these groups no shock had ever been received.

These finding have been replicated (Elliot, 1966; Deane,

1969; Hodges & Spielberger, 1966). Threat of electric shock

has also been found to produce increases in SCR (Bowers,

1971a, 1971b). Positron emission tomography (PET)

measurements of regional blood flow have also been obtained

during anticipation of electric shock (Reiman, Fusselman,

Fox, & Raichle, 1989). Reiman and colleagues found that

during anticipatory anxiety, there was significant blood

flow increases to both temporal poles.

The investigation of the psychophysiology of pleasant

and appetitive anticipation has received minimal attention

in the experimental human literature. Consequently,

psychophysiological responding during pleasant anticipation

must be inferred from other studies. Based on the results

of the above literature, it is likely that anticipation of

pleasant stimuli would evoke physiological changes similar

to those found during presentation of pleasant stimuli

(i.e., increased zygomatic EMG and SCR). Also, based on the

above studies of anticipation during nonaversive

anticipation (Simons, et al., 1979; Klorman & Ryan, 1980),

HR is primarily deceleratory.

Summary

Psychophysiological measures of heart rate (HR), skin

conductance responding (SCR), corrugator electromyography

(CEMG), and zygomatic electromyography (ZEMG) have all been

used as indices of emotional psychophysiology. Alone, each

of these measures has been associated with various

psychological phenomenon. For example, SCR has been

associated with mental effort, attentive movements or

attitudes, painful stimuli, variations in respiratory rate,

along with emotional arousal and various other psychological

phenomenon Cacioppo & Tassinary, 1990). Increased heart

rate has also been associated with various psychological

phenomenon including startle, mental effort, and defensive

responding Cacioppo & Tassinary, 1990). In addition,

corrugatcr electromycgraphy has been associated with

concentration as well as unpleasant emotional experience

,Cacioo, e, t & Morris, 1985 .

Because chances in heart rate, skin conductance, and

facial EMG have all been fiund to be associated with

psychological phenomenon other than emotional experience,

changes in one cf these variables is not necessarily

indicative of emotional experience. However, examination of

multiple variables over time has revealed specific

51

physiological response patterning which results in a one-to-

one relationship with experience of emotional states. Thus,

it is necessary to investigate patterns of physiological

behavior over time in order to infer the presence of a

psychological phenomenon based on physiological responding

(Cacioppo & Tassinary, 1990).

Critical Issues

As reviewed earlier, there are two cc:rsing views of

how the cerebral hemispheres differ in their contributions

to emotional processing. However, the precise role played

by each hemisphere remains unclear. Some investigators have

proposed that the right hemisphere is globally involved in

all aspects of emotional processing including evaluation,

expression, activation, and experience of emotion (Heilman

et al., 1985). Others researchers have suggested that each

hemisphere is specialized fcr a different type of emotion

(Fox & Davidson, 1984; Kinsbourne & Bemporad, 1984; Tucker,

1981; Heller, 1990). The mcso popular version of the

bivalent view is that the lef- hemisphere is dominant for

positive/approach emotions, while the right hemisphere is

dominant for negative/avoidance emotions.

Alona with differences in laterality of emotional

processing, investiga-trs have speculated about differences

in emotional processing based on caudality, i.e., anterior

versus posterior regions cf the brain. For example, studies

of interpretation of emotional information implicate the

right temporal and parietal regions (e.g., Bowers et al.,

1987), whereas studies of emotional mood have implicated the

frontal lobes (e.g., Davidson, 1984).

Heller (1990) has interpreted the literature in terms

of type of emotional processing, such that "cold" or

nonexperienced emotional processing is modulated by the

right posterior region. In addition, she posited that

"warm" or experienced positive emotion is processed

predominantly by the left hemisphere, whereas "warm or

experienced negative emotion is processed predominantly by

the right hemisphere. According to Helier, the majority of

evidence in support of the right hemisphere model of emotion

comes from studies which have investigated cognitive

processing of information in brain damaged and normal

subjects, whereas most evidence in support of the bivalent

models cf emotion has been derived from lateralization of

mood states. Heller suggested that there is no reason to

assume that because a hemisphere is associated with a

particular mood state, that it must be specific for

cognitive representations of that emotion.

In order to distinguish among the ability of the global

and bivalent models nc explain emotional experience, it is

necessary to evoke emccion with both positive/approach and

negative/withdrawal emotions. Because RHD patients have

difficulty interpreting emotional stimuli, including faces

and prosody (e.g., Bowers et al., 1987), it is difficult to

53
evoke emotional states in the laboratory. Thus, using an in

vivo situation in which nonverbal emotional stimuli do not

have to be interpreted would be useful in evoking emotion in

RHD patients. Ideally, it is crucial for positive and

negative emotions to be equally arousing. Unfortunately, it

is difficult to equate in vivo positive and negative

emotional experiences in emotional arousal because highly

arousing negative emotional experience is much easier to

experimentally induce than highly arousing positive

emotional experience.

It is important to define emotional experience and how

it can be measured. As mentioned above, emotional

experience is defined as a phenomenon which can be

indirectly measured using physiological measures (e.g., HR

and SCR), overt behavior (e.g., facial expression, in this

case measured using CEMG and ZEMG), and verbal report (e.g.,

paper and pencil assessment measures In normal subjects

these three response systems have usually been found to be

concordant; however, discordant responses have been revealed

in pathological populations "Patrick, Bradley, & Lang,

1991). These discordant results may imply that the three

response systems are mediated by different subsystems. In

brain damaged patients discordance is often observed. For

example, patients with pseudobulhar laughter display cvert

behaviors cf emotion, but verbally deny feelings associated

with emotion ,Heilman, Bowers, & Valestein, 1993) These

54
results imply that there is a defect in the mediation output

systems, such that behaviorally the patient responds, but

without the corresponding subjective experience of emotion.

Due to the inability in directly measuring subjective

experience, the ability to interpret discordance in response

systems is weakened. To illustrate, two groups, A and B,

are investigated during emotion-eliciting experiences. Both

A and B verbally report experiencing emotion. However,

group A does not exhibit psychophysiological measures

indicative of emotion. Are the subjective emotional

experiences of group A and B different? There are two

possible interpretations: (1) they are experiencing

qualitatively different emotional experiences, such that

group A's experience of emotion is more "cognitive" than

group B's experience, or (2) they are experiencing the same

emotional experiences, but group A has a problem with the

feedforward system of emotional psychophysiological

responding. Because interpretation includes inferences

about subjective experiences, neither interpretation can be

proven correct or rejected as invalid. It is unclear, at

this time, how patients with unilateral damage experience

emotion based on the interaction of these three response

systems. Specifically, it is unknown whether unilateral

lesions would produce concordance or discordance of

emotional experience.

It is important to consider the constraints that are

placed on evaluating emotional experience in patients with

focal lesions. For example, left hemisphere damaged

patients often have difficulty with language, which may

affect their verbal report data. To minimize this problem

in the present study, severely aphasic patients would not be

used and only verbal report measures with simple language

were used. Also, right hemisphere damaged patients often

have difficulties with visual attention, neglect, and

vigilance. Consequently, adequate attention to the task at

hand must be insured among RHD patients.

To study emotional experience, it is important to

measure all three response systems; verbal report, overt

behaviors, and physiological indices. One way to better

understand the neuropsychology of emotional experience is to

use paradigms which are highly sensitive to emotional

responding. The present study focused on an anticipation

paradigm (Reiman et al., 1989) designed to investigate

verbal report, heart rate, skin conductance, and facial

responses associated with emotion. In order to examine the

psychophysiology of emotional experience, an "in vivo"

situation was used. Using anticipation of "in vivo"

aversive and pleasant stimuli, it was easier for patients to

interpret the emotional meaning of the situations because

they did not have to analyze the affective quality of

various perceptual stimuli.

CHAPTER 2
STATEMENT OF THE PROBLEM

The purpose of the present study was to broadly examine

emotional responsivity of RHD and LHD patients in affect-

evoking situations and determine whether the pattern of

responses obtained from these patients was more in line with

predictions of a global right hemisphere model versus a

bivalent hemisphere emotion model. To examine this verbal

report, autonomic measures of arousal (SCR, HR), and indices

of facial muscle movement (EMG) were be collected in

situations that are known to elicit negative (anticipation

of shock) and positive responses (anticipation of reward) in

normals.

To date, few neuropsychological studies of emotion with

focal lesion patients have concurrently investigated more

than one component of emotional responsivity. That is,

either autonomic indices have been obtained (Heilman,

Schwartz, & Watson, 1978) or verbal report of mood states

have been obtained (Robinson & Price, 1982). No study to

date has used facial EMG to examine emotional responsivity

in focal lesion patients. Facial EMG may potentially be a

useful tool in that it has been shown to be sensitive to

changes in the reported experience of valence in normal

individuals (Greenwald et al., 1989).

Further, those patient studies that have examined

psychophysiological indices of arousal in response to

emotional stimuli have typically used perceptual stimuli

(i.e., affective scenes) which must be accurately

"interpreted" in order to induce emotion. Patients with RHD

are known to have an array of visuoperceptual and

hemispatial attentional scanning difficulties which can

potentially interfere with their interpretation of such

stimuli. Consequently, findings that RHD patients are

autonomically hypoaroused in response to emotional scenes

may, in part, be secondary to difficulties in interpreting

these stimuli.

To avoid such confounding, the present study used "in

vivo" situations to elicit negative and positive emotions

among focal lesion patients. An anticipatory anxiety

paradigm adopted from Reiman et al. (1989) was used to

induce negative emotion (i.e., anxiety). In this paradigm,

subjects are told that they would sometimes receive a mild

shock. Findings with normals reveal changes in autonomic

arousal during the period that the subject is awaiting shock

in conjunction with self reports of increased levels of

anxiety (as measured by the State-Trait Anxiety Inventory).

An anticipatory reward paradigm was used to induce positive

emotion. Here, subjects were told that they would sometimes

receive monetary reward (i.e., dollar bills or lottery

tickets).

The specific objectives of this study are to determine:

(a) whether patients with RHD or LHD become autonomically

aroused in these in vivo emotional situations (as indexed by

HR and SCR changes); (b) whether they display contraction of

facial muscles (as measured by EMG indices) that correspond

to the positive-negative nature of the emotional situation;

and (c) whether they explicitly report subjective changes in

their emotional experiences (as measured by their responses

to questionnaires).

According to the global right hemisphere emotion model,

the RHD patients should display attenuated responsivity

across all three response domains (arousal, facial, verbal

report) in both the negative and positive emotion-eliciting

situations. In other words, relative to the LHD group, the

RHD patients should be less autonomically aroused, show

minimal facial muscle contractions, and report less intense

changes in their subjective experience of emotion.

Diminished responding by RHD patients would be observed in

both the anticipatory anxiety paradigm, as well as the

anticipatory reward task.

According to the bivalent hemisphere emotion model, the

responses of the RHD and LHD patients would vary as a

function of the positive-negative nature of the induced

emotional situation. Specifically, the RHD group would show

diminished autonomic responsivity and less intense reports

of emotional experience in the anticipatory anxiety task

relative to the anticipatory reward task, whereas the LHD

group would show the opposite pattern.

Overview of Experimental Design

Patients with RHD, LHD, and NHD participated in two

experiments. Both experiments consisted of two parts, an

anticipatory anxiety task and an anticipatory reward task.

In the first experiment, a two-stimulus paradigm (see Vrana,

Cuthbert, & Lang, 1989) was used in both the anticipatory

anxiety and anticipatory reward tasks. Specifically, one

warning tone signaled that the subject would receive shock

stimulation during the subsequent six seconds, whereas the

other tone signaled that shock would not occur. Prior to

the beginning of the task, subjects learned which tone would

be associated with shock and which with no shock. An

analogous two-stimulus paradigm was used in the anticipatory

reward task. Psychophysiological measures of arousal (HR,

SCR) and facial E'- corrugatorr and zygomatic) were obtained

during the six second anticipatory interval.

In the second experiment, a slightly different paradigm

was used to examine anticipatory anxiety and reward in RHD

and LHD patients. Specifically, there was a 5 minute

interval (versus 6 seconds in Experiment 1) during which the

subject awaited shock (or reward). Five-minute control

trials were also be given in which the subject is told that

no shock (or reward) would be presented. During these 5-

minute anticipatory intervals, subjects were administered

brief mood questionnaires (i.e., Positive and Negative

Affect Schedule and Self-Assessment Manikin).

The use of the 5-minute paradigm in Experiment 2 is

more suitable for obtaining self-report information, whereas

the use of 6-second two-stimulus paradigm in Experiment 1 is

more suitable for obtaining reliable brief

psychophysiological indices of emotion.

Hypotheses and Predictions

Overall Hypotheses

According to the global right hemisphere model, emotion

is modulated predominantly by the right hemisphere.

Consequently, the global model hypothesizes that patients

with RHD will display attenuated responsivity, relative to

the LHD group, across all three response domains (arousal,

facial, and verbal report) in both negative and positive

emotion-evoking situations.

In contrast, the bivalent model posits that

positive/approach emotions are mediated by the left

hemisphere and negative/avoidance emotions are mediated by

the right hemisphere. According to the bivalent model, the

responses of the RHD and LHD patients would vary as a

function of valence (positive-negative nature) of the

induced emotion. Specifically, the RHD group would show

diminished responses in all three response domains (arousal,

facial, and verbal report) during the anticipatory anxiety

(negative emotion) situation relative to their responses

during anticipatory reward (positive emotion). The LHD

group would show the opposite pattern.

Specific Predictions for Experiment 1: Psychophysiological
Arousal and Facial EMG during Anticipatory Anxiety and
Anticipatory Reward in Patients with RHD and LHD

Normal control group (NHD)

In line with previous research, it is anticipated that

the normal control group (NHD) will experience unpleasant

emotion (anticipatory anxiety) during the shock anticipation

condition and more pleasant emotion (anticipatory reward)

during the prize anticipation. Specific predictions

regarding psychophysiological responsivity (HR, SCR) and

facial EMG are derived from empirical research with emotion-

inducing stimuli. A replication of previous findings is

expected such that:

1. Compared to baseline HR, a HR triphasic response (Dl,

Al, D2) will be observed during shock anticipation and

prize anticipation. The Al, acceleratory peak, is

expected to be greater during shock than during prize

anticipation. In some subjects, however, deceleration

only may be observed during prize anticipation.

Relative to the experimental trials, attenuated HR

change will occur during control trials.

2. Compared to baseline SCR, SCR will be greater during

shock and prize anticipation compared to no shock/no

62
reward control trials. Additionally, SCR will decrease

over trials.

3. Compared to baseline corrugator EMG, corrugator EMG

(CEMG) will be elevated during shock anticipation and

will remain relatively unchanged during prize and

control trials.

4. Compared to baseline zygomatic EMG, zygomatic EMG (ZEMG)

will increase during prize anticipation. Additionally,

a smaller increase may be revealed during shock

anticipation. Also, ZEMG will not change from baseline

during control trials.

Focal Lesion Patients (RHD and LHD)

Predictions for the RHD and LHD patients differ

depending on the global right hemisphere emotion model

versus the bivalent model. Specific predictions for the

right hemisphere emotion model will be first presented and

then followed by those from the bivalent model.

A) Global Right Hemisphere Emotion Model: According to this

view, patients with right hemisphere damage are relatively

blunted in their emotional responsivity and experience of

emotion. Thus, RHD patients will experience less anxiety

and positive feelings during the shock and prize conditions,

respectively, relative to the NHD and LHD subjects. In

contrast, LHD patients may experience more intense emotional

responsivity than NHD subjects. Specific predictions are as

follows:

1. During shock and prize anticipation, LHD subjects will

display similar or accentuated HR response patterns

compared to normal controls, whereas RHD subjects will

display decreased HR responding rel-tive to normal

controls. HR responding will be greater for the LHD

and NHD groups during shock and prize trials compared

to control trials. HR responding for the RHD group-

will not differ between shock, prize and no shock/no

reward control trials.

2. During both shock and prize anticipation, LHD patients

will display greater SCR than the normal controls. For

the RHD patients, SCR will be smaller than that of the

LHD and NHD patients. SCR will be greater during shock

and prize trials than control trials for the LHD and

NHD groups, whereas the difference in SCR for the RHD

group between shock, prize, no shock/no reward control

trials will be smaller.

3. During shock anticipation, corrugator EMG reactivity

will be similar or greater for LHD compared to the NHD

patients, whereas RHD patients will show smaller

corrugator EMG compared to NHD patients. For LHD and

NHD patients corrugator EMG will be greater for shock

trials than no shock trials. However, differences in

corrugator EMG will be smaller between shock and no

shock control trials in RHD patients.

64

4. During prize anticipation, zygomatic EMG reactivity will

be similar or greater for LHD compared to NHD patients,

whereas RHD patients will show smaller zygomatic EMG

compared to NHD patients. For LHD and NHD groups,

zygomatic EMG will be greater for prize compared to no

reward trials. However, differences in zygomatic EMG

will be attenuated between prize and no prize control

trials in RHD patients.

B) Bivalent Emotion Model: According to this view, patients

with RHD should demonstrate attenuated anxiety during the

shock anticipation condition (relative to NHD controls), and

either normal or enhanced pleasant feelings during

anticipatory reward condition. In contrast, patients with

LHD should demonstrate attenuated pleasant feelings during

the anticipatory reward condition (relative to NHD subjects)

and either normal or enhanced negative feelings during the

anticipatory shock task. These results may be most

pronounced in patients with anterior-extending lesions.

Specific predictions are as follows:

1. During shock anticipation, the LHD subjects will have

greater or similar HR responding compared to the NHD

group, whereas RHD subjects will have smaller HR

responding relative to NHD subjects. Additionally, LHD

and NHD patients will display greater HR responding

during shock relative to no shock trials, whereas HR

responding in RHD patients will not differ between

shock and no shock trials. During prize anticipation,

RHD subjects will have greater or similar HR responding

compared to normal controls, whereas LHD patients will

have smaller HR responding relative to NHD patients.

Also, RHD and NHD groups will display greater HR

responding during prize relative to no reward control

trials, whereas LHD patients will not differ between

prize and no prize trials.

2. During shock anticipation, the LHD patients will have

greater or similar SCR compared to NHD controls and RHD

patients will have smaller SCR compared to NHD

controls. Also, LHD and NHD subjects will have greater

SCR during shock compared to no shock trials, whereas

SCR will not differ between shock and no shock trials

in RHD patients. During prize anticipation, however,

RHD patients will have greater SCR than NHD patients,

while the LHD patients will have smaller SCR than NHD

subjects. Similarly, RHD and NHD patients will have

greater SCR during prize compared to no prize trials,

whereas SCR will not differ between prize and no reward

trials in LHD patients.

3. During shock anticipation, the RHD subjects will have

smaller CEMG compared to NHD patients, while the LHD

group will have greater or equal corrugator EMG

compared to NHD patients. Compared to control trials,

LHD and NHD groups will show accentuated corrugator EMG

66

during shock trials, whereas RHD patients will exhibit

no differences.

4. During prize anticipation, the RHD will have greater or

equal ZEMG compared to the NHD group which will have

greater zygomatic EMG compared to LHD group. Relative

to no reward control trials, RHD and NHD subjects will

display increased zygomatic EMG during reward trials,

whereas LHD patients will show no differences.

Specific Predictions for Experiment 2: Subjective Report of
Emotion during Anticipatory Anxiety and Reward Tasks by RHD,
LHD, and NHD Patients

The hypotheses and predictions for this experiment are

similar in kind to those of Experiment 1.

Normal control group (NHD)

1. In line with previous research, it is expected that the

NHD group will report greater state anxiety during the

shock than no shock control trials.

2. Similarly, during prize anticipation, NHD group will

report more intense positive emotions than during the

no prize control trials

Focal lesion patients (RHD and LHD)

A) Global Right Hemisphere Emotion Model: The predictions

of this model are as follows:

1. During shock anticipation, the LHD and NHD groups will

report greater anxiety (based on state anxiety scores

on the State-Trait Anxiety Inventory, dimensional

ratings of unpleasantness, arousal, powerlessness on

the Self Assessment Manikin, and the negative affect

factor score of the Positive and Negative Affect

Schedule) than the RHD group. The LHD and NHD groups

will report greater state anxiety during shock than no

shock control trials. The difference in reported

anxiety will be attenuated in RHD patients between

shock and no shock control trials.

2. During prize anticipation, the LHD and NHD subjects will

report greater positive emotions (based on dimensional

ratings of pleasantness, arousal, and dominance on the

Self Assessment Manikin, and the positive affect factor

score of the Positive and Negative Affect Schedule)

compared to the RHD. LHD and NHD groups will report

more positive emotions during prize compared to no

reward trials. The difference in reported positive

emotions will be smaller during prize compared to no

reward trials in RHD patients.

B) Bivalent Emotion Model: Predictions based on the

bivalent view are:

1. During shock anticipation, the LHD subjects will report

greater or equal anxiety compared to the NHD patients,

whereas RHD subjects will report less anxiety than the

NHD group. More anxiety will be reported during shock

compared to no shock trials for LHD and NHD patients.

68
RHD will report no differences in anxiety between shock

and no shock trials.

2. During prize anticipation, the RHD will report more or

equal positive emotion compared to the NHD patients,

whereas LHD subjects will report less more positive

emotions than the NHD group. Also, RHD and NHD

subjects will report more positive emotion during prize

compared to no reward control trials. LHD patients

will report no differences in positive affect between

prize and no reward trials.

CHAPTER 3

Subjects

A total of 48 right handed patients were included in

the study. Handedness was determined by Briggs and Nebes

(1975) abbreviated version of Annett's (1970) questionnaire.

The stroke patients were recruited through clinics,

laboratories, and medical records at Shands Teaching

Hospital at the University of Florida and the Veteran's

Administration Hospital in Gainesville. Additionally, other

subjects were recruited through neurologists, physical

therapists, and stroke clubs in the north central Florida

region. Control subjects were recruited through

laboratories at Shands Hospital and the VA, volunteer

services at the VA hospital, as well as from other local

senior groups.

All subjects were alert, cooperative, and oriented to

time, place, and person. The population consisted of four

groups; 12 patients with right hemisphere ischemic

infarctions (RHD), 12 patients with left hemisphere ischemic

infarctions (LHD), and 24 patients without neurologic

disease (12 were controls for the RHD group and 12 were

controls for the LHD group). Attempts were made to match

sex, age, and level of education across groups. There were

12 males in both the RHD and the RH NC groups. In the LHD

and LH NC groups there were 11 males and 1 female within

each group. Separate analyses of variance (ANOVA) were

used with group (LHD, LH NCS, RHD, RH NCS) as the between

subject factor to determine if there were any group

differences in age and education. There was no significant

difference in the age of the subjects between each group.

The means and standard deviations for age of each group are

as follows: RHD=63.01(9.74), RH NCS=63.92(10.63),

LHD=66.75(7.59), LH NCS=68.67(7.35).

There was also no significant differences between the

number of years of education for subjects between each

group. The means and standard deviations of years of

education for each cr:i..p are as follows: RHD=13.08(3.97),

RH NCS=14.25(2.83), LHD=12.79(2.60), LH NCS=13.83(3.95).

The ANOVA tables for the analyses examining age and

education are presented below.

AGE

SS DF MS F SIG of
F
GROUP 238.729 3 79.576 .996 .404

ERROR 3514.750 44 79.881

EDUCATION

SS DF MS F SIG of
F
GROUP 16.182 3 5.394 0.468 0.706

ERROR 507.563 44 11.536

Any patient with a pacemaker was excluded. All

subjects were questioned about hearing and visual defects.

All medications taken by the subjects on the day of the

psychophysiological measurements were recorded and a list of

these medications is provided in Table B-l, B-2, B-3, and B-

4 of Appendix B.

All subjects were administered the Zung Depression

Rating Scale. No group differences were found in their self

report of depression on the Zung [F(3,41) = 2.134, P =

.1107]. The mean scores and standard deviations on the Zung

are as follows: LHD (mean=38.636, sd=5.29); LH NCS

(mean=36.091, sd=5.28); RHD (mean=40.167, sd=7.814); RH NCS

(mean=34.091, sd=5.991).

The RHD and LHD subjects all had a CT or MRI performed

for clinical purposes. To be included, patients had a

discrete abnormal area compatible with cerebral infarction

on the head scan. Patients with tumors, hemorrhages,

trauma, or bilateral cerebral infarcts were excluded. All

subjects were tested at least 5 months post stroke in order

to control for possible changes in autonomic responsivity

over time. A t-test was conducted to examine group

differences in the amount of time since the last cortical

stroke. No differences were found between the groups

[T(1,22) = .588, P = .5626]. The average time in months for

the LHD group was 78, sd=72.72 and the average time in

months for the RHD group was 60.92, sd=69.59.

_ _

72

Using the atlas of Damasio and Damasio (1989), lesions

from the patients' CT scans were projected onto templates by

a board certified neurologist (K.H.), who was unaware of

patients' performance. Based on their scans, the

neurologist divided the stroke patients into anterior,

posterior, and mixed groups. Lesions were termed

"posterior" if located behind the central fissure or within

the posterior temporal lobe. Lesions located in front of

the central sulcus or involving the anterior temporal lobe

were considered "anterior." Lesions were considered

"primarily anterior" if they also involved the primary

sensory areas or Heschl's gyrus and "primarily posterior" if

they involved the primary motor areas. Lesions involving

both anterior and posterior regions, and/or regions between

them were considered "mixed."

All of the scans were then ranked from largest to

smallest lesion by the neurologist. The rankings were

analyzed using an independent samples Wilcoxon Test of

Ranked Sums to explore the group differences in size of

lesion. No significant differences was found between the

RHD and LHD groups [W = 141.0, P = 0.6075].

A summary of the neurological information for each

subject is provided in separate tables for each group. See

Tables B-5 and B-6 in Appendix B.

Baseline Evaluation

The baseline evaluation included a review of

neurological records along with a neuropsychological and

psychophysiological screening. All patients' neurological

records were reviewed by a neurologist prior to acceptance

into the study. All patients psychophysiological responses

to a series of 60db tones was assessed. The

psychophysiological screening is described more fully in the

procedure section for experiment 1. The neuropsychological

screening is described below.

All patients were administered the Information and

Similarities subtests of the Wechsler Adult Intelligence

Scale-Revised (WAIS-R), Wechsler Memory Scale-Revised

(Orientation, Digit Span, Logical Stories I,II and Visual

Reproductions I, II subtests), Benton Facial Recognition

Test, Western Aphasia Battery (Spontaneous Speech, Auditory

Comprehension, Repetition, and Naming subtests), Florida

Neglect Battery (shortened version including line bisection,

cancellation, visual extinction, tactile extinction, and

draw/copy a clock). The average performance on these

measures by group is presented in Table B-7. Individual

subjects' performance on these measures are presented in

Tables B-8, B-9, B-10, and B-ll in Appendix B.

T-tests were conducted to examine group differences in

neuropsychological functioning. Examination of the WAIS-R

subtests revealed that the LHD subjects had signicantly

74
lower scores on information compared with the CONS, but not

the RHD subjects. There were no significant group

differences on the similarities subtest of the WAIS-R. Both

the LHD and RHD subjects had significantly decreased digit

span forward and backwards compared to the CONs.

Results of memory testing revealed that the LHD

subjects scores on immediate recall on Logical Memory were

significantly lower than controls. However, after a delay,

the RHD subjects had significantly poorer recall compared to

the CONs. On both Logical Memory I and II, there were no

differences between the LHD and RHD subjects. RHD subjects

performed worse on Visual Reproductions I and II compared

with CONs, but not LHD subjects.

Language testing revealed that the LHD subjects had

more difficulty with comprehension and had a lower overall

Aphasia Quiotent compared with CONs and RHD subjects.

All Ss were also administered the Florida Affect

Battery. Their results on this test are provided in Tables

B-12, B-13, B-14, and B-15 in Appendix B.

Experiment 1: Psychophysiological Arousal and Facial EMG
during Anticipatory Anxiety and Anticipatory Reward in
Patients with RHD and LHD

This experiment consisted of two parts, an anticipatory

anxiety and an anticipatory reward task. In both, a two

stimulus paradigm was used whereby subjects were told that

one target tone would signal the occurrence of shock or

reward during the following 6 seconds, whereas a second

75

target tone indicated that nothing would occur during the 6

second interval. Autonomic measures of arousal (HR, SCR)

and facial EMG measures were obtained. The order of the

anticipatory anxiety task and the anticipatory reward tasks

was counterbalanced across subjects in each group.

Stimuli and Apparatus

The electrical stimuli was delivered by a Grass S88

Stimulator and Isolation Unit. A Zenith Data Systems AT

clone computer was programmed to deliver one high tone

(usually 800 or 1000 Hz) as a warning stimulus at 60 db for

one second. The computer also interacted with the

stimulator such that six seconds after presentation of a

specific tone, a shock was administered. The presentation

of a low tone (usually 400 or 600 HZ) was not followed by a

shock. For the reward task, the computer produced one high

and one low tone. Six seconds after the high tone, the

screen produced a message stating how many dollars or

lottery tickets the subject had won so far and a picture of

a smiling face. Six seconds after the low tone, nothing

occurred.

Stimulus presentation and data storage was controlled

by customized application software. Equipment for recording

heartbeat (HR), skin conductance rate (SCR), corrugator

electromyography (CEMG), and zygomatic electromyography

(ZEMG) included a set of Colbourn Instruments data

acquisition modules, a DT2805 Multifunction Board, and a

Zenith Data Systems AT clone computer.

Heartbeat was monitored by a Colbourn Instruments EKG

Coupler recorded from standard lead II. Colbourn

Instruments Bipolar comparator was used to detect the R-peak

of the EKG. Sampling occurred at 200 Hz. The output of the

Schmitt trigger was sampled at the digital input port of a

DT2805 Multifunction Board installed in a Zenith Data

Systems AT clone computer.

Skin conductance was measured by attaching 4-mm Ag/AgCl

electrodes to the thenar and hypothenar eminences of the

palm ipsilateral to the lesion. To control for possible

hand effects NHD subjects were divided into left hemisphere

normal control (LH NC) and right hemisphere normal control

(RH NC) groups. The LH NC group had electrodes placed on

their left hand and the RH NCs had electrodes placed on

their right hand. One LHD subject had skin conductance

measured on his right hand because his left arm had been

amputated. Since recent evidence (Tranel & Damasio, 1994)

suggests that brain damage subjects do not display

differential skin conductance between their right and left

hands, it was decided to include this subject in the SCR

analyses. A 0.05 m NaC1 electrolyte (Johnson & Johnson K Y

Jelly) was used. Colbourn Instruments Skin Conductance

module S71-22 was used to condition the SC signal. This is

a constant voltage system which passes 0.5v across the palm

during the recording. Sampling occurred at 20 Hz. The

analog SC signal was then be digitized by the Multifunction

board, which physically resides in the backplane of the

Compaq computer. Software control was accomplished by

customized programs.

Corrugator and zygomatic EMG was recorded using 2-mm

Ag/AgCl electrodes placed unilaterally over the corrugator

and bilaterally over the zygomatic muscle regions after the

skin was cleansed with 70% EtOH. Zygomatic EMG was

collected bilaterally because motoneuron pathways which

innervate the lower face are largely contralateral (Rinn,

1984). On the other hand, corrugator EMG was collected

ipsilaterally because motorneurons innervating the upper

face muscles are for the most part, bilateral (Rinn, 1984).

Additionally, to control for possible laterality effects,

the LH NCs had electrodes placed over their left brow and

the RH NCs had electrodes placed over their left. Muscle

regions were designated using the placement specified by

Tassinary, Cacioppo, & Geen (1989). Four Colbourn model

S75-01 High Gain Bioamplifiers with bandpass filters were

used to record the signals. Filter level was set at 90-1000

Hz and coupling at 10 Hz (Fridlund & Cacioppo, 1986). Data

was integrated with Colbourn model S76-01 Contour Following

Integrator with a time constant set at 500 milliseconds.

Sampling rate was 20 Hz.

Procedure

At the beginning of each test session, there was

approximately a 5 minute adaption period during which the

recording electrodes had been applied and the subject

relaxed while sitting in a comfortable chair in a climate

controlled shielded room. Following this adaption period,

basic physiologic reactivity (HR, SCR) to a series of 24

tones, in 8 blocks of three with two tones at 400 Hz and one

at 100 Hz (each at 60 db for .5 seconds) was measured and

the course of orienting and habituation was assessed.

The anticipatory anxiety paradigm adopted from Reiman

et al. (1989) to induce negative emotion and reward portion

of the study to evoke positive emotion were given

independently and the order in which they were given was

counterbalanced by subject for each group. For both the

anticipatory anxiety and anticipatory reward, there was 40

trials: 20 control and 20 experimental shock or reward

trials. Each trial began with a meditation period of 2 to 3

seconds, where subjects repeated the number one silently to

themselves, followed by one of four tones (between 500 and

1500 Hz for 1 second at 60db). Physiological measurements

were recorded during the last second of each baseline period

through the six second interstimulus interval and through

the six second stimulus and recovery period.

Anticipatory shock task

Before beginning the anticipatory anxiety task, each

subject choose the intensity of shock. This was done by

increasing voltage from 0 volts in five volt increments.

Half second shocks were administered after each increase in

voltage until the subject found the shock intensity

"uncomfortable but not painful."

Before the onset of the session, subjects were told

which of two tones corresponded to shock trials and which

corresponded to control trials. This two-stimulus paradigm

is similar to that used by Vrana, et al. (1989) in which a

warning signal is followed six seconds later by an electric

shock. The subjects were instructed that during the shock

trials at tone offset (after hearing the high tone), there

will be a 6 second interstimulus interval which will be

followed by a shock. In addition, the subjects were told

that when they heard the low tone, it signaled that in six

seconds, nothing would happen.

Anticipatory reward task

At the beginning of the sessions, subjects were told

which tone would indicate that they would receive a dollar

(or lottery ticket) and which tone was not associated with

reward. The higher tone always designated reward. The

designated tone for the reward trials was followed by a six-

second interval after which a message appeared on the

computer screen. The message read "You have won -- dollars"

and a smiling face. The number on the message corresponded

to the total number of dollars and/or lottery tickets won.

As in the anticipatory anxiety task, the tone designating

the control trials, the low tone, was not followed by

anything.

For both the shock and reward tasks, a square appeared

on the screen, during the 6 second period between tone and

stimulus. A cross gradually enlarged within the square. By

the end of the six seconds, the cross would touch each side

of the square and the screen would go blank. Since it is

unclear how patients with cortical strokes estimate time,

the square and growing cross were used to control for time

estimation by helping all of the subjects keep tract of time

during the six second period.

During both the anticipatory shock and anticipatory

reward tasks, the procedure was interrupted after each block

of 10 trials. At that time, the experimenter entered the

room and administered to the subjects the three-item Self-

Assessment Manikin (SAM) (Hodes, Cook, & Lang, 1985). The

SAM, which is described below, is designed as a self-report

measure of valence (pleasantness-unpleasantness), arousal,

and dominance (control).

The Self-Assessment Manikin (SAM) measures subjective

ratings of three independent affective dimensions which have

been derived from factor analytic studies (Hodes, Cook, &

Lang, 1985). The three dimensions include valence (pleasant

to unpleasant), arousal (aroused to calm), and control

(dominance to submission). There are both computer and

paper and pencil versions of SAM. In this study, a paper

and pencil version of SAM in which each dimension was

presented as a series of five cartoon characters was be

used. For the valence dimension, SAM's facial expression

gradually changes from a smile to a frown. Arousal is

denoted by increased activity in the abdomen to no activity

and wide eyes to closed eyes. Control is represented from a

very large character who gradually shrinks in size to a very

small character.

During both the anticipatory shock and the anticipatory

reward tasks, subjects were asked to rate how they felt

using the SAM. This was done after each block of 10 trials,

so that two ratings were obtained after each of the

following conditions: shock anticipation, anticipation of

no shock, reward anticipation, anticipation of no reward.

Data Reduction

Heart rate

First raw data was examined. Based on the subjects

data as a whole, missing beats and double beats were

estimated and corrected. Next, a computer program was used

to more thoroughly determine missed beats and double beats.

Double beats were removed from the data set. The computer

used the surrounding beats to estimate the missing beats.

Average half second beats/minute were obtained for each

condition for four blocks, each containing five control

trials and five stimulus trials. An average baseline score

was derived for the high and low tones for each trial block.

Beats per minute change was then determined by subtracting

the baseline value from each half second beats/minute

average for each trial block. Those values were then used

to designate average Dl, Al, and D2 for each subject for the

stimuli and control blocks within each condition. D1 was

designated as the lowest point within the first 3 seconds.

The highest point following D1 was considered Al. D2 was

the lowest point following Al. If the last value in the six

second period was the Al, D2 and Al were the same.

Skin conductance

A computer program calculated baseline, skin

conductance response (change from baseline), range-corrected

skin conductance response scores (minimum and maximum values

within each experimental condition was used in the

calculations), and half recovery time. Data was divided

into four blocks, each containing five control trials and

five stimulus trials. One average range-corrected SCR was

calculated for each stimuli and control block within each

condition. Additionally, range corrected SCR was also

recorded by changing all values under .02 micro ohms to zero.

An average recorded range corrected SCR was calculated for

each stimulus and control block within each condition.

Facial electromyoqraphy

A computer program calculated baseline corrugator

electromyography (CEMG), left zygomatic electromyography

(ZGL), and right zygomatic electromyography (ZGR) along with

average CEMG, ZGL, and ZGR over the six second period for

each block within each experimental condition. As a

consequence, for each trial block, there was one baseline

and one average score for each stimulus and control trial

for each of the three facial muscles regions: CEMG, ZGL,

ZGR. Difference scores for each of the variables was

obtained for each block by subtracting the average score

from the average baseline score for each subject.

Experiment 2: Subiective Report of Emotion During
Anticipatory Shock and Reward Tasks by RHD, LHD, and NHD
Patients

This experiment also consisted of two parts, an

anticipatory anxiety task and an anticipatory reward task.

Both were similar in kind to those of Experiment 1 except

that a 5 minute anticipatory interval was used in this study

in order to give the subjects time to complete verbal report

questionnaires about their affective state during

anticipation. In this experiment, the anticipatory shock

task and the anticipatory reward task were counterbalanced

by subject within each group.

Stimuli and Apparatus

The stimuli and apparatus used to dispense the shocks

were identical to used in Experiment 1. Additionally, two

verbal report measures of affective states were given.

These included the Self Assessment Manikin and the Positive

and Negative Affect Schedule (PANAS) (Watson, Clark, &

Tellegen, 1988). The SAM was described in the methods for

Experiment 1 above. The Positive and negative affect

schedule is comprised of two 10-item mood scales. Using

factor analysis positive affect (PA) and negative affect

(NA) factors have been identified. The directions used

were, "How are you feeling right now?" The experiment

inserted each item into the blank. Subjects were asked to

rate the intensity of each feeling on a scale of 1 to 5,

with 1 corresponding to "not at all" and 5 corresponding to

"extremely."

Procedure

This study consists of two parts, an anticipatory shock

task and an anticipatory reward tasks condition which were

counterbalanced, and described below.

Anticipatory shock task

This task had two parts, a shock and a no-shock

condition. In the shock condition, the subject waited five

minutes to receive a shock. Subjects were told that they

would receive a shock five minutes after hearing the warning

tone and that the strength of this shock was either the same

or greater than that previously given in Experiment 1. At

the end of the five minutes, subjects were given the same

intensity of shock they had previous received in Experiment

1. During the five minute anticipation period, negative

emotions were assessed using the Positive and Negative

Affect Schedule (PANAS) (Watson, Clark, & Tellegen, 1988),

and the Self-Assessment Manikin (SAM) (Hodes, Cook, & Lang,

1985). The experimenter read each item to the subjects and

recorded the responses.

In the no-shock task, subjects waited for a five minute

period with the understanding that they would not receive a

shock. The no-shock control condition consisted of a five

minute period during which time the subjects were

administered the PANAS and SAM.

Anticipatory reward task

The reward condition consisted of counterbalanced

reward and no-reward conditions. In the reward condition,

subjects waited to receive a reward. During the reward

condition, subjects were informed that they would receive

between 5 and 8 dollars, lottery tickets, or a combination

of both. Subjects were administered the PANAS and SAM while

waiting for the reward. In the no-reward condition,

subjects were informed that they were not receiving a

reward. The same questionnaires were administered during

the five minute no-reward condition.

Design Issues

A few problems inherent in the design of this project

are presented here. First, it is presumed that the positive

and negative emotions experienced in the anticipatory prize

and anticipatory shock situations will not be equal in

intensity even for the NHD group. Specifically, intensity

of anxiety/negative affect in anticipation of electric shock

will probably be greater than the intensity of joy/positive

affect in anticipation of a dollar or a lottery ticket.

However, due to financial constraints, it is not possible to

raise the financial value of the reward. Yet, by giving

each subject the choice between a dollar and a lottery

ticket, hopefully the intensity of the reward will be

maximized as each subject chooses the reward that is most

salient to him or her. In considering this problem, it may

be that autonomic, facial, and/or verbal responding are not

as pronounced as expected during the reward tasks. Yet, to

assume that differences in intensity of emotion contributed

to the lowered responsivity in the measured response

systems, attenuated responding would need to be evidenced in

all three subject groups.

Secondly, differences in baseline autonomic responding

may exist between the RHD patients, LHD patients, and normal

controls. This would make it difficult to separate deficits

in baseline autonomic responding, per se, from affective

autonomic responding. Consequently, baseline autonomic

responding will be examined in the psychophysiological

screening procedure.

Thirdly, it may be that the results of this study

provide partial support for both the global and bivalent

87

models of emotion. For example, autonomic arousal (SCR and

HR) may be mediated by the right hemisphere and hence RHD

patients would show diminished responding during both shock

and reward tasks. At the same time, facial activity, a more

accurate index of valence, may provide support for the

valence hypothesis such that RHD patients show reduced

corrugator muscle activity during negative emotions, but

accentuated zygomatic activity during positive emotion,

whereas LHD patients would show the reverse pattern. The

above is only one example of support for both the bivalent

and global models. There are other possible outcomes

indicating support for both models.

CHAPTER 4
RESULTS

First, analyses of the heart rate and skin conductance

responding during the psychophysiological orienting

procedure are presented. Next, primary analyses for

Experiment 1 are presented for heart rate, skin conductance,

ipsilateral corrugator EMG, bilateral zygomatic EMG, and

verbal report ratings separately for the shock and reward

conditions. Third, the analyses of the verbal report data

from Experiment 2 are presented.

Following the analyses of group data, data from

anterior and posterior subgroups and individual cases are

examined.

GrouD Data

Ps ych .'phE I : .. i 1 J L- r e en i nq

To review, the orienting, or physiological screening

procedure, consisted of an approximately 10 minutes period

where the subjects were instructed to sit quietly and listen

to tones. There were 8 block of three tones (24 tones

total). Two of every three tones were 1000 hz and one was

400 hz. Heart rate and skin conductances responding was

measured during the second before and six seconds following

presentation of each tone. One subject was removed from the

heart rate analyses due to unusually high and variable heart

88

rate. Additionally, one subject was removed from the skin

conductance analysis due to faulty electrode connections.

Heart rate

Average heart rate change from baseline was examined

using a Repeated Measures Analysis of Variance (ANOVA) with

group (LHD, LH NCS, RHD, RH NCS) as the between subject

factor and tone (low, high) as the within subject factor.

The low tone was the novel tone. The main effect for group

[F(3,43) = .419, P = .740], tone [F(1,43) = 1.634, P =

.208], and the interaction between group and tone [F(3,43) =

.218, P = .884] were all nonsignificant. See Table C-l in

Appendix C.

A repeated measures analysis of variance (ANOVA) was

employed to examine D1 using group as the between subject

factor (LHD, LH NCS, RHD, RH NCS) and tone (high, low) and

block (1 to 8) as the within subject factors. Results

revealed a main effect for tone [F(1,43) = 8.63, P < .011

such that there was a greater D1 for the low tone (the novel

tone) compared to the high tone (the repeated tone). The

mean D1 for the low tone was -3.4 (sd=4.40) bpm change from

baseline whereas the mean D1 for the high tone was -2.5

(sd=3.82) bpm change from baseline. None of the other

effects were significant. See Table C-2, the full ANOVA

table, in Appendix C.

Skin conductance

The percentage of responses greater than .02 micro

sieman was analyzed using a repeated measures analysis of

variance with group (LHD, LH NCS, RHD, RH NCS) as the

between subjects factor and tone (low and high) as the

within subject factor. One RHD subject was excluded due to

faulty electrode connections which resulted in corrupt data.

Results revealed that the main effect of group, tone, and

the group by tone interaction were not significant. The

mean percentage of responses and standard deviations for

each group were as follows: LHD, mean=8.07%, sd=25.73; LH

NCS, mean=25.52%, sd=36.16; RHD, mean=7.67%, sd=19.19; RH

NCS, mean=29.69%, sd=26.47. The full ANOVA table, Table C-

3, is presented in Appendix C.

The recorded range corrected skin conductance response

(SCR) was analyzed using a repeated measures analysis of

variance (ANOVA) with group (LHD, LH NCS, RHD, RH NCS) as

the between-subject factor and tone (low, high) and block (1

to 8) as the within subject factors. As mentioned above,

one subject was excluded due to corrupt data. The main

effect for group [F(3,43) = 1.91, P =.1421], block [F(7,43)

= 1.20, P = .3017], and tone [F(1,43) = .21, P = .6495] were

not significant. The interactions between block and group

[F(21, 301) = .70, P = .8305], tone and group [F(3,43) =

.33, P = .80], and between block, tone, and group [F(21,301)

= .87, P = .6244] were also not significant. The full A::'.A

91

table, Table C-4 is presented in Appendix C. The means and

standard deviations for each group collapsed across tone and

block were: LHD mean=3.881, sd=13.690; LH NCS mean=14.691,

sd=27.809, RHD mean=3.895, sd=14.528; RH NCS mean=13.549,

sd=23.073.

To sum, during the psychophysiological screening

procedure, subjects had a greater heart rate D1 to the novel

tone. There were no differences between the tones in

overall heart rate, percentage of SCR responses, or amount

of skin conductance responding. Additionally, there were no

group differences found for either heart rate or skin

conductance.

Experiment 1

Experiment 1 consisted of two tasks (shock or reward).

During each condition, heart rate, skin conductance,

ipsilaceral corrugator EMG, and bilateral zygomatic EMG were

recorded during a three second baseline, tone onset, and a

six second anticipation period. Within each task, the tone

onset signaled either a stimulus or control trial. High

tones always signaled stimulus trials (i.e., shock and

reward) and low tones always signaled control trials. There

were 40 trials within each task which were divided into four

10-trial blocks. Within each block there were 5 stimulus

and 5 control trials. Subjects were administered the Self

Assessment Manikin at the end of each 10-trial block.

Shock task

As mentioned above, subjects received the shock in the

forearm ipsilateral to their lesions. Additionally, RH NCS

and LH NCS received the shock on their right and left arms

respectively. Subjects were asked to determine the level of

shock that was "uncomfortable, but not painful." The level

of shock chosen by the subjects was examined using a 1

factor ANOVA with group (LHD, LH NCS, RHD, RH NCS) as the

between subject factor. There were no group differences in

the voltage of shock chosen [F(3,44) = 1.79, P = .1622].

The means and standard deviations for each group in volts

are as follows: LHD group (mean=68.75, sd=14.79), LH NCS

(mean=57.08, sd=12.70), RHD group (mean=64.17, sd=12.58), RH

NCS (mean=64.58, 9.40). The ANOVA table is presented below.

Table 4-1 ANOVA Table of Amount of Shock

SS DF MS F Sig
of F

Group 843.229 3 281.07 1.79 .1622

Residual 6893.750 44 156.67

Heart rate. A series of separate analyses were

conducted to examine several heart rate variables. These

included overall heart rate change from baseline, D1 (the

greatest deceleration within the first 3-seconds after tone

offset), Al (the greatest acceleration following D1 within

the 6-second period), and D2 (the greatest deceleration

following Al within the 6-second period). One subject was

excluded from the LH NC group due to unusually high and

variable heart rate. Figure C-1, C-2 and C-3 depict the

heart rate wave forms in half second intervals for the NCS,

RHD, and LHD subjects respectively.

Average heart raze change from baseline was examined

using repeated measures analyses of variance (ANOVAs) for

the shock condition with group as the between subjects

factor (LHD, LH NCS, RHD, RH NCS) and condition (shock, no-

shock) as the within subject factor. The analyses

revealed no group differences [F(3,43) = 1.55, P = .214] as

well as no differences between the shock and no-shock

conditions [F(1,43) = .050, P = .824]. The interaction of

group and condition was also not significant [F(3,43) =

.927, P = .436]. The means for each group were as follows:

LHD mean=-.558, sd=.985; LH NCS mean=-.130, sd=.650; RHD

mean=.139, sd=1.556; RH NCS mean=-.121, sd=l.01. The

complete ANOVA table is depicted in Table C-5 of Appendix C.

Heart rate D1 was examined using repeated measures

analyses of variance ANOVAS) with group (LHD, LH NCS, RHD,

RH NCS) as the between subject factor. The two within

subject factors were block (1 to 4) and condition (shock and

no-shock). Analysis of D1 revealed that there was a

significant three way interaction between group, condition,

and block [F(9,43) = 2.09, P < 05]. Money of the other

interactions or main effects were significant. The full

	Title Page
	Acknowledgement
	Table of Contents
	Abstract
	Review of the literature
	Statement of the problem
	Methods
	Results
	Discussion
	Appendix A: Psychological...
	Appendix B: Demographic inform...
	Appendix C: Statistical inform...
	References
	Biographical sketch
	Back Matter

MILLISECOND	CLASS.METHOD	MESSAGE
0	sobekcm_page_globals.constructor
0	sobekcm_page_globals.constructor	Application State validated or built
0	sobekcm_database.verify_item_lookup_object
0	sobekcm_page_globals.constructor	Navigation Object created from URI query string
0	sobekcm_database.verify_item_lookup_object
0	sobekcm_page_globals.display_item	Retrieving item or group information
0	sobekcm_page_globals.get_entire_collection_hierarchy	Retrieving hierarchy information
0	sobekcm_assistant.get_entire_collection_hierarchy
0	cached_data_manager.retrieve_item_aggregation
0	cached_data_manager.retrieve_item_aggregation	Found item aggregation on local cache
0	item_aggregation_builder.get_item_aggregation	Found 'all' item aggregation in cache
0	system.web.ui.page.page_load (ufdc.page_load)
0	sobekcm_page_globals.constructor.on_page_load
0	html_echo_mainwriter.add_style_references	Adding style references to HTML
0	html_echo_mainwriter.add_text_to_page	Reading the text from the file and echoing back to the output stream
77	html_echo_mainwriter.add_text_to_page	Finished reading and writing the file