UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Emotion and psychopathy: a three-component analysis Forth, Adelle E. 1992

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1993_spring_forth_adelle2.pdf [ 4.93MB ]
JSON: 831-1.0086378.json
JSON-LD: 831-1.0086378-ld.json
RDF/XML (Pretty): 831-1.0086378-rdf.xml
RDF/JSON: 831-1.0086378-rdf.json
Turtle: 831-1.0086378-turtle.txt
N-Triples: 831-1.0086378-rdf-ntriples.txt
Original Record: 831-1.0086378-source.json
Full Text

Full Text

EMOTION AND PSYCHOPATHY:A THREE-COMPONENT ANALYSISByADELLE ELIZABETH FORTHB.A., University of British Columbia, 1982M.A., University of British Columbia, 1987A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIES(Department of Psychology)We accept this thesis as conformingto the required standard.UNIVERSITY OF BRITISH COLUMBIADecember, 1992© Adelle Elizabeth Forth, 1992In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.Department of ^PsychologyThe University of British ColumbiaVancouver, CanadaDate 'Ne.. DE-6 (2/88)AbstractThe study was designed to examine the hypothesis thatpsychopathy is associated with an affective deficit. Subjectswere 42 incarcerated offenders divided into nonpsychopathic andpsychopathic groups based on their scores on the Hare RevisedPsychopathy Checklist (Hare, 1991). Facial expressions, centraland peripheral physiological activity, and subjective ratingsof affective valence and arousal were measured during exposureto a series of slides and film clips designed to elicit eitherpositive or negative affective states. The results indicatethat psychopaths do not differ from criminal controls in theiraffective self-report, autonomic nervous system response, orobserved facial expressions to emotional stimuli. However, withrespect to cerebral asymmetry, psychopaths failed to showrelative right frontal activation during exposure to thedisgust film. This result is discussed in relation to recentattempts to explain psychopathy in terms of lateralizedcerebral dysfunction.Table of ContentsAbstract  ^iiTable of Contents ^  iiiList of Tables List of Figures ^  viiAcknowledgements  ixIntroduction  ^1Psychopathy  ^4The Assessment of Psychopathy  ^4Affective Processing in Psychopathy  ^6Clinical Observations  ^6Empirical Studies  ^8Psychophysiology of Emotion  ^16Facial Activity  ^18Autonomic Activity ^  23Cerebral Asymmetry and Emotion ^  26Purpose of Current Study ^  28Experimental Hypotheses  30Experiment 1 ^  31Method  32Subjects  ^32Stimulus Materials  ^34Apparatus and Response Measurement ^  34Procedure  ^37Self-report Questionnaires  ^39Data Analysis  ^40-iv-^Results  ^43Self-report Measures  ^43Skin Conductance  ^50Heart Rate  ^53Facial Activity  ^55Electrocortical Data  ^58Recognition Reaction Time Task  ^61Self-report Questionnaires  ^66Discussion  ^68Experiment 2  ^69Method ^  70Subjects  ^70Stimulus Materials  ^70Apparatus and Response Measurement ^  71Procedure ^  72Data Analysis  ^73Results ^  75Self-report Data  ^75Affect Grid ^  81Physiological Data  ^84Facial Expressions  ^88Electrocortical Data  ^93Discussion  ^98General Discussion ^  101References ^  113Appendix A  126Appendix B ^  128-v-List of TablesTable 1.Table 2.Table 3.Table 4.Table 5.Table 6.Table 7.Table 8.Table 9.Individual slides, rank-ordered by affectivevalence from most pleasant to least pleasantaveraged across subjects from each group ^ 44Individual slides rank-ordered by arousallevel from most arousing to least arousingaveraged across subjects from each group ^ 45Mean (SD in brackets) number of upper andlower facial actions for each group as afunction of slide valence  ^56Mean (SD in brackets) reaction times (msec)and error rates for familiar slides for eachgroup as a function of slide valence ^ 62Mean (SD in brackets) reaction times (msec)and error rates for distracter slides for eachgroup as a function of slide valence  ^64Percentage of nonpsychopaths and psychopathswho correctly classified slides  ^65Mean scores (SD in brackets) on self-reportquestionnaires for psychopaths andnonpsychopaths ^  67Emotional ratings in the film conditions,for all subjects  ^79Percentage of nonpsychopaths and psychopathswho rated the target emotion the strongest ^ 82-vi-Table 10. Percentage of nonpsychopaths and psychopathsrating positive or negative emotions thestrongest across film conditions  ^83Table 11. Number of nonpsychopaths and psychopathsdisplaying discrete facial expressionsduring film clips  ^90-vii-List of FiguresFigure 1. Self-ratings of valence (upper panel) andarousal (lower panel) for pleasant, neutral,and unpleasant slides, by diagnostic group ^ 47Figure 2. Self-ratings of valence (upper panel) andarousal (lower panel) for low, medium, andhigh arousal slides, by diagnostic group ^ 49Figure 3. Mean log-transformed skin conductanceresponse magnitude to low, medium, and higharousal slides, by diagnostic group ^ 51Figure 4. Mean log-transformed skin conductanceresponse magnitude to pleasant, neutral,and unpleasant slides, by diagnostic group ^ 52Figure 5. Mean heart rate change (beats per minute)to pleasant, neutral, and unpleasant slides,by diagnostic group ^  54Figure 6. Mean log-transformed alpha power (in )V2 /Hz)for the left and right frontal (F3 and F4,respectively: upper panel) and parietal(P3 and P4, respectively: lower panel)regions during the pleasant, neutral, andunpleasant slides ^  60Figure 7. Mean self-report ratings to the amusing film(upper panel) and happy film (lower panel)by diagnostic group ^  76Figure 8. Mean self-report ratings to the sad film(upper panel) and fear film (lower panel)by diagnostic group ^  77Figure 9. Mean self-report ratings to the anger film(upper panel) and disgust film (lower panel)by diagnostic group ^  78Figure 10. Film clips scaled on the dimensions ofpleasure-displeasure and low to high arousal(based on Affect Grid) ^  85Mean heart rate change (in beats per minute)for nonpsychopaths and psychopaths duringfilm clips ^  86Mean skin conductance response for nonpsychopathsand psychopaths during film clips ^ 87Mean number of facial expressions of enjoymentand disgust displayed by nonpsychopaths andpsychopaths to the disgust film ^ 92Figure 14. Mean log-transformed alpha power (in )V2 /Hz)for the left and right frontal (F3 and F4,respectively) region during the amusing anddisgust films for nonpsychopaths andpsychopaths ^  95Figure 15. Mean log-transformed alpha power (in )1V 2 /Hz)for the left and right parietal (P3 and P4,respectively) region during the amusing anddisgust films for nonpsychopaths andpsychopaths ^  96Figure 11.Figure 12.Figure 13.AcknowledgementsI would like to express my thanks to the many people whohave helped directly and indirectly to complete this research.First, I wish to express my gratitude to my parents, brother,sisters, and to all my friends for their constant support.Second, I would like to thank those who contributed tothis study: my committee members, Wolfgang Linden and JohnYuille for their useful suggestions and advice; Susan Nurse andMichael Laycock for doing the psychopathy ratings; Gary Birch,Charlie Seeburg, and Gary Maclssac for their technical help;the inmates and staff at Matsqui Institution for theircooperation; Shelley Parlow and John Logan who have kept mehappy and sane; and all my friends at the Psychophysiology Labwho have given me support and help.I would also like Stephen Hart to know how much Iappreciate his friendship. It has been a privilege to work withhim on our many joint research projects.Finally, and most importantly, I would like to givespecial thanks to my supervisor, Robert Hare, who nurtured myinterest in the area of psychopathy and whose ideas helped formthe basis of this and much of the other research I have beeninvolved in. I would also like to thank him for his generoussupport, guidance, and encouragement throughout my graduatecareer.-1-IntroductionFirst clinically defined in the early 19th century (Pinel,1809; Prichard, 1835; Rush, 1812), the concept of psychopathyis still surrounded by controversy. Nevertheless, mostclinicians and researchers agree that psychopathy is a seriouspersonality disorder that first manifests itself early in lifeand that, in most cases, persists throughout most of thelifespan. The disorder is associated with a constellation ofaffective, interpersonal, and behavioral characteristics (seeCleckley, 1976; Craft, 1966; Hare, 1991; McCord & McCord,1964). One of the most salient characteristics is an apparentaffective deficit in which the individual seems unable to formlasting bonds with others, displays shallow and short-livedemotions, is lacking in empathy, and has a general lack ofconcern or remorse for his actions. Interpersonalcharacteristics include egocentricity, manipulativeness,glibness, and dominance. Behaviorally, psychopaths aresensation-seekers, impulsive, and irresponsible. Because ofthese traits many psychopaths fail to fulfill socialobligations and come into frequent contact with the criminaljustice system.Many clinical descriptions of the psychopath haveemphasized the inability to experience or to appreciate theemotional significance of events. Several of the major featuresof the disorder--glibness, lying, manipulation, lack ofempathy, shallow affect--may be reflections of a core affective-2-deficit. Cleckley (1976), who has provided one of the mostcomprehensive descriptions of the syndrome, hypothesized thatthe psychopath suffers from a deep-seated deficit in which adissociation exists between affect and cognition. He proposedthat this profound affective disturbance, which he termedsemantic dementia, resulted in "a selective defect orelimination or attenuation of those strong affective componentswhich ordinarily arise in major personal and social issues"(Cleckley, 1976, p. 230). He considered this deficit to becentral rather than peripheral, affecting the appreciation ofthe affective substance of most life experiences, but "withoutobvious damage to the outer appearance or superficial reactionsof the personality" (p. 237). Although the notion of anaffective deficit has been considered central to the concept ofpsychopathy, there actually have been few empirical studies onthe topic.While an extensive body of literature is devoted toclinical descriptions of psychopathy and to assessment issues,it is only recently that researchers have begun tosystematically direct their attention to the disorder, usuallyfrom the point of view of learning theory, particularly theclassical conditioning of autonomic responses and passive-avoidance learning (e.g., Gorenstein & Newman, 1980; Gray,1975; Hare, 1970; Lykken, 1957; Schmauk, 1964). In many cases,the findings are interpreted in terms of affective processes.For example, the frequent finding that psychopaths give-3-relatively small electrodermal responses in anticipation of anaversive stimulus led Hare (1978) to conclude that "situationsthat have great emotional impact for most people may be ofrelatively little consequence to psychopaths (p. 137)."In the past few years the emphasis has become morecognitive in nature, with interesting results. For example,several studies have found that psychopaths appear to differfrom others in the processing, use, and cerebral organizationof language (Hare & McPherson, 1984; Hare & Jutai, 1988). Thereis some evidence that these differences may involve theaffective components of language (Gillstrom & Hare, 1988; Hare& Jutai, 1988; Williamson, Harpur, & Hare, 1991).The results of several studies raise the possibility thatpsychopaths are characterized by poor integration of thereferential and affective components of language (Williamson,Harpur, & Hare, 1990, 1991). This "deficit" appears to berelated to an insensitivity to emotional valence. Although mostof the research on affective processing in psychopaths hasinvolved linguistic stimuli, the possibility exists that theysuffer from a more generalized affective deficit.Given that between 15 and 25% of inmate populations arepsychopaths (Hare, 1991), and that psychopaths are responsiblefor a disproportionate amount of crime (Hare & McPherson,1984), the emotional and financial costs to society areenormous. Because they play an important role in the mediationand regulation of behavior, the study of affective processes-4-may provide us with a better understanding of psychopathy andmay provide important insights into the development andmaintenance of the disorder.The research described below provides the background forthe current study on psychopath's affective reactions toemotionally eliciting stimuli. A brief review of diagnosticissues and the literature relevant to psychopathy, affect, andthe cerebral organization of language is presented. Theliterature on the association between emotions and facialexpressions is also summarized, followed by summaries ofresearch on the autonomic correlates of emotional processingand the cerebral asymmetry of emotion.The present study is an initial step in the systematicinvestigation of affective reactions in psychopaths. Towardsthis end, all three components of emotion--physiological,subjective, behavioral--were included as dependent variables.The purpose of the study was to evaluate the hypothesis thatpsychopathy is associated with an affective deficit.PsychopathyThe Assessment of PsychopathyPerhaps the largest obstacle to research in this area isthe unsystematic and variable way in which psychopathy has beenassessed. Until recently, a variety of different assessmentprocedures, many with poor reliability and validity, have been-5-used. Because of the lack of any consistent assessment criteriaand the absence of any unitary theoretical perspective, it isdifficult to assess the validity of the research or to comparethe findings from different studies.The terms sociopathy, psychopathy, and antisocialpersonality disorder (APD; American Psychiatric Association,1987) are often used to refer to the same underlying concept.However over the years, the terms sociopathy and APD havebecome to rely almost exclusively on easily measured behavioralcriteria, particularly delinquent and criminal behaviors,presumably relevant to psychopathy. Such procedures are oftenreliable but they lack demonstrated validity. They fail tomeasure the affective and interpersonal characteristics thathave traditionally been considered central to this disorder.Other procedures, such as the Psychopathy Checklist (PCL) andits revision (PCL-R; Hare, 1991; Hart, Hare, & Harpur, inpress), make explicit use of inferences about personalitytraits. Because of the extensive literature on its reliabilityand validity with forensic populations (e.g., Hare, 1985; Hare,1991; Harpur, Hare, & Hakstian, 1989; Kosson, Smith, & Newman,1990; Wong, 1984), the PCL-R was used in the present study.The PCL-R consists of two stable correlated factors (Hareet al., 1990; Harpur, Hakstain, & Hare, 1988). Both areconsidered essential in the diagnosis of psychopathy. Factor 1reflects interpersonal and affective characteristics, such asegocentricity, callousness, and manipulativeness, that are-6-considered fundamental to the traditional clinical conceptionof psychopathy. This factor is most strongly correlated withself-report measures of machiavellianism, narcissism, andempathy (Hare, 1991). Factor 2 reflects the behavioralcharacteristics of psychopathy, such as impulsivity, chronicinstability, an irresponsible lifestyle, and antisocialbehavior. This Factor is related to the DSM-III-R diagnosis ofAPD, criminal behaviors, and self-report measures of sensation-seeking and substance abuse (Harpur et al., 1989; Hemphill,Hart, & Hare, 1991).Affective Processing in Psychopaths Clinical Observations Early clinical accounts of what we now refer to aspsychopathy assigned an important role to a core affective"disturbance." Pinel (1809) described a condition he called"manie sans delire" (madness without confusion) in which reasonremained intact but emotion or affect was disturbed. Rush(1812), an American psychiatrist, described cases characterizedas "morally deranged", where there was antisocial behavior inthe apparent absence of thought disorder. Prichard (1835) usedthe term "moral insanity" to describe those in whomthe moral and active principles of the mindare strongly perverted or depraved; the powerof self-government is lost or greatlyimpaired and the individual is found to be-7-incapable not of talking or reasoning uponany subject proposed to him, but ofconducting himself with decency and proprietyin the business of life (p. 15).The more recent clinical literature is full ofdescriptions about the psychopath's emotional coldness and lackof feeling. Cleckley (1976), who has provided one of the mostcomprehensive and influential descriptions of the psychopath,lists 16 defining characteristics of the psychopath. Of these,four pertain directly to affect: lack of remorse or shame;pathologic egocentricity and incapacity for love; generalpoverty in the major affective reactions; and unresponsivenessin general interpersonal relations. Cleckley (1976) believedthat the psychopath "can learn to use ordinary words... (and)will also learn to reproduce appropriately all pantomime offeeling...but the feeling itself does not come to pass(p.230)."McCord & McCord (1964), after a review of the literature,provide a succinct and effective description of thepsychopath's emotional status. They say:The psychopath feels little, if any, guilt.He can commit the most appalling acts, yetview them without remorse. The psychopath hasa warped capacity for love. His emotionalrelationship, when they do exist, are meager,fleeting, and designed to satisfy his owndesires. These last two traits, guiltlessnessand lovelessness, conspicuously mark thepsychopath as different from other men (pg.17).-8-Although these descriptions of affective reactions arebased on clinical observations, they suggest that thepsychopath is lacking in an important dimension of experience,namely, affect. As Cleckley (1976) states, "...a good deal ofthe affective substance that people find in life experience islacking in the psychopath's responses..." (p. 231).The PCL-R is conceptually and empirically closely linkedto Cleckley's conception of psychopathy, and includes severalitems related to affect: lack of remorse or guilt; shallowaffect; and lack of empathy.Empirical Studies Early research into the autonomic correlates of emotionalreactivity in psychopaths focused on the arousal of fear oranxiety. The results from a number of studies (Hare, 1970; Hare& Craigen, 1974; Hare, Frazelle, & Cox, 1978; Hare & Quinn,1971; Lykken, 1957), indicate that psychopaths tend to displaysignificantly smaller skin conductance responses (SCR), butlarger heart rate acceleration, in anticipation of anunpleasant event (e.g., a loud tone, electric shock) than dononpsychopaths.In order to explain this dissociation betweenelectrodermal and cardiac responses, Hare (1978) used Lacey's(1967; Lacey & Lacey, 1974) hypothesis concerning cardiacchanges and "sensory intake" and "sensory rejection." Accordingto this theory, cardiac responses reflect the way sensory input-9-is modulated. Lacey (1967) argued that cardiac deceleration,which is accompanied by decreased pressure in baroreceptors inthe carotid sinus, is associated with increased corticalarousal and sensory intake. Cardiac acceleration is associatedwith increased pressure in the carotid sinus, a decrease incortical arousal, and sensory-rejection. Hare (1978, 1982)suggested that the psychopath's pattern of skin conductance andheart rate may be indicative of a unique coping mechanism. Theheart rate acceleration is part of a defensive response thatacts to lessen the impact of the unpleasant stimulus, whereasthe skin conductance hyporesponsivity reflects the success ofthis coping strategy.Support for this coping hypothesis comes from two otherstudies. Hare (1982) found that psychopaths did not showcardiac acceleration in anticipation of an aversive tone ifthey were given the opportunity to distract themselves (bylistening to a night-club comic). More recently, Ogloff & Wong(1990) provided psychopaths and nonpsychopaths with two tasks.One was identical to the countdown task used by Hare and hiscolleagues (the aversive stimulus was unavoidable), whereas inthe other the subject could prevent the aversive stimulus frombeing delivered by pressing a button. The psychopaths showedanticipatory heart rate acceleration when the aversive stimuluswas unavoidable but not when it could be avoided.Research on the psychophysiology of memory retrieval foremotional text (Vrana, Cuthbert, & Lang, 1989) indicates that-10-physiological responding varies as function of processing mode(i.e., imagining a sentence vs. silently repeating it) andsentence type (i.e., fearful or neutral). Heart rateacceleration is greatest during imagery of fearful material. Arecent study, by Patrick, Cuthbert, & Lang (1990), tested agroup of sex offenders with this paradigm. Despite similaraffective ratings of imaginal experience, psychopathic sexoffenders (defined by the PCL-R) displayed smallerphysiological responses during the imagery of fearful materialthan did nonpsychopathic sex offenders.Several studies have reported that the size of thereflexive eyeblink to a startle stimulus varies as a functionof the affective valence of the foreground (Bradley, Cuthbert,& Lang, 1990). The startle response is larger while viewingnegative affective slides than while viewing positive affectiveslides or neutral slides. The startle reflex is a defensive orprotective response to unexpected stimuli. The modulation ofthis reflex is dependent on general matching principles. Whenthe foreground is aversive, the startle reflex is enhanced;conversely, when the foreground is pleasant, the startleresponse is diminished. The startle response paradigm thusprovides a potentially useful method for studying emotionalprocessing in psychopaths.Preliminary results from a study of the blink-startlereflex in sex offenders (Patrick, Bradley, & Lang, 1991)indicated that the startle responses of nonpsychopaths werelarge when they viewed unpleasant slides, and small when theyviewed pleasant slides (in each case, relative to when theyviewed neutral slides). Psychopaths (defined by the PCL-R), onthe other hand, gave smaller responses when they viewedunpleasant and pleasant slides than when they viewed neutralslides. Although preliminary, these results provide support forthe notion that psychopaths differ from others in theirphysiological responses to affective stimuli.Recent research by Hare and his colleagues has looked atthe processing of affective linguistic stimuli in psychopaths.The hypothesis that psychopathy is associated with abnormalprocessing of affective verbal material was investigated byWilliamson et al. (1991). Criminal psychopaths andnonpsychopaths performed a lexical decision task (decidingwhether or not a letter-string is a word) with affective andneutral words, and nonwords. Psychopaths failed to show thebehavioral or electrocortical differentiation between affectiveand neutral words shown by the nonpsychopaths. The latterresponded significantly faster, and had larger amplitude event-related potentials (ERPs), to affective words than to neutralwords. This pattern of results is one that has been found instudies with noncriminals. Psychopaths failed to show fasterreaction times to affective than to neutral words, and showedno ERP differentiation between word types. Their failure todifferentiate between neutral and affective words suggests thatemotional words do not have the same affective or motivational-12-significance for psychopaths as they do for others.Additional support for language-based affective anomaliescomes from a recent experiment by Williamson et al. (1990) inwhich criminal psychopaths and nonpsychopaths were presentedwith a group of three words and asked to choose which two words"go together the best". Words could be grouped on the basis oftheir denotative or connotative meaning. The word pairs wereclassified into six types: antonym, domain, metaphor, polarity,domain and polarity, and no relation. If psychopaths areinsensitive to the affective components of words, it washypothesized that they would use connotation (polarity andmetaphor) less often when grouping them than wouldnonpsychopaths.The results partially supported the hypothesis. Bothgroups were similar in their use of denotation (antonym,domain) to group words. However, psychopaths grouped words onthe basis of their connotative meaning less frequently than didnonpsychopaths. Further analysis revealed that the differencewas confined to the use of polarity; psychopaths used metaphorto group words to the same extent as did the nonpsychopaths.The ability of psychopaths to use metaphor may reflect thefact that affective judgments are not necessarily required formaking metaphorical associations. Thus psychopaths may havebased their judgments on a learned association between thewords, and not on the affective significance of the words.The Williamson et al. (1990) study involved the processing-13-by psychopaths of linguistic stimuli. Psychopaths have alsobeen found to differ from others in the use of language-relatedhand gestures (Gillstrom & Hare, 1988; Hare & Gillstrom, 1990).In both studies videotaped interviews were used to code handgestures. Language-related gestures were divided into two maincategories: iconic gestures, and beats. Iconic gestures arerelated to the semantic content of speech whereas beats arenot. Psychopaths used more beats than did nonpsychopaths, agroup difference that was larger during segments of theinterview that dealt with family life than during segments thatdealt with criminal activities. In interpreting this findingthe authors made use of the evidence that excessive use of beatgestures is related to speech-encoding difficulties. Theysuggested that the psychopaths had difficulty inconceptualizing and encoding material about their family lifebecause of the interpersonal/affective components involved.This interpretation is consistent with previous research(discussed above) indicating that psychopaths are relativelyinsensitive to the affective components of language.These studies support the notion that psychopaths may havedifficulty in integrating the semantic and affective componentsof language. It is of course possible that psychopathy ischaracterized by a more generalized affective deficit, one thatis not confined to language. In an initial attempt to study abroader range of affective processes and abilities inpsychopaths, Hare, Forth, & Hart (1989) administered a battery-14-of affective tests to a sample of male offenders, psychopathsand nonpsychopaths. The first three tests were affect-matchingtasks using pictures, sentences and facial expressions. Thesubject's task was to choose from four stimuli the one thatmost closely matched the emotional tone of a target stimulus.Because depressed mood has been found to affect performance onthis test (Williamson, Crockett, Hurwitz, & Remick, in press),the Beck Depression Inventory (BDI; Beck, Ward, Mendelson,Mock, and Erbaugh, 1961) was also administered to each subject.There were no group differences on any of the tasks, a resultthat remained unchanged when scores on the BDI were taken intoaccount. These results suggested that psychopaths are notcharacterized by a deficit in the matching of affect.In a more detailed analysis of the picture and sentenceaffect-matching task with a larger sample of offenders,Williamson et al. (1990) found that differences betweenpsychopaths and nonpsychopaths emerged when the type of errorsmade by the subjects was analyzed. Errors could be made in anumber of ways: opposite (matching emotions on the basis of theopposite valence), descriptive (matching the target on thebasis of shared visual/content features), or neutral (choosingthe stimuli that did not match in emotional tone or indescriptive features). If psychopaths are insensitive toaffective valence, they should tend to make more oppositeerrors than do normal individuals. The results indicated thatpsychopaths tended to confuse emotions of opposite polarity in-15-both tasks, but the effect was greater for the linguistic taskthan for the picture task.A study by Day (1991) focused on the lateral processing ofemotional words and faces in psychopaths. He found thatpsychopaths and nonpsychopaths employed different processes toevaluate the emotional significance of language but not facestimuli. Nonpsychopaths, like normal controls, displayed theexpected right hemisphere advantage, whereas psychopathsdisplayed a left hemisphere advantage in processing emotionalintensity of negatively-toned words. However, both psychopathsand nonpsychopaths displayed a right hemisphere advantage inprocessing facial stimuli. He concluded that the reduced righthemisphere performance to language stimuli observed inpsychopaths was due to difficulty in decoding the emotionalcomponents of linguistic information.Taken together, the results of these studies suggest thatpsychopathy is characterized by poor integration of, ordissociation between, the affective and semantic components oflanguage. Whether or not psychopathy is also characterized by adissociation between affect and nonlinguistic processes isunknown at this time. If a more generalized affective deficitdoes exist with psychopaths, it could be tied in with thesuggestion by Ax (1962) that in order to accurately perceiveanother's feelings or to empathize with him it is necessary toconstruct an "emotional facsimile" of the other person. Thisinvolves the higher cortical and subcortical centres of the-16-autonomic nervous system. Hare et al. (1988) speculated thatpsychopaths are only able to empathize with others in theintellectual sense, since the affective components of theemotional facsimile are absent.Clinical observations suggest that the psychopath fails toappreciate the emotional significance of events. Both thepsychopath's words and behaviors often appear to reflect someform of an affective deficit. Although the empirical literatureis still sparse the studies described above suggest that thestudy of affective functioning in psychopaths is likely to befruitful.Psychophysiology of EmotionAlthough they differ in their precise definition ofemotion, most major theorists agree that there are threeaspects or components to an emotional reaction: (1)physiological/autonomic; (2) behavioral/expressive; and (2)cognitive/experiential (for a comprehensive review see Izard,Kagan & Zajonc, 1984; Lang, 1978). However, the nature of therelationship among these components is not yet well understood.The purpose of this section is to briefly review recentresearch in this area, with emphasis on facial and autonomicnervous system (ANS) activity.There are two major models used to conceptualize how humanemotions are organized: the discrete model and the dimensionalmodel (Levenson, 1988). According to the discrete model, there-17-are well-defined "primary" emotions, such as fear, anger, andsadness. Researchers using this model have focused theirefforts on uncovering correlates of these emotions, rather thanon examining intercorrelations among emotions. In contrast,dimensional models propose that there is a limited number ofunderlying dimensions, such as negative-positive or active-passive, upon which emotional states are located. Researchersusing this model have typically focused on ways in which peoplethink about and form judgments about emotional phenomena.Schlosberg (1954) was the first to demonstrate that facialexpressions of emotions were characterized by a relativelysimple underlying structure. He suggested that a two or threedimensional scheme could account for much of the variance. Morerecent research, using a wide variety of stimuli, populations,and procedures, has also found that two dimensions consistentlyemerge (Russell & Mehrabian, 1977; Smith & Ellsworth, 1985).One dimension reflects an evaluation of the emotion in terms ofpleasant or unpleasant and the other reflects the intensity ofthe emotion from arousal to sleepiness.A circular model of emotion, based on the two abovedimensions, was first proposed by Schlosberg (1954) and todayconsiderable evidence exists to support this model (Plutchik,1980; Russell, 1980; Russell & Bullock, 1986). Current evidencesuggests that emotion categories are systematically and highlyinterrelated (see review by Russell, 1989). The circumplexmodel specifies the degree to which emotions are interrelated.-18-For example, happy and delighted are closer together on thecircumplex then are happy and sad.Facial Activity. The overt expression of emotions ismediated by the skeletal muscle system. These muscles arefinely regulated by the brain to produce precise and highlycomplex patterns of activity across muscles and over time. Theface in particular, with its rich neural innervation (and itslarge number of muscles), is capable of reflecting differentfundamental emotions and patterns of emotions (Ekman & Friesen,1978). It therefore provides an excellent window for studyingaffective processes. The capability of the facial musculatureto express emotional states was first systematicallyinvestigated by Duchenne, a French anatomist who electricallystimulated specific facial muscles to assess their role in theproduction of facial expressions (Duchenne, 1862/1990).The face can be divided into two main regions: the upperface (brows, forehead, and eyes) and the lower face (cheeks,mouth and chin). These regions function in relativeindependence of each other, and they also differ in the degreeof fine motor control available to the muscles. The muscles ofthe lower face are exclusively controlled by the contralateralhemisphere, and are bilaterally independent in their function.This region is controlled by many small muscles that can movein almost any direction. In contrast, the upper face iscontrolled by fewer muscles that can only be moved medially.-19-These behavioral differences between the upper and lower faceare the result of differences in neural innervation of the tworegions (Rinn, 1984).The available evidence suggests that the motor pathwaysresponsible for the mediation of voluntary facial expressionsdiffer from those responsible for involuntary, spontaneouslyinduced, facial expressions (Rinn, 1984). Volitionalexpressions are more strongly expressed on the left side of theface, whereas spontaneous expressions are either symmetrical orlateralized depending on affective valence (Sackeim & Gur,1983). The neuroanatomical distinction between voluntary andspontaneous facial expressions is a well-established principleof clinical neurology. The control of voluntary expressions isbelieved to be via the corticobulbar pathway beginning in themotor cortex. However, different pathways have been proposed tomediate spontaneous expression. Rinn (1984) suggests thatspontaneous facial expressions arise from a phylogeneticallyolder motor system known as the extrapyramidal motor system.This system consists of a group of highly interactive neuralcircuits that involve primarily subcortical nuclei but thatalso include some cells of the frontal and prefrontal cortex.In recent years, researchers have increasingly used facialactivity to differentiate among emotions. There is asubstantial body of literature concerning the relation betweenfacial expression and inferred or reported emotional feelings.Results from these studies have found that distinctive facial-20-expressions correspond to different feelings and thatintensities of expressions and feelings are correlated with oneanother.Two methodological approaches have been used to examinethe relation between facial expression and the subjectiveexperience of emotion. One method is to manipulate facialexpressions experimentally and to assess changes in feelings.This approach provides a test of the facial-feedback hypothesisof emotional feelings, which states that facial expressionsprovide feedback to the expresser, which in turn affectsemotional experience (Izard, 1977; Laird, 1974, 1984; Tomkins,1962a; Zajonc, 1984, 1985; see recent criticism by Matsumoto,1987). For example, according to this view, frowning would beexpected to make people feel more angry than sad or afraid.Some authors believe that facial muscle activity may affectbrain functioning via efferent feedback loops (Izard, 1977;Tomkins, 1962b) or by altering the blood flow to the brain(Zajonc, 1985).Several studies have examined the effects of facialalterations on emotions. The primary methodological issue inthese studies is how to manipulate the facial expressionwithout alerting subjects to their emotional meanings. Mostresearchers have used elaborate cover stories to disguise thetrue purpose of their studies. Extensive postexperimentalquestionnaires are used to exclude subjects who are aware ofwhat the study was about. The majority of studies have reported-21-findings consistent with the hypothesis that facial muscleactivity contributes to the experience of emotion (Duclos etal., 1989; Fridlund & Izard, 1983; Kappas, McHugo, & Lanzetta,1989; Laird, 1974, 1984).The other approach to demonstrating an association betweenfacial expression and feelings is to elicit emotional feelingsand to measure facial expressions and subjective reports.Schwartz and others (Cacioppo & Petty, 1981; Fridlund,Schwartz, & Fowler, 1984; Schwartz, Weinberger, & Singer, 1981;Voglmaier & Hakerem, 1989; see Fridlund & Izard, 1983, for acomprehensive review of the facial-EMG and emotion research)have recorded EMG activity in paradigms in which subjects arerequired to imagine different emotional scenes. These studieshave consistently reported an increase in the EMG activity ofthe zygomatic muscle (elevates the cheeks to a smile) duringthe imagination of happy scenes and an increase in EMG activityof the corrugator muscle (pulls the eyebrows together whenfrowning) during sad or unpleasant scenes. The use of EMGenables the detection of changes that are so small that theymay be missed by, or invisible to, an individual observing thefacial activity.In a recent study, Friedlund et al. (1984) appliedmultivariate pattern-classification strategies to facial EMGdata. The EMG data reliably discriminated among four emotions:happiness, sadness, anger, and fear. Subsequent studies byother researchers (Friedman & Thayer, 1989; Uijtdehaage &-22-Thayer, 1988) have also reported the ability of thisstatistical procedure to differentiate among emotionalresponses.Using a different paradigm, Dimberg (1982) measured facialEMG activity when subjects were shown faces with a happy or anangry expression. Happy faces elicited an increase in zygomaticactivity, whereas angry faces evoked an increase in corrugatoractivity. Dimberg (1986) obtained similar results with fear-relevant (snakes/spiders) and fear-irrelevant(mushrooms/flowers) slides.Ekman, Friesen, & Ancoli (1980) conducted a study in whichfacial activity was covertly videotaped while subjects viewed apositive (designed to elicit happiness), negative (designed toelicit disgust) and a neutral film. Subjects rated theiremotional reactions on a series of scales for interest, anger,disgust, fear, happiness, pain, sadness, surprise, and arousal.The Facial Action Coding System (FAGS; Ekman & Friesen, 1978)was used to measure facial activity during the film clips.Compared with the other clips, the positive clips wereassociated with greater reports of happiness and increasedactivity of the zygomatic major muscle. Facial activity(levator labii superioris) to the negative films correlatedwith reported feelings of disgust but not with reports of othernegative emotions.The above review suggests that facial activity reliablyreflects emotional reactions. There is some debate about the-23-exact number of emotions that can be reliably distinguishedusing facial expression, but most theorists have suggested thatthe number is less than 10 (e.g., Ekman, Friesen, & Ellsworth,1972; Izard, 1971). In any case, it is apparent that measuringfacial activity has the potential for providing us with newinsights into the affective processes of psychopaths.Autonomic Activity. The use of autonomic activity tomeasure emotion is complicated by the as yet unresolved issueof whether or not different emotions are in fact associatedwith distinctive patterns of ANS activity. On one side of thedebate are theorists who, following the tradition of Cannon(1927), claim that the same general and undifferentiatedphysiological arousal is part of all emotions, and that theemotion experienced by an individual is dependent on hiscognitions about the causes of the arousal (Mandler, 1975;Schachter, 1975). Arousal determines only the intensity ofemotional states, while cognition determines their quality. Onthe other side of the debate are those who, following in thetradition of Darwin (1872) and James (1884), assert that thereare emotion-specific physiological patterns (e.g. Izard, 1977;Leventhal & Tomarken, 1986; Plutchik, 1980).There are a number of difficult methodological problemsinvolved in studying ANS specificity (see review by Levenson,1988). These include difficulties in obtaining a reasonablebaseline to compare emotion-related ANS changes, failure to-24-verify the presence of the intended emotion, and difficultiesin obtaining physiological measurements before and after theemotion is felt by the subject. These problems notwithstanding,studies in the 1950s provided some initial evidence ofautonomic differentiation between fear and anger (Ax, 1953;Schachter, 1957). More recent investigations have providedadditional empirical support for the view that various emotionsare associated with specific patterns of ANS activity (Ekman,Levenson, & Friesen, 1983; Levenson, Ekman, & Friesen, 1990;Roberts & Weerts, 1982; Schwartz et al., 1981, 1983; Stemmler,1989; Uijtdehaage & Thayer, 1988). For example, Ekman and hiscolleagues (Ekman et al., 1983; Levenson et al., 1990) havestudied six different emotions in two different experimentaltasks: A posed facial expression task and an imagery (relivedemotions) task. The results from these studies indicate thatautonomic activity differentiated between positive and negativeemotions and among negative emotions.Further support for autonomic-specificity of emotionscomes from several recent studies investigating the autonomicchanges people report to have experienced during variousemotional states. In these studies, subjects are asked toremember in detail an instance of a given emotion. They arethen asked to describe any physiological changes associatedwith this experience. Clearly differentiated patterns ofphysiological changes are typically reported for each specificemotion (Pennebaker, 1982; Shaver, Schwartz, Kirson, &-25-O'Connor, 1987).Like discrete models, dimensional approaches havedemonstrated that the pattern of physiological activitycovaries with both affective valence and arousal. There isevidence that phasic heart rate change is associated with theaffective valence of pictorial stimuli (Sartory, 1983), butthat the direction and magnitude of heart rate change isdependent on characteristics of the subjects (Hare, 1972;Klorman, Weissberg, & Wiesenfeld, 1977). In normal subjectsheart rate acceleration occurs in response to pleasant slides(e.g., landscapes, nudes, children) and heart rate decelerationoccurs in response to unpleasant slides (e.g., starvingchildren, mutilated bodies; Eisenberg et al., 1988; Winton,Putnam, & Krauss, 1984). In contrast to heart rate, skinconductance activity (neuoranatomical innervation is entirelysympathetic) primarily reflects arousal. Skin conductanceactivity is also related to stimulus intensity, regardless ofaffective valence (Greenwald, Cook, & Lang, 1989; Manning &Melchiori, 1974).Greenwald et al. (1989) used a dimensional analysis ofemotion to examine the relation between subjective judgmentsand physiological responses. Heart rate and electrodermalresponses were recorded as subjects viewed slides that variedacross the dimensions of affective valence and arousal. Usingregression analyses heart rate acceleration was specific toratings of valence, whereas SCRs were specific to ratings of-26-arousal. This pattern of association between subjective ratingsof valence/arousal and physiological activity has recently beenreplicated using a different set of affective stimuli(Fitzgibbons & Simons, 1990).The present study measured skin conductance and heart rateactivity during the processing of affective stimuli. In thefirst experiment the pattern of phasic cardiac andelectrodermal responses to affective stimuli that vary invalence and arousal was examined. In order to determine whetheror not psychopaths are characterized by differences inautonomic specificity across emotions, a second experimentmeasured autonomic activity during the induction of sixdiscrete emotions.Cerebral Asymmetry and EmotionA growing body of literature indicates that the twocerebral hemispheres are differentially lateralized for theexpression of certain positive and negative emotions (seereviews by Davidson, 1984; Kinsbourne, 1978; Leventhal &Tomarken, 1986; Tucker & Frederick, 1989). The evidence onhemispheric asymmetry and emotion comes from a variety ofsources, including brain electrical activity and regional bloodflow studies on normals in response to affective stimuli(Davidson, 1987; Davidson, Ekman, Saron, Senulis, & Friesen,1990; Ekman, Davidson, & Friesen, 1990), unilateral brain-damaged subjects (Kolb & Milner, 1981; Robinson, Kubos, Starr,Rao, & Price, 1984), unilateral injections of sodium amytal in-27-patients prior to neurosurgery (Lee, Loring, Meador, &Flanigin, 1988), and from the study of clinically depressedpopulations (Davidson, Schaffer, & Saron, 1985). These studiesindicate that frontal EEG asymmetries differentiate betweenpositive and negative emotions; left-sided activation occursduring positive affect and right-sided activation duringnegative affect. In a recent series of studies, Davidson andhis colleagues have extended the evidence for differentialasymmetries to newborn and 10-month-old infants (Davidson &Fox, 1982; Fox & Davidson, 1988).On the basis of the above evidence, Davidson (1984, 1987)has proposed a developmental model of the hemisphericsubstrates of emotion. This model postulates that the leftanterior region is associated with an approach system, and thatthe right anterior region is associated with a withdrawalsystem. It is further proposed that approach and withdrawalcomponents are differentially associated with differentemotions. For example, fear and disgust have been associatedwith behavioral components of withdrawal (Ekman & Friesen,1978; Plutchik, 1980), whereas happiness has been associatedwith approach. Some emotions (e.g. sadness) involve bothapproach and withdrawal components (Fox & Davidson, 1988).There is some evidence that psychopaths differ from othersin their cerebral organization of language (see review by Hareet al., 1988). Results from dichotic and tachistoscope studiesof psychopaths suggest that their left hemisphere is not as-28-specialized for language processing as it is for the majorityof people (Hare & Jutai, 1987; Hare & McPherson, 1984; Jutai,Hare, & Connolly, 1987; Raine, O'Brien, Smiley, Scerbo, & Chan,1990). Results from a recent lateralization study of affectiveprocessing (Day, 1991), have been interpreted as indicatingthat psychopathy is associated with a right hemisphereimpairment.To date no study has used brain wave activity to examinethe laterality of affective reactions in psychopaths. In thisstudy, hemispheric differences between psychopaths andnonpsychopaths in brain wave activity were measured during theelicitation of different emotional states.Purpose of the Current StudyThe results of several studies raise the possibility thatpsychopaths are characterized by poor integration of thereferential and affective components of language. This"deficit" appears to be related to an insensitivity toemotional valence. Although most of the research on affectiveprocessing in psychopaths has involved linguistic stimuli, thepossibility exists that they suffer from a more generalizedaffective deficit.The present study is an initial step in the systematicinvestigation of affective reactions in psychopaths. Both adimensional and a discrete model approach were used to studyaffective functioning. For the former, emotion was assessed in-29-both nonpsychopathic and psychopathic subjects by sampling allthree components of emotion--physiological, subjective,behavioral--during and subsequent to viewing of colour-slidestimuli that varied across the dimensions of valence andarousal. Physiological measures included skin conductance,heart rate, facial muscle activity, and brain wave activity,while subjective measures consisted of direct subject ratingsof valence and arousal. A measure of behavior was obtained byhaving subjects engage in a recognition reaction time task.Data showing that psychopaths give small electrodermalresponse during anticipation of a noxious stimulus and are poorpassive avoidance learners have been used to support thehypothesis that psychopathy is related to a specific deficit inneurophysiological systems modulating fear behavior (Fowles,1980; Gray, 1975). In order to study psychopaths' emotionalreactions across different negative emotional states a discretemodel approach was used. A series of film clips was used toelicit various negative and positive emotional states: fear,sadness, anger, disgust, happiness, and amusement. Psychopathsand nonpsychopaths were compared on the pattern ofphysiological, facial, and self-report responses associatedwith each emotional state.As noted above, the study also provided data on the issueof cerebral organization (lateralization) of affect inpsychopaths.-30-Experimental HypothesesClinical observation of speech and behavior, as well assome recent empirical research, suggest that psychopathsexhibit abnormalities in the processing of affectiveinformation. However, the extent to which these abnormalitiesare related to the emotional valence (positive or negative) isuncertain. For this reason, the present study evaluatesresponsivity to pleasant (positive affect) and unpleasant(negative affect) slides and films.The ability of psychopaths to use and manipulate otherssuggests that there may be a desynchrony between verbalizedreports of their affective experience and the associatedphysiological activity. Both subjective reports and indices ofphysiological variables were therefore obtained from eachsubject as he viewed the slides and films.The following predictions were tested:1. Assuming that psychopaths are insensitive to emotionalvalence, their heart rate responses (which in normalindividuals covary with affective valence) will be the sameto the positive slides and films as they are to the negativeslides and films.2. Skin conductance responsivity is positively correlated witharousal level in normal individuals. The same pattern ispredicted for psychopaths. That is, the SCRs of both-31-psychopaths and nonpsychopaths will be positively associatedwith the reported intensity of arousal in response to theslides and films.3. Results from several studies have found that facialexpressions and reports of feelings correlate with oneanother. In experiment 1, psychopaths will show littlevariation in facial activity across affective and neutralslides. In experiment 2, if psychopaths have an affectivedeficit specific to aversive material, than differencesbetween psychopaths and nonpsychopaths will be limited tothe negatively-toned films. In response to these films,psychopaths will display fewer facial expressions ofnegative emotions than will nonpsychopaths.4. Research with normal populations indicates that positiveemotions differentially activate the left frontal cortex,whereas negative emotions differentially activate the rightfrontal cortex. It is predicted that this lateralizedeffect--in response to slides and film clips--will occur innonpsychopaths but not in psychopaths.Experiment 1To begin the investigation of emotion-processing inpsychopathy, a dimensional conceptualization of emotiondescribed by Lang (1984, 1985) was adopted for the present-32-study. According to this view emotions are action dispositionsthat can be characterized in terms of two orthogonaldimensions: affective valence and arousal. Valence refers tothe directionality of the elicited action disposition, varyingfrom avoidance to approach; arousal refers to the intensity ofactivation of the response. The following paradigm was modelledafter research conducted by Peter Lang and his colleagues whohave reported strong and specific relationships betweenphysiological measures and verbal reports of valence or arousalwith college samples (Greenwald et al., 1989). In thisexperiment facial expressions, central and peripheralphysiological activity, and subjective ratings of affectivevalence and arousal were measured during exposure to a seriesof pleasant, neutral, and unpleasant slides. A measure ofbehavioral response was obtained by having subjects engage in arecognition reaction time task.MethodSubjects Subjects were 42 male inmates incarcerated at a mediumsecurity federal institution in British Columbia who hadvolunteered to participate in ongoing research projects.Psychopathy was assessed with the the PCL-R (Hare, 1991), usingextensive interview and file information. The interviewer whoperformed the initial PCL-R assessments recruited subjects to-33-participate in the present research excluding subjects whosuffered from seizures, hallucinations, or those who hadexperienced a loss of consciousness due to a head injury.The PCL-R is a 20 item symptom checklist that measures thepersonality traits and behaviors characteristic of thepsychopath. Each item is scored on a 3-point scale according tothe extent to which it applies to the inmate. The individualitems are then summed to yield a total score. The total scorecan range from 0 to 40 and represents the degree to which anindividual resembles the prototypical psychopath. There is alarge body of evidence attesting to the reliability andvalidity of the PCL-R. The interrater reliability and alphacoefficients for total scales are usually above .80 (Hare,1991). To assess the reliability of the ratings in the presentsample, two raters (not the experimenter) completed the PCL-Ron a subsample of 37 inmates. The intraclass correlationcoefficient of interrater reliability for the averaged ratings(Bartko, 1976) was .92.The mean PCL-R total scores for the sample was 27.0 with astandard deviation of 4.6; the mean was about 3 points higher,and the standard deviation about 2 points lower, than thoseobtained with much larger samples of male offenders (Hare,1991). Previous research (Hare, 1991) in forensic populationshas used a cutoff score of 30 or above on the PCL-R for acategorical diagnosis of psychopathy. Using this criterion, thesubjects in the sample were divided into 21 nonpsychopaths-34-(PCL-R score less than 30; range 5-27) and 21 psychopaths (PCL-R 30 or greater; range 30.5-40). The mean PCL-R total score forthe nonpsychopaths and psychopaths was 20.86 (SD = 5.90) and33.24 (SD = 2.62), respectively. Nonpsychopaths and psychopathsdid not differ significantly in age (M = 26.29 years and 26.81years, respectively) or in formal education (M = 9.14 years and8.90 years, respectively).The hand preference of each subject was assessed with theAnnett Handedness Questionnaire (Annett, 1970).Stimulus Materials Affect-relevant stimuli were colour photographs selectedfrom the International Affective Picture System (IAPS; Lang,Ohman, & Vaitl, 1988). These slides have previously been shownto vary across dimensions of valence and arousal in normalsubjects (Bradley et al., 1990). Each subject viewed 36 colourslides depicting pleasant (e.g., smiling face, childrenplaying, food), unpleasant (e.g., sad face, human injury,starving child, aimed gun), and neutral (e.g., neutral face,household objects) objects or scenes.Apparatus and Response MeasurementThe presentation and collection of stimuli were controlledon-line by a Compaq 286 micro-computer. The Affect Grid(Russell, Weiss, & Mendelson, 1989) was used to assess twodimensions of affect: pleasure-displeasure and arousal--35-sleepiness. This scale was chosen over other measures of affectbecause it is short and easy to fill out. The Affect Grid hasbeen shown to have adequate reliability and validity in theassessment of affect (Russell et al., 1989).Slides were presented for 5 s using a Kodak projector withan electronic shutter. A 12-channel Beckman polygraph was usedto record EEG, electrooculographic (EOG), electrodermal,cardiac, and respiration activity as the subject viewed theslides.EEG activity was recorded from the left and right frontaland parietal regions (F3, F4, P3, P4) all referenced to thevertex (Cz). Subjects wore a stretchable lycra cap (Electro-Cap) that contained all of the 10/20 electrode placements. Eachof the four EEG leads was directed into its separate Beckmantype 9806A A.C. coupler with low-pass band filters set at 30Hz. The EEG and EOG channels were calibrated before eachtesting session by inserting a 50-3IV sine wave into thepreamplifier for each channel. This signal was digitized andeach channel was scaled by the calibration signal prior to dataanalyses. To ensure low impedances all electrode sites wereabraded lightly with a blunt needle. Electrical impedances atall sites were recorded before and at the end of theexperiment. All were well below 5K ohms.EOG was recorded via Beckman miniature electrodes placedon the supra-orbital ridge and the outer canthus of the righteye; the couplers and time constant were the same as those used-36-for EEG. EEG and EOG were digitized on-line using a micro-computer (sampling rate of 128 samples per second). All datawere stored for later off-line analysis.Skin conductance (SC) was recorded using Beckman standardAg/AgC1 cup electrodes filled with .05 molar NaCl electrodepaste (Lykken & Venables, 1971). The electrodes were attachedto the palmar side of the medial phalanges of the first andsecond fingers of the subject's non-dominant hand using double-sided adhesive collars that permit a contact area of 1 cm indiameter. A Beckman Type 9844 skin conductance coupler provideda constant 0.5 volt potential across the electrodes.Heart rate was measured using Ag/AgC1 disposableelectrodes attached to the sternum (at the 4th rib) and at thelateral margin of the chest. Electrode sites were first abradedwith Redux paste and the signal was fed into a Beckman Type9806A coupler.Thoracic respiration was obtained by a pneumatic tubepositioned around the thorax, which was attached to a GrassPT5A Volumetric Pressure Transducer. The signal was then fedinto a Beckman 9853A Voltage/Pulse/Pressure coupler.EEG and EOG were recorded from 2 s prior to slide onset(to establish a baseline value), throughout the slide-viewingperiod, and for 2 s post-slide. Autonomic nervous systemactivity was recorded 5 s prior to slide onset, throughout theslide-viewing period, and for 10 s post-slide offset.A Sony Super-VHS Camcorder, located behind a black-37-curtain, recorded facial activity continously during alltrials.ProcedureAll subjects were tested individually by a femaleexperimenter in a quiet research room at the institution. Thesubject was paid $8.00 for his participation in a singlesession lasting approximately 2 hours.When the subject arrived he was told that some of thematerial he would view would be unpleasant. Any subject wishingto withdraw from the experiment was given the opportunity to doso at this time. None of the subjects withdrew. The subject wastold that his physiological and subjective reactions to bothpleasant and unpleasant slides would be recorded, and he wasthen asked to sign two consent forms, one describing thepresent study and the other giving permission for portions ofthe study to be videotaped. He was told that videotaping wasnecessary in order to ensure that the electrodes did not falloff and to check for movement artifact when analyzing the EEG.During the experiment the subject's facial activity wasvideotaped.In order to control for the effect of mood and anxiety onemotional responding, the subject was asked to complete theBeck Depression Inventory (BDI) and the State form of theState-Trait Anxiety Inventory (STAI-S; Spielberger, Gorsuch, &Lushene, 1970) while the experimenter checked the physiological-38-recordings. Baseline recordings were made during 2-minuteperiods immediately preceding and following presentation of theslides.The slides were projected onto a screen 2.5 m away fromthe subject. Prior to beginning the experiment, each subjectwas shown how to use the Affect Grid. He was instructed tofocus his attention on the slide for the entire time it wasdisplayed and to relax between the trials.On each trial, the subject viewed the slide for 5 s.Immediately after the slide offset, he reported (with theAffect Grid) his subjective response (valence, intensity) tothe slide. A relatively long intertrial interval, rangingrandomly from 15 to 25 s, permitted autonomic responses torecover and the Affect Grid to be completed.During the experiment proper, the experimenter remainedout of the subject's sight, behind a screen that separated therecording equipment from the subject.After the post-slide baseline period, the experimenterreturned and administered the Affect Intensity Measure and theInterpersonal Reactivity Inventory (described below). Thesubject then performed an incidental memory task and alabelling task. He viewed a set of 72 slides (36 slides that hehad seen previously mixed with 36 new slides), and was asked todecide as quickly and accurately as possible, after each slide,whether or not he had previously seen the slide. He used hisdominant hand to press one of two buttons: the right button if-39-the slide was from the series he previously had seen, and theleft button if he did not recognize the slide. Reaction time(RT) was recorded in msec after slide onset. The subject'sbutton press removed the slide from the screen. After eachslide the subject categorized the slide according to which ofseven emotions (fearful, disgusted, sad, angry, happy,surprised, neutral) he felt it most closely evoked. Theintertrial interval varied randomly between 10 and 20 s. Thesubject than had a 10 min break prior to beginning the secondexperiment.Self-report Ouestionnaires The Affect Intensity Measure (AIM: Larsen, Diener, &Emmons, 1986) is a 40-item self-report questionnaire thatmeasures the strength of affective reactions to lifesituations. Larsen et al. (1986) have proposed that affectintensity is a stable individual difference dimension, withadequate test-retest stability (.75 over 2 years). In addition,Larsen and his colleagues have shown that scores on the AIM arenot contaminated by response bias, and have presented evidencefrom several sources documenting the validity of the AIM as ameasure of individual differences in the intensity of affect(for a review, see Larsen & Diener, 1987).The Interpersonal Reactivity Index (IRI; Davis, 1983) is a28-item self-report questionnaire that consists of foursubscales, each of which assesses a specific aspect of empathy.These subscales are Perspective-Taking, a measure of the-40-tendency to adopt the point of view of other people; theEmpathic Concern scale, a measure of the tendency to experiencefeelings of warmth, compassion, and concern for others; theFantasy scale, a measure of the tendency to imagine oneself asa character in fictitious settings; and the Personal Distressscale, a measure of feelings of unease and discomfort in tenseinterpersonal settings.Data Analysis Skin conductance responses (SCRs) were hand-scored anddefined as the largest increase occurring between 1 and 6 safter slide onset. They were subjected to a log transform(log[SCR+1]) to eliminate positive skew in their distribution.The mean SCR for the unpleasant, neutral, and pleasant slideswas calculated by averaging across the 12 trials.Heart-rate data, in beats per minute, were hand-scored foreach stimulus presentation. For data analysis purposes, heartrate activity was defined as change during the 5 s slideviewing period from the 5 s baseline immediately precedingslide onset. The mean heart rate change was calculated byaveraging across the 12 unpleasant, neutral, and pleasanttrials.EEG data were screened off-line for eye movement artifact.Trials contaminated by artifact on any channel were excludedfrom data analysis. EEG was aggregated from all artifact-freeperiods to form one set of data for each of the slide types-41-(pleasant, unpleasant, and neutral).Two s chunks of EEG data were extracted using a Hammingwindow and were overlapped by 50% to capture data at the tailsof the chunks. A Fast Fourier Transform (FFT) was applied toeach chunk of EEG. The FFT is an algorithm that decomposes anycomplex waveform into its underlying sine wave components. TheFFT permitted the computation of the amount of power atdifferent frequencies in the EEG. Power values from all thechunks across the slide types were averaged.In the present study, the dependent measure selected forsubsequent analysis was the power density (in pV2 /Hz) in thealpha (8-13 Hz) band. Previous studies have reported that powerin this band has been most consistently linked to EEGasymmetries associated with concurrent emotional state (e.g.,Davidson et al., 1990a) and with individual differences inaffective predispositions (Tomarken, Davidson, & Henriques,1990). Research indicates that decreased power in the alphaband in a given region is associated with increased corticalactivation in that region (e.g., Lindsley & Wicke, 1974; forreview see Davidson, Chapman, Chapman, & Henriques, 1990b). TheEEG data were log-transformed to normalize their distributionsbecause the power values were positively skewed.Facial activity was measured using the Facial ActionCoding System (FACS; Ekman and Friesen, 1978). FACS is ananatomically based, comprehensive, and objective coding systemfor measuring all visible facial movements. Elemental facial-42-muscles are categorized under what are called "action units,"each of which has a number code. FACS allows a trained scorerto code all facial movements without making any inferencesabout underlying emotional states or using descriptions thatmix inferences and description, such as smile or frown.Intercoder reliability for these scoring procedures hasbeen established in a number of laboratories and studies (cf.Ekman, Friesen, & O'Sullivan, 1988; Fox & Davidson, 1988;Krause, Steimer, Sanger-Alt, & Wagner, 1989). All facialactivity was scored by the author who had previously beentrained using FACS and whose reliability had been establishedagainst a standard criterion (Ekman's and Friesen's ownscoring).To check reliability, a second coder with extensiveexperience in using FACS and blind to the purpose of the studyrandomly scored 12 of the subjects. Interscorer reliability wasevaluated by using a ratio in which the number of facialactions on which the two coders agreed was multiplied by twoand then divided by the total number of facial actions scoredby the two coders (Ekman et al., 1980). Agreements betweencoders that no scorable behavior occured were not included inthe ratio. The mean ratio of agreement across all scored eventswas .73, which is comparable to the level of reliabilityreported in the FACS manual (Ekman & Friesen, 1978). A separateratio was calculated for agreement between scorers that "noscorable" facial actions had occurred. The mean ratio for no-43-scorable behavior was .91.The overall number of facial movements (action units) wascalculated for each subject and summed for each group acrosseach of the slide types. In addition, the frequencies of upperand lower face action units were separately calculated.ResultsSelf-Report Measures Data analysis began with the calculation of group meansfor the valence and arousal ratings for each slide. Valencemeans for the total sample were then ranked from most pleasantto least pleasant, and these values were used to divide the 36-slide set into 12 pleasant, 12 neutral, and 12 unpleasantslides. Table 1 presents the listing of the slide content,ranked from most pleasant to least pleasant, for psychopathsand nonpsychopaths. For comparison purposes, rankings of thesame slides by male college students are included; theserankings were calculated from ratings obtained from Lang et al.(1988). The two offender groups and the students produced verysimilar rankings of the slides. The same process was used torank and group the slides on arousal and these values were usedto group the slides into 12 low, 12 medium, and 12 higharousing slides (see Table 2). The offenders and studentsamples were very similar in their ratings of the pleasantslides. However, several of the unpleasant slides (e.g., baby-44-Table 1Individual slides, rank-ordered by affective valencefrom most pleasant to least pleasant averaged acrosssubjects from each groupContentpuppiesbunniesfemale on beachsailingman with childrenkissing couplenude femalenude femaleman with babysmiling babyturkey dinnersunset at beachagateman climbing rockbuildingflowermale faceelectric outletmushroombeesrolling pinforkumbrellaneutral faceangry facebarbed wiretoxic cleanupwar victimaimed gunsick mandead cowmutilated bodymafia hitmutilated bodystarving childbaby with tumorSlideOrderSlideNo.27 17114 1753 42215 80817 23428 46621 43034 4312 21520 2048 72331 5837 7821 81625 75032 50233 2226 61513 55036 13919 70024 70829 71510 2204 21223 90135 92712 92511 62316 32226 9145 31222 3019 30018 90430 317MalelStudents NP P1 1 32 2 23 4 94 8 65 5 116 10 87 7 108 11 79 3 410 6 511 13 1212 9 113 15 1514 12 1415 19 2116 14 1317 24 2318 18 1819 16 1620 22 2421 23 1722 20 2023 17 1924 21 2525 27 2626 29 2827 26 2728 30 3129 25 2230 28 2831 31 3032 32 3333 33 3234 35 3435 34 3536 36 36Note. NP = nonpsychopaths; P = psychopaths1 = Taken from norms provided by Lang et al. (1988)-45-Table 2Individual slides rank-ordered by arousal level frommost arousing to least arousing averaged across subjectsfrom each groupSlide SlideOrder^No.^Content^21^430 nude female15 808^sailing22^301 mafia hit11 623^aimed gun28^466 kissing couple1 816 man climbing rock34^431 nude female30 317 baby with tumor3^422 female on beach9 300 mutilated body12^925 war victim5 312 mutilated body36^139^bees8 723 turkey dinner35^927 toxic cleanup27 171 puppies17^234 man with children18 904 starving child16^322^sick man33 222 male face4^212 angry face31 583 sunset at beach26^914 dead cow2 215 man with baby7^782 agate20 204 smiling baby23^901 barbed wire14 175 bunnies6^615 electric outlet25 750 building13^550 mushroom10 220 neutral face24^708^fork32 502^flower19^700 rolling pin29 715 umbrellaMalelStudents NP P1 1 12 2 63 8 84 6 55 5 46 7 77 3 28 33 299 4 310 17 1411 14 1712 29 1513 9 2014 16 1015 15 1316 22 1617 10 1118 35 3619 32 2620 12 1221 11 922 25 2423 34 2824 23 1825 18 2726 13 2227 21 2128 36 3229 26 3430 19 2531 20 1932 27 2333 24 3134 31 3535 28 3036 30 33Note. NP = nonpsychopaths; P = psychopaths1 = Taken from norms provided by Lang et al. (1988)-46-with tumor, starving child) were given low arousal ratings bythe offender groups, whereas the student group rated them asbeing high in arousal. The offenders' low arousal ratings forcertain unpleasant slides may reflect a higher threshold ofarousability for these types of stimuli.The data for the valence and arousal ratings were analyzedin a 2 (group: nonpsychopath or psychopath) X 3 (slide type:pleasant, neutral, or unpleasant) multivariate analysis ofvariance (MANOVA) with group as a between-subjects factor andslide type as a within-subject repeated measures factor. Theonly multivariate main effect to reach significance was forslide type (F (2,39) = 121.91, p < .001). Univariate analysisof variance (ANOVA) tests on the two dependent variablesindicated that the effects for both valence and arousal ratingswere significant. Valence and arousal ratings for each of theaffective slide types for each group are presented in Figure 1.There was a significant main effect for slide type (F (2,39) =124.71, p < .001). Not surprisingly, there was a highlysignificant linear relation between valence ratings and theirrespective slide groups (F (1,40) = 250.37, p < .001), withpleasant slides being rated the most pleasant and unpleasantslides the most unpleasant. The psychopathic andnonpsychopathic groups did not differ in their ratings of thevalence of the slides.A univariate ANOVA performed on the arousal ratingsrevealed that pleasant slides were reported as more arousing2354768-47 -Pleasant^Neutral^UnpleasantSlide Valence CategoryFigure 1. Self-ratings of valence (upper panel) and arousal(lower panel) for pleasant, neutral, and unpleasantslides, by diagnostic group.-48-than unpleasant slides (linear, F (1,40) = 35.29 p < .001), andthat both the pleasant and unpleasant slides had higher arousalratings than did the neutral slides (quadratic, F (1,40) =20.22, p < .001). The same general pattern of results wasobtained for both psychopaths and nonpsychopaths.The data for the arousal and valence ratings were alsoanalyzed in a 2 (group) X 3 (slide arousal level; low,medium,or high) MANOVA. There was a significant main effect forslide arousal level (F (2,39) = 15.73, p < .001). UnivariateANOVAs on the two dependent variables yielded significanteffects for both valence and arousal ratings. The expectedsignificant linear relation between arousal ratings and therespective slide groups was obtained (linear, F (1,40) =116.61, p < .001), but the quadratic trend was also significant(quadratic, F(1,40) = 19.54, p < .001). Arousal and valenceratings for each group across the slide arousal levels arepresented in Figure 2. The psychopaths and nonpsychopaths didnot differ in their ratings of slide arousal.A univariate ANOVA indicated that high arousal slides wererated as more pleasant than either neutral slides or lowarousal slides (linear, F (1,40) = 49.01, p < .001). There wereno significant differences between groups.Correlations between mean valence and arousal ratings werecomputed for the entire sample, and then separately for eachgroup to establish whether the rating dimensions wereindependent. The valence and arousal dimensions for this slide-49-Low^Medium^HighSlide Arousal LevelFigure 2. Self-ratings of valence (upper panel) and arousal(lower panel) for low, medium, and high arousalslides, by diagnostic group.-50-set were relatively independent for both nonpsychopaths (r = -.05, n.s.) and psychopaths (r = -.23, n.s.). The correlation inthe overall sample was -.20 (n.s.), with greater pleasuretending to be associated with higher arousal.Skin Conductance There were no group differences in tonic SC level at anypoint in the experiment. There was a significant main effectfor slide arousal level (F (2,39) = 16.46, p < .001). Aspredicted, SCR magnitude covaried significantly with slidearousal, with high arousal slides being associated with thelargest SCRs. Consistent with findings for college students,there was a strong linear effect (linear, F (1,40) = 22.06, p <.001), but the quadratic effect was also significant(quadratic, F(1,40) = 12.29, p < .005). Figure 3 depicts thisrelationship and shows that there were no group differences inSCR to slides differing on arousal levels. This pattern ofresults replicates the pattern which has been observed inseveral previous studies with normal subjects (Greenwald etal., 1989; Lang et al., 1990).Slide valence had a significant effect on the amplitude ofthe SCR F(2,39) = 8.08, p < .005). As Figure 4 shows, SCRs werelarger to unpleasant and pleasant slides than to neutralslides, quadratic trend (F(1,40) = 16.15, p < .001). The groupX slide valence interaction approached significance (F(2,39) =3.07, p = .07), with psychopaths tending to give larger SCRs0.230.22 • Nonpsychopaths■ Psychopathsc 0.18 -a)20.17 -0.160.15Low^Medium^HighSlide Arousal LevelFigure 3. Mean log-transformed skin conductance responsemagnitude to low, medium, and high arousal slides,by diagnostic group.Slide ValenceFigure 4. Mean log-transformed skin conductance responsemagnitude to pleasant, neutral, and unpleasantslides, by diagnostic group.-53-than those given by the nonpsychopaths.Examination of the SCRs of individual subjects revealedthat eight subjects displayed very little skin conductanceactivity across the entire sequence of slides. Of thesesubjects four were nonpsychopaths and four were psychopaths.Because the nonreactivity of these subjects may haveaffected the group means, the data were reanalyzed withoutthese subjects; the results were unchanged.Heart RateDue to a failure of the heart rate coupler, no heart ratedata were available for one subject from the psychopathicgroup. There were no group differences in tonic heart rate atany point in the experiment. Mean heart rate during the 2-minpre-slide baseline period was 73.46 (SD = 9.03) for thepsychopaths and 73.96 (SD = 8.86) for the nonpsychopaths.Mean heart rate change scores as a function of affectiveslide types are presented in Figure 5. Heart rate change scoreswere significantly related to the emotional valence of theslide, F(2,38) = 8.47, p < .005: heart rate deceleration wasgreatest when subjects viewed the unpleasant slides, andsmallest when they viewed the pleasant slides (linear, F (1,39)= 12.03, p < .005). There were no significant group differencesin heart rate responsivity as a function of slide valence.Heart rate change scores were unrelated to slide arousal(F (2,38) < 1), confirming past research that heart rate is a-3-2-10rfa Nonpsychopaths■ Psychopaths1Pleasant^Neutral^UnpleasantSlide ValenceFigure 5. Mean heart rate change (beats per minute) topleasant, neutral, and unpleasant slides, bydiagnostic group.-55-sensitive and specific measure of affective valence but notaffective arousal (Greenwald et al., 1989). There were no groupdifferences in heart rate responsivity as a function of slidearousal.In order to control for the possible confounding effectsof respiration on heart rate, the heart rate data werereanalyzed after deleting those trials in which a deep breathwas taken during the baseline period or during the period whenthe slide was on the screen. This resulted in deletion of thecomplete data from five subjects (two from the nonpsychopathicgroup and three psychopaths). Due to equipment problems, norespiration data were available for one subject in thenonpsychopathic group. Analysis of the data for the remaining18 nonpsychopaths and 17 psychopaths produced results that werevery similar to those reported above: that is, the linearrelationship between slide valence and heart rate change scoresremained significant (linear, F(1,33) = 50.22, R < .001).Facial ActivityExperimenter error resulted in the loss of data for one ofthe psychopaths. Overall, little codeable facial activity wasobserved in Experiment 1. A 2 X 3 (group X slide valence type)between-within MANOVA was performed on the upper and lowerfacial action units. The mean number of upper and lower facialaction units to the pleasant, neutral, and unpleasant slides ispresented in Table 3. The only multivariate F to reach-56-Table 3Mean (SD in brackets) number of upper and lower facial actionsfor each group as a function of slide valenceNonpsychopaths PsychopathsSlide type Lower Upper Lower UpperPleasant 2.38^(2.91) 1.67^(2.48) 4.45^(3.76) 3.55^(4.32)Neutral 1.38^(2.44) 1.57^(2.75) 2.55^(3.17) 2.95^(3.79)Unpleasant 1.43^(1.91) 3.24^(4.48) 3.40^(3.41) 4.15^(4.52)-57-significance was for slide type (F(2,38) = 9.38, p < .001).Univariate tests on each of the two dependent variablesindicated that upper facial activity was related to slidevalence, F (2,38) = 6.35, p < .005; facial activity was greaterduring presentation of both unpleasant and pleasant slides thanduring presentation of neutral slides (quadratic F(1,39) =6.70, p < .05).Slide valence also had a significant effect on lowerfacial activity, F (2,38) = 5.30, p < .01). As with upperfacial activity, both pleasant and unpleasant slides elicitedgreater facial activity than did neutral slides (quadraticF(1,39) = 8.11, p < .01). A significant main effect for groupwas obtained (F(1,39) = 4.52, p < .05. As can be seen in Table3, psychopaths displayed more lower facial activity across allslide stimuli than did nonpsychopaths.Much of the previous EMG research on facial expression hasfocused on activity in the corrugator and the zygomaticmuscles. A strong linear relationship between slide valence andcorrugator activity typically is found, with corrugatoractivity being greatest to slides with negative valence andsmallest to slides with positive valence (Bradley et al., 1990;Greenwald et al., 1989). In contrast, zygomatic activity isgreater to slides with positive valence than to slides withnegative valence. Action unit 4 (bringing the eye browstogether) reflects the activity of the corrugator muscle andAction Unit 12 (lifting the cheeks) reflects activity of the-58-zygomatic muscle.In the present study, there was a significant linearrelationship between slide valence and the frequency of ActionUnit 4 (F (1,39) = 11.84, p < .005). As expected, meanfrequency of Action Unit 4 was greatest during viewing of theunpleasant slides, and least during viewing of pleasant slides.Consistent with previous research, Action Unit 12 occurred mostoften in response to the pleasant slides (quadratic F(1,39) =11.43, p < .005), although some activity was seen in responseto unpleasant slides. The psychopathic and nonpsychopathicgroups did not differ in the frequency of Action Units 4 or 12.Electrocortical Data Because left-handers are known to differ from right-handers in functional organization of the brain (Springer &Deutsch, 1989), only subjects who were strongly right-handedwere included in these analyses. The Annet HandednessQuestionnaire asks the individual to indicate which hand he orshe prefers to use to perform a series of activities (e.g.,writing, using scissors). Scores can range from -4 (strongpreference for left hand) to +4 (strong preference for righthand). In this study, only subjects who indicated that theywrote with their right-hand and who scored +3 or higher wereconsidered strongly right-handed. Of the 42 subjects, 33(78.6%) met this criterion for right-handedness: 19 werepsychopaths and 14 were nonpsychopaths. One nonpsychopathic-59-subject was dropped due to recording problems with the EOGchannel. Two psychopathic subjects were dropped because ofexcessive noise on the parietal channel.A group (psychopath, nonpsychopath), X slide type(pleasant, neutral, unpleasant) X hemisphere (left, right) Xregion (frontal, parietal) ANOVA was performed on log-transformed alpha power. Neither the main effect for groups, F(1,28) = .87, nor any of the interactions involving groups wassignificant. The expected slide type X hemisphere interactionwas not obtained. There were no hemispheric differences inactivation in either the frontal or parietal regions across theslide types. However, there was a highly significant hemisphereX region interaction, F (1,28) = 63.17, p < .001.The means for log alpha power in the left and righthemisphere leads for the frontal and parietal regions aredisplayed in Figure 6. It is clear that the pattern of EEGasymmetry in frontal regions was opposite to that in parietalregions. To explore this interaction, separate ANOVAs wereperformed on frontal and parietal data. The ANOVA of thefrontal data revealed no significant interactions. The onlymain effect to reach significance was for hemisphere, F(1,28) =50.36, p < .001. As Figure 6 indicates, the right frontalregion was significantly more active (i.e., less alpha power)than the left.For the parietal region, the only significant effect wasfor hemisphere (F(1,28) = 400.68, p < .001). Across all slides2.54.0Pleasantei Left Hemisphere■ Right HemisphereNeutral Unpleasant-60 -Frontal RegionParietal RegionPleasant^Neutral^UnpleasantSlide ValenceFigure 6. Mean log-transformed alpha power (in iV 2 /Hz) for theleft and right frontal (F3 and F4, respectively:upper panel) and parietal (P3 and P4, respectively:lower panel) regions during the pleasant, neutral,and unpleasant slides.-61-there was significantly more left activation (i.e., less alphapower) than right activation.Recognition Reaction Time Task RT data were analyzed for only trials on which the subjectresponded correctly. Mean RT for the pleasant, unpleasant, andneutral slides was calculated for both series of slides. Thenumber of errors (subject incorrectly indicates that herecognized a slide or subject failed to correctly identify aslide) was calculated.Error rate and reaction time data were analyzed separatelyfor familiar (those slides previously seen) and distractorslides. A 2 X 3 ANOVA was performed on both measures with groupas a between-subject factor and slide type as a within-subjectsfactor. Due to equipment malfunction, data from two subjects(one psychopath and one nonpsychopath) were not recorded.Error rates and reaction times for familiar slides areshown in Table 4. Neither group made many errors. Analysis ofthe error rates revealed a significant main effect for slidetype, F(2,37) = 5.39, p < .05; the error rate for theunpleasant slides was generally lower than was the error ratefor the other slides. The group by slide type interaction wasnot significant.There were no group differences in RT to the familiarslides. RTs were not significantly related to slide type,(F(2,37) = 2.95, p = .07).-62-Table 4Mean (SD in brackets) reaction times (msec) and errrorrates for familiar slides for each group as afunction of slide valenceNonpsychopaths PsychopathsSlide type RT ER RT ERPleasant 1662 (486) .85 (1.18) 1744 (390) 1.15 (^.98)Neutral 1634 (422) .75 (1.02) 1594 (353) .75 (1.12)Unpleasant 1597 (432) .30 (^.47) 1713 (398) .55 (^.76)Note. RT = reaction time; ER = error rate-63-Very few errors were made to the distractor slides. Table5 presents the error rate data for both groups across the slidetypes. Analysis of the error rate yielded a significant effectfor the slide type (F(2, 37) = 4.52, p < .05), with both groupsbeing less accurate with the neutral slides than with thepleasant or unpleasant slides.RTs to the distractor slides was strongly related to slidetype, F(2,37) = 6.15, p < .005. For both groups RTs to thepleasant and unpleasant slides were slower than they were tothe neutral slides.To determine if both groups perceived the slides in thesame way, subjects were asked to categorize the slides asrepresenting one of seven different emotions. Prior to theexperiment the slides had been grouped (on the basis of ratingsby noncriminal subjects) as pleasant, unpleasant, or neutral.Responses were collapsed into three categories: negativeemotions, positive emotions, and neutral. A subject's responsewas considered correct if a negative emotion was chosen for theunpleasant slide, if a positive emotion was chosen for thepleasant slide, and if neutral response was chosen to theneutral slides. Table 6 presents the percentage of correctresponses for the three emotion categories. An ANOVA wasperformed on the mean number of correct responses. The onlysignificant main effect was for the emotion category F(2,36)4.42, p < .05. Tukey tests indicated that subjects weresignificantly more accurate (p < .05) when coding the pleasant=-64-Table 5Mean (SD in brackets) reaction times (msec) and errorrates for distracter slides for each group as afunction of slide valenceNonpsychopaths^PsychopathsSlide type^RT^ER^RT^ERPleasant^1672 (555)^.65 ( .99)^1749 (515)^.70 ( .80)Neutral^1609 (476) 1.25 ( .78)^1585 (392) 1.10 ( .79)Unpleasant^1770 (802)^.85 (1.27)^1931 (690)^.75 (1.16)Note. RT = reaction time; ER = error rate-65-Table 6Percentage of nonpsychopaths and psychopathswho correctly classified slidesPCL-R GroupEmotion^Nonpsychopaths PsychopathsPositive^82.0^85.6Neutral 80.2 77.9Negative^78.0^72.7-66-slides than when coding the neutral or unpleasant slides. Therewere no significant group differences in accuracy with whichthe subjects coded the slides.Self-report Ouestionnaires There were no group differences in AIM or IRI scores (seeTable 7). The mean score on the AIM (3.52) was somewhat lowerthan that reported for male noncriminal populations (3.80;Ouimette, Klein, Clark, & Margolis, 1992).None of the correlations between PCL-R scores and scoreson the IRI or the AIM was significant. There has been only oneother study that has administered the AIM to a sample ofpsychopathic and nonpsychopathic offenders (Day, 1991). InDay's study, PCL-R total scores and AIM scores were negativelycorrelated (r = -.37).There were no group differences in self-reportedpsychopathology. The mean BDI and STAI-S scores for the groupswere, respectively, as follows (SD in brackets): psychopaths10.19 (7.42) and 35.76 (6.02); nonpsychopaths 8.90 (5.65) and35.05 (4.92). In order to insure that group differences inpsychopathology did not obscure any differences in affectiveresponding, the analyses described above were repeated with BDIand STAI-S scores as covariates. No significant groupdifferences emerged from these analyses.-67-Table 7Mean scores (SD in brackets) on self-reportquestionnaires for psychopaths and nonpsychopathsPCL-R GroupScale^ Nonpsychopaths^PsychopathsIRIPerspective-taking^14.00 (3.67)^13.71 (5.14)Empathic concern^18.33 (4.00)^18.76 (4.95)Fantasy^13.33 (4.39)^14.81 (5.78)Personal distress^8.67 (3.99)^7.86 (5.00)AIM^ 3.58 ( .41)^3.48 ( .45)Note. IRI = Interpersonal Reactivity Inventory;AIM = Affect Intensity Measure.-68-DiscussionThe results from Experiment 1 lend no support to theproposition that psychopathy is associated with a generalaffective deficit. The three-system analysis of emotionindicates that psychopaths' responses to affective slides wereappropriate, given what we know about normal subjects, and wereconsistent across the self-report, physiological, andbehavioral domains. The psychopath's deficit for processinglinguistic affective material does not appear to extend tononlinguistic material.In a more general sense, the data provide support for thedimensionality approach to research on emotion. With theexception of electrocortical activity, the affective slideselicited responses from the criminals that were very similar tothe responses elicited from male college students (Bradley etal., 1990). Both the men in prison and college students rankedthe pleasantness of the slides in much the same way. Further,consistent with research on noncriminals, facial activity andheart rate in criminals were related to the valence of thestimuli, while skin conductance was related to self-reportedarousal.The lack of reliable differences in brain asymmetry as afunction of the valence of the slides was somewhat surprising;however previous failures to find reliable asymmetries havebeen reported with the use of still photographs (Davidson etal., 1985). Most of the studies that have found EEG asymmetries-69-as a function of affective valence have used much more intensestimuli. Data from the next experiment, in which films wereused to elicit emotional reactions, were expected to producecerebral asymmetries.There was one area in which the psychopaths differed fromthe nonpsychopaths: Psychopaths displayed more lower facialactivity than did nonpsychopaths across pleasant, unpleasant,and neutral slides.Experiment 2 was designed to investigate affectivereactions across a range of specific positive and negativeemotions. In an attempt to elicit stronger emotional reactionsthan those observed in Experiment 1, Experiment 2 used filmclips as stimuli.Experiment 2A considerable amount of research has looked at theresponses of psychopaths in situations involving punishment oranxiety. A consistent finding is that psychopaths showrelatively small electrodermal responses in anticipation of anaversive stimulus and fail to develop passive avoidanceresponses to unpleasant stimuli (Hare, 1978; Lykken, 1957;Schmauck, 1970). More recently, Patrick et al. (1991) foundthat psychopaths did not show facilitation of the blink startleresponse during the viewing of unpleasant slides, addingsupport to the argument that psychopathy is associated with anaffective deficit involving aversive stimulation. Findings of-70-this sort have been interpreted as consistent with theoriesthat relate psychopathy to a specific deficit in theneurophysiological systems modulating fear (Fowles, 1980; Gray,1971).However, a limitation of this research is that there havebeen no attempts to assess the reactivity of psychopaths in avariety of emotional situations. Experiment 2 thereforeinvestigated the emotional reactions of psychopaths insituations designed to elicit either positive or negativeaffective states.MethodSubjects The same subjects as in Experiment 1.Stimulus Materials Each subject viewed seven films, each approximately 80 slong. The first film was neutral in valence in order toacclimatize the subject to the procedure. The next two filmswere intended to evoke positive emotions (amusement andhappiness), followed by four films designed to evoke negativeemotions (disgust, fear, sadness, and anger). A description ofeach film is presented in Appendix A. Studies with collegestudents have indicated that these films succeed in elicitingthe targeted emotion (Levenson, personal communication, July,-71-1990).The films were in colour with a sound track. The positivefilms were always presented first because of the finding byEkman and his associates (Ekman et al., 1980; Ekman et al.,1988) that the affect elicited by negative films tends topersist longer than the affect elicited by the positive films.Within types of film (positive, negative) the order ofpresentation was counterbalanced.Apparatus and Response MeasurementThe physiological measures, amplification, and samplingrate were the same as in Experiment 1. Physiological signals(heart rate, skin conductance, and electrocortical activity)were sampled 10 s prior to the onset of the film clip, andthroughout each of the film clips. Although these films wereintended to arouse only one emotion, normal subjects oftenreport an array of emotions if given the opportunity to do so.In order to determine which emotions were elicited, after eachfilm clip the subject was asked to complete the Affect Grid anda series of affective ratings describing how he felt duringexposure to the film. The following "differential" ratingscales were used: interested, content, happy, amused,surprised, sad, afraid, disgusted, and angry. Each rating wasmade on a 10-point scale, with 0 representing "Do not feel atall" to 9 representing "Feel very strongly" (see Appendix B).To reduce possible demand characteristics, the subject was-72-asked to report what he actually felt during the film and notwhat he thought he should feel. He was told that people varyconsiderably in their responses to the films and that therewere no right or wrong responses.ProcedureThe experimenter told the subject that he would now watcha series of films (displayed on a monitor), and that this partof the experiment was designed to examine the relation amongphysiological variables and subjective experience. The subjectsat approximately 1.3 m from a colour TV monitor. Theexperiment began with a 10 s baseline recording of EEG and ANSactivity, after which each of the film clips was presented.Another 10 s baseline period followed each film. The subjectwas instructed to relax with eyes open during the baselineperiods.After each film clip he was asked to rate the emotions hehad just experienced. The next film clip started approximately4 min after the end of the preceding film. The experimenterremained on the other side of a sound screen (out of sight ofthe subject) during presentation of the film clips.The subject's facial activity was videotaped while heviewed the film clips. Following the presentation of the filmclips the experimenter removed all electrodes and debriefed thesubject.-73-Data Analysis Prior to the study three students in psychology unaware ofthe purpose of the research identified the section in each filmdeemed the most emotion-evoking.Two electrodermal variables were coded: the amplitude ofthe SCR immediately following the critical point in the film;and, as a measure of overall electrodermal arousal, the meannumber of SCRs throughout each film clip. All SCRs larger than.05 micromhos were counted.Heart rate (beats per minute) was coded by hand directlyfrom the polygraph records. Heart rate change scores werecomputed by subtracting the mean rate during the last five-s ofbaseline for a given film from the mean rate during ten simmediately following the critical point (i.e., the mostemotion-evoking part) in the film. This difference score, whichattempts to control for individual differences in heart rate,was used as the dependent variable in the statistical analyses.All EEG data were screened off-line for eye movementartifact, and trials contaminated by artifact were excludedfrom data analysis. Analyses were performed on EEG data fromall artifact-free periods during and immediately following thecritical point in the film. Chunks of 2 s of EEG were extractedusing a Hamming window and were overlapped by 50%. A FFT wasapplied to each chunk of EEG and power values from successiveepochs within a film were averaged. The dependent measure thatwas extracted from this analysis was power density (in )1V2/Hz)-74-in the alpha (8-13 Hz) band. The data were log-transformed tonormalize the distributions because power values werepositively skewed.The FACS system was used to code facial activity duringthe most emotion-evoking part of the film: the 3 s prior to thecritical part and the subsequent 7 s. As a reliability check, asecond rater (the same as used in Experiment 1) coded 12randomly selected subjects. The mean ratio (described inExperiment 1) was .78 across all scored events, and .93 for "noscorable" facial actions. The Emotional Facial Action CodingSystem (EMFACS; Friesen & Ekman, 1984) was used to convert thefacial action scores into the following categories: primaryemotions (happiness, anger, sadness, disgust, contempt, fear,surprise, and interest), blends, and unspecific negativeemotions. The program provides frequency scores for each of theprimary emotions and scores for the co-occurrence of two ormore of these emotions in blends. Unspecific negative emotionsare facial expressions that cannot be assigned to anyparticular negative emotion.Ekman (1985) has described several different types ofsmiles, but it is only the Duchenne ("enjoyment") smile that isreliably associated with reports of happiness; these are smilesthat are produced when the zygomatic major muscle lifts thecheeks, and the skin around the eye is gathered by theorbicularis oculi muscle. In a recent report, Ekman et al.(1990) found that left-sided anterior temporal activation was-75-greater during Duchenne smiles than during other types ofsmile. In addition, the amount of Duchenne smiling waspositively correlated with subjective reports of positiveemotions.ResultsPreliminary analyses indicated that the order in which thefilms were presented had no effect on the results. The order offilm presentation was not included in the analyses reportedbelow.Self-report DataResponses to each of the nine differential rating scalesused to assess the subject's emotional responses to each filmwere analyzed in nine separate 2 X 6 ANOVAs, with group as abetween-subjects factor and film type as a within-subjectsfactor.Group means for each of the rating scales are presented inFigure 7 (positive films) and in Figures 8 and 9 (negativefilms). None of the film type X group interactions weresignificant. There were significant differences between filmconditions for all of the ratings. The differences betweenindividual means were tested using post-hoc comparisons (TukeyHSD test). The results, collapsed across groups, are summarizedin Table 8. The highest rating corresponded to the putativecontent of the film. That is, subjects reported feeling theEa Nonpsychopaths■ Psychopaths-00I ,e0-81z0C00-o0toE<Y0(1)i5.-mco86ill -, 4E40)cfa. 4-mcocu2 212mU2a)C.C-0C.O0-o00E<-76-Amusing Film8ciiicas•-0^r^-0m m mUc^com^<0aHappy Filmy^:0^y^?.."^m0 cmm hcTi.^;74^7 <7 0V) 6Rating ScaleFigure 7. Mean self-report ratings to the amusing film (upperpanel) and happy film (lower panel) by diagnosticgroup.• Nonpsychopaths■ Psychopaths-77-Sad Film0E0.aC00Fear Film3Ea0000a300"0C.)053rnco0860Iccoccccaa)42210^:0^0El 153I.^...1t-^=m2.W^g%JOC8 -6 -co-..-2 4 -CZ .cco0 2 -M0) 0Rating ScaleFigure 8. Mean self-report ratings to the sad film (upperpanel) and fear film (lower panel) by diagnosticgroup.86laa)Wemca)C0C.)>.aaas=El Nonpsychopaths■ Psychopaths-omNE<L"o)C<86-a0mE<V00r.se—mWva)U,=m06-78-Anger FilmDisgust Film_^ ,..-o^C^am m a1-4^Z^0T m00^oCRating ScaleFigure 9. Mean self-report ratings to the anger film (upperpanel) and disgust film (lower panel) by diagnosticgroup.-79-Table 8Emotional ratings in the film conditions, for all subjectsF RatioFilm condition^ (dfs=5,36)Rating^Happy Amused Sad^Fear Anger Disgust EmotionHappy^4.81b^5.33b^1.09a^1.00a^1.09a^.67a^23.87*Sad^.67a^.12a^4.52b^.33a^1.43a^.45a^28.11*Content^2•93ab 3.86 a^1.17bc 1.50bc 1.33bc^.38c^20.22*Surprised^.98ab 2.38ab^.59a^2.62b^1.17ab 5.28c^32.52*Afraid^.05a^.02a^.67a^1.81b^.45a^.19a^5.17Disgusted^.28a^.07a^1.12a^.38a^3.38b^7.91c 141.58*Interested 4.05 ab 5.26b^3.31abc 5.48b^4.33ab 1.29c^21.38*Amused^2.69a^6.10b^1.02a^2.49a^2.07a^1.58a^20.95*Angry^.29a^.05a^2.24ab 1.09a^4.54b^1.69a^15.04*Note. Means with different subscripts differ significantly atp < .005 * p < .001-80-most happy when watching the happy film, most amused whenwatching the amusing film, most fearful when watching thefearful film, most angry when watching the anger film, most sadwhen watching the sad film, and most disgusted when watchingthe disgusting film.For the positive emotions, the amused ratings weresignificantly higher in the amused film condition than in anyof the other film conditions. For the happy ratings, the happyand amused films did not differ significantly from each otherbut were significantly different from all the other filmconditions.For the negative emotions, the sad ratings weresignificantly higher in the sad film condition, and the fearratings were significantly higher in the fearful filmcondition, than in any of the other film conditions. Theresults for angry and disgust were slightly more complex. Thedisgust ratings were significantly higher in the disgust andanger films than as the other film conditions. The angerratings to the angry film differed significantly from all theother film conditions, except the sad film condition.Although the fear film elicited higher fearful ratingsthan did the other films, it also elicited higher ratings ofinterest, surprise, and amusement than fear. Clearly this filmdid not elicit strong feelings of fear in this subjectpopulation.In addition to these differences for the target feelings,-81-there were also differences for some of the other ratings. Inthe disgust film, subjects reported feeling more surprised andless interested than in the other film conditions, and in theamused condition, they reported feeling more content than inthe other film conditions.The percentage of subjects in each group that reportedexperiencing the target emotion as the strongest is presentedin Table 9. Except for the fear film, the majority of subjectsreported feeling the target emotion the strongest. Very cleardiscrimination across the positive and negative films occurredwhen the positive (happy, content, amused) and negative (anger,disgust, fear, sad) ratings were collapsed. As can be seen inTable 10, for the happy and amused films, nearly all subjectsrated the films are being positive. For the negative films,except the fear film, the most strongly felt emotion wasnegative.Affect GridThe Affect Grid was administered as an additional methodof assessing the subject's emotional reactions to the filmclips. The pleasure-displeasure and low-high arousal dimensionswere analyzed in a 2 X 6 MANOVA, with group as the between-subjects factor and film type the within-subjects factor.The dimensional ratings varied across the films. The onlymultivariate F to reach significance was the main effect forfilm type (F(5,36) = 35.75, p < .001). Univariate tests on the-82-Table 9Percentage of nonpsychopaths and psychopathswho rated the target emotion the strongestPCL-R GroupFilm type^Nonpsychopaths PsychopathsAmused^76^86Happy 95 62Sad^ 76^67Anger 62 62Fear^24^19Disgust 90 95-83-Table 10Percentage of nonpsychopaths and psychopaths ratingpositive or negative emotions the strongest acrossfilm conditionsNonpsychopaths PsychopathsFilm type Positive Negative Positive NegativeAmused 95 0 100 0Happy 100 0 90 0Sad 14 86 24 76Anger 24 71 24 71Fear 24 33 24 48Disgust 05 95 0 95Note. Positive emotions included happy, content, andamused. Negative emotions included anger, disgust,fear, and sadness. Numbers do not add up to 100percent since subjects could have rated the strongestemotion as neutral or surprise.-84-two dependent variables indicated that the main effect wassignificant for both pleasure-displeasure, (F(5,36) = 41.41, p< .001, and arousal ratings, (F(5,36) = 24.20, p < .001). Therewere no significant group X film type interactions. Figure 10orders the film clips around a circumplex. With the exceptionof the fear film, the films' position in the circumplex isconsistent with previous research. For example, the amused filmfell into the quadrant of pleasure/high arousal, whereas thesad film fell into the quadrant of displeasure/low arousal. Itwas expected that the fear film would have been located in thedispleasure/high arousal quadrant; however, ratings ofemotional reactions to the fear film fell in the pleasure/higharousal quadrant.Physiological DataEach set of physiological data (SCR, heart rate) wasanalyzed with a 2 X 6 (group X film type) ANOVA, with group asa between-subjects factor and film type as a within-subjectfactor. Heart rate and skin conductance activity of thepsychopaths and nonpsychopaths during exposure to each film arepresented in Figures 11 and 12, respectively. There was asignificant film type main effect for heart rate, (F(5,34) =7.92, p < .001), mean number of SCRs, (F(5,34) = 17.78, p <.001), and mean amplitude of the SCR (F(5,34) = 10.78, p <.001). None of the group differences were significant. Thenature of this emotion-specific ANS activity was explored with-85-High Arousal•Angry•Fear•Amused•HappyCcoC•Disgust0Cu05:•SadLow ArousalFigure 10. Film clips scaled on the dimensions of pleasure-displeasure and low to high arousal (ratings basedon Affect Grid).-4^I^I^I^I^I^IAmusing Happy Sad^Fear Angry DisgustFilm TypeFigure 11. Mean heart rate change (in beats per minute) fornonpsychopaths and psychopaths during film clips.0to^1.0-0^0.8-iLE^0.6 -1C0.4-i2Amusing Happy Sad^Fear Angry DisgustFilm TypeFigure 12. Mean skin conductance response for nonpsychopaths andpsychopaths during film clips.-88-Bonferroni-adjusted t tests, revealing the followingsignificant differences across films: heart rate increased morein amusement than in anger, disgust, sadness, or happiness anddecreased more in disgust than in sadness, fear, or anger. SCRswere larger in amusement than in happy, sadness, or fear, werelarger in anger than in happy or fear and were larger indisgust than in happy, sadness, or fear.Data from the number of SCRs during the film clipsindicated that arousal was greatest during the amusing film (M= 5.15), followed by the angry film (M = 2.80), disgust film (M= 2.52), happy film (M = 2.10), sad (M = 1.22), and fear film(M = 1.02).Facial ExpressionsThe psychopaths and nonpsychopaths did not differ in thetotal number of facial events elicited by the films(psychopaths M=22.57, SD=18.99; nonpsychopaths M=18.09,SD=12.70).In the analyses that follow only those facial events thatare interpreted according to the EMFACS lexicon were examined.Contrary to prediction, the total number of interpretable lowerfacial events was higher for the psychopaths (M=14.72,SD=13.20) than for the nonpsychopaths (M=11.61, SD=8.36), butthe difference was not significant. Similarly, psychopaths hadmore upper facial actions (M=8.14, SD=7.00) than didnonpsychopaths (M=6.91, SD=4.82), but the difference was not-89-significant.The two groups were compared on those facial events thatcould be interpreted as being primary emotions: happiness,disgust, contempt, fear, sadness, anger, and surprise. Facialevents not included in the analyses were: blends and masks, thecategory of "unfelt happiness", and those expressions thatcould not be clearly classified. There were no groupdifferences in the overall number of primary emotions expressed(psychopaths: M = 5.71, SD = 5.11; nonpsychopaths: M = 5.52, SD= 3.68).The films varied dramatically in their ability to elicitfacial expressions of primary emotions. Table 11 presents thenumber of subjects who displayed different discrete facialexpressions of emotions during each of the films. Only twofilms elicited significant facial expressions concordant withthe content of the target film: the amused and disgust films.Because few facial expressions were elicited by the otherfilms, the remaining analyses will focus on the amused anddisgust films.During the amused film, 86% of nonpsychopaths and 67% ofpsychopaths displayed facial signs of enjoyment (Duchennesmile), and none of the subjects displayed signs of disgust. Inresponses to the disgust film, 52% of nonpsychopaths and 33% ofpsychopaths displayed facial signs of disgust. However, thedisgust film also elicited facial signs of enjoyment in manysubjects (19% of the nonpsychopaths and 33% of the-90-Table 11Number of nonpsychopaths and psychopaths displaying discretefacial expressions during film clipsJoy^Sadness^Anger^Fear^DisgustFilm type NP^P^NP^P^NP^P^NP^P^NP^PAmused^18^14^0^0^0^1^0^0^0^0Happy^3^3^0^0^0^1^0^0^0^0Sad 0^0^1^0^1^2^1^0^1^0Anger^4^4^1^0^0^0^0^0^2^2Fear^3^1^0^0^1^1^0^0^0^2Disgust^4^7^0^0^2^1^1^0^11^7Note. NP = nonpsychopaths; P = psychopaths. Note that more thanone facial expression could have been contributed by a subjectfor a particular film clip. None of the subjects showed facialsigns of surprise or contempt.-91-psychopaths). There was a relative absence of facial signs ofanger, fear, or sadness during each of these film types.To determine if more Duchenne smiles were elicited by theamusing film than by the disgusting film, the data were enteredinto a 2 X 2 X 2 group (psychopath or nonpsychopath) by filmtype (amusment or disgust) by expression type (Duchenne smileor disgust expression) ANOVA.The main effect for film type was significant, (F(1,40) =27.38, p < .001), indicating that the the film types differedin the overall number of expressions elicited: the amusing filmproduced more facial expressions than did the disgusting film.The main effect for expression type was also significant,F(1,40) = 49.59, p < .001; regardless of film type there weremore Duchenne smiles than disgust expressions. The significantinteraction between film type and expression type, F(1,40) =47.09, P < .001, showed that different types of expressionsoccurred in the amusement and disgust films. There were moreDuchenne smiles during the amused film than during the disgustfilm (t = 6.27, p < .001), and there were more disgustexpressions during the disgust film than during the amusingfilm (t = 5.19, p < .001). There were no significant groupeffects. Figure 13 shows the mean frequency of each type ofexpression during the disgust film.If Duchenne smiles occur in response to positive emotionalexperiences then their frequency should be correlated withsubjective reports of positive emotions. The frequency of0.8Oe 0.6 -1aO>. 0.4 -IC)o-wLt 0.2caca20.0Duchenne Smiles^Disgust ExpressionDisgust FilmFigure 13. Mean number of facial expressions of enjoyment anddisgust displayed by nonpsychopaths and psychopathsto the disgust film.-93-Duchenne smiles during the amusing film was significantlycorrelated (R < .001 in each case) with ratings of happiness(r=.51), contentment (r=.52), amusement (r=.50), and interest(r=.52). Similarly, for the disgust film, the frequency ofdisgust expressions should be correlated with subjectivereports of negative emotions. However, although in the expecteddirection, none of the correlations between frequency ofdisgust expressions and ratings of negative emotions reachedsignificance.Electrocortical Data Only subjects who were strongly right-handed were includedin these analyses (see Experiment 1). Recent research (Davidsonet al., 1990a) indicates that cerebral asymmetries in responseto positive and negative films occur only when facial actionsare used to verify the presence of the targeted emotion. In thepresent study, the self-report data indicated that the mostintense ratings of emotional arousal occurred during theamusing and disgusting films. In addition, it was only duringthese films that facial signs of emotion occurred. In order toensure that the targeted emotion had been generated, analysesof EEG data were limited to only those subjects who displayedfacial expressions of the emotions that were consistent withthe film type presented to them. Subjects were included if theydisplayed facial signs of enjoyment (i.e., Duchenne smiles)during the amusing film and facial signs of disgust during the-94-disgust film. Although the majority of subjects (86.7%)displayed facial signs of enjoyment during the amusing film,only 13 (43%) of these displayed facial signs of disgust duringthe disgust film. Of these 13 subjects eight werenonpsychopaths and five were psychopaths.An ANOVA with group (psychopath or nonpsychopath) as thebetween-factor and film type (amusing or disgusting), region(frontal or parietal), and hemisphere (left or right) aswithin-factors was performed on log alpha power. Significanteffects were obtained for film type, F(1,11) = 18.82, p < .005,hemisphere, F(1,11) = 28.17, p < .001, region, F(1,11) = 7.32,R < .05, and the hemisphere X region interaction, F(1,11) =114.87, p < .001. The film type effect was due to less alphapower (i.e., greater activation) during the disgust film thanduring the amusing film. The region effect was primarily afunction of less alpha power in frontal than in parietalregions. The hemisphere effect was due to less alpha power inthe left than in the right hemisphere. The means for alphapower in the left and right hemisphere for each group aredisplayed in Figure 14 (frontal region) and in Figure 15(parietal region). On the basis of a priori predictionsconcerning the locus of the asymmetrical effects, separateANOVAs were computed on the frontal and parietal EEG data.Considering the frontal region first, there was asignificant main effect for film type F(1,11) = 11.64, p < .05,hemisphere (F(1,11) = 23.19, p < .001) and a significant groupAmusing Film Disgust FilmNonpsychopaths Psychopaths Nonpsychopaths PsychopathsFigure 14. Mean log-transformed alpha power (in pV2 /Hz)for the left and right frontal (F3 and F4,respectively) region during the amusing anddisgust films for nonpsychopaths and psychopaths.t; 4.0-30a.coo.-c 3.5 -"0103.01• Left Hemisphere■ Right Hemisphere4.52.5Nonpsychopaths Psychopaths Nonpsychopaths PsychopathsAmusing Film^Disgust FilmFigure 15. Mean log-transformed alpha power (in uV 2 /Hz) for theleft and right parietal (P3 and P4, respectively)region during the amusing and disgust films fornonpsychopaths and psychopaths.-97-X hemisphere X film type interaction F(1,11) = 4.89, p < .05.There was less power during the disgust film than during theamusing film, and greater power in the left hemisphere (i.e.,right-sided activation) than in the right hemisphere. Post hoccomparisons for the frontal ANOVA revealed that during thedisgust film nonpsychopaths showed significantly less right (p< .05) than left frontal power, indicative of right-sidedactivation. All eight nonpsychopaths who displayed facialexpressions of disgust exhibited right frontal activationduring the disgust film, whereas only two of the fivepsychopaths exhibited this pattern. None of the other pair-wiseeffects reached significance at the .05 level.The parietal ANOVA revealed significant effects for filmtype F(1,11) = 19.25, p < .005 and hemisphere (F(1,11) =120.19, p < .001). No significant group interactions werefound. Consistent with the findings from the frontal regionthere was less power during the disgust films than during theamusing film. In contrast to the frontal area, power wasgreater (less activation) in the right hemisphere than in theleft hemisphere.In summary, no significant difference in left and rightfrontal alpha power was found to the amusing film for bothpsychopaths and nonpsychopaths. The disgust film was associatedwith significantly less frontal alpha power (greateractivation) in the right hemisphere than in the lefthemisphere, but only for the nonpsychopaths. In the parietal-98-region, both the amusing and the disgust films were associatedwith more left-sided than right-sided activation, for bothgroups.DiscussionExperiment 2 was designed to examine psychopaths'affective reactions across a variety of both positive andnegative emotions. The results of this experiment indicate thatpsychopaths do not differ from criminal controls in theiraffective self-report, autonomic nervous system response, orobserved facial expressions to emotional stimuli. The twogroups were very similar in their ratings of the intensity oftheir emotional responses to the target emotions. Thesefindings are consistent with those of Experiment 1, in whichaffective slides failed to produce any group differences insubjective self-report or physiological response.In a more general sense, Experiment 2 provides somesupport for the proposal that different emotions are associatedwith different patterns of cardiovascular activity (Ekman etal., 1983; Levenson et al., 1991). The emotion-specific ANSactivity of the criminal subjects bears a strong resemblance tothat reported for noncriminal populations.Although psychopaths displayed more lower facial actionsin Experiment 1 than did nonpsychopaths, this group differencewas not evident in Experiment 2. A problem in evaluating thisfinding is that there were relatively few codable facial-99-expressions in either group. In most of the previous researchon facial expressions the subjects were videotaped withouttheir knowledge. A recent exception is a study on physiologicalreactions to emotions in the elderly (Levenson et al., 1991) inwhich subjects were informed that they were being videotaped.In the present study it was not possible to use a hidden camerato record facial activity. It is possible therefore that thesubjects may have attempted to inhibit their facial expressionssince they knew they were being videotaped. Attempts were madeto minimize the impact of the camera by placing it in a cornerand telling subjects that it would be used to screen formovement-related EEG artifact.The small number of facial expressions may have also beenrelated to the instructions given to subjects to not movearound during the presentation of the slides and film clips inorder to minimize the amount of movement artifact in the EEGchannels. However, it may have had the added effect of reducingthe amount of facial activity to the emotional stimuli. Therewere few instances of facial expressions of fear, sadness, oranger in response to the film clips. However, a relatively highpercentage (79%) of subjects displayed facial signs ofenjoyment to the amusing film, and nearly half (43%) displayedfacial expressions of disgust in response to the disgustingfilm. This would suggest that there was emotion-specificinhibition of facial expressions. Incarcerated offenders mayhave learnt to inhibit the expression of certain types of-100-negative emotions. For example, in prison, the expression ofanger or fear may bring about unwanted confrontations withother offenders or staff.The amusement and disgust films produced greater activityin the frontal lobes than in the parietal lobes, supporting thehypothesis that the frontal lobes are involved in theprocessing of both negative and positive emotions (see reviewby Davidson, 1984; Kinsbourne, 1978; Silberman & Weingartner,1986). This finding is consistent with the evidence indicatingthat the frontal lobe has extensive anatomical connections withthe limbic circuits that have been implicated in the controlover emotion (Nauta, 1971; Rolls, 1986).With respect to cerebral asymmetry, both psychopaths andnonpsychopaths displayed more left-sided than right-sidedparietal activation during the amusing and disgust films.Asymmetry in the frontal region was confined to the disgustfilm and to the nonpsychopaths.One methodological concern is the number of subjects thatcould be used in the EEG analyses. Fifty-seven percent of theoriginal sample was eliminated because of failure to show atleast one instance of a codeable happy and disgust expressionduring the amusing and disgusting films. By way of comparison,other (similar) research projects typically drop only about 30%of the subjects because of the absence of codeable responses(Davidson et al., 1990a). However, the subjects in thesestudies have usually been female and noncriminal. Several-101-studies have found that affective stimuli produce greaterfacial expressivity in women than in men (Greenwald et al.,1989; Schwartz et al., 1981). Moreover, offenders probably aremore likely to try to control their emotional expressions thanare nonoffenders.Although the subjects who were retained rated the targetemotions about the same as did those who were dropped, it isconceivable that the former differed in some important way fromthe latter. Future studies may have to use either large samplesof subjects or stimuli capable of eliciting a greater number ofexpressions of emotion than was the case in the present study.General DiscussionAlthough there have been some attempts to measure emotionsin psychopaths, most researchers have focused on only oneresponse domain at a time, typically the autonomic nervoussystem or self-report measures of fear or anxiety. The purposeof the present experiments was to look at several components ofemotion elicited by a range of positive and negative stimuli.The dependent variables were subjective reports of emotionalreactions, indices of peripheral and central nervous systemfunctioning (electrodermal, cardiovascular, andelectrophysiological activity), and facial expressions.Dimensional and discrete models of emotion were used.It should be noted that the separation between the twogroups of subjects--in terms of the independent variable,-102-psychopathy--was not as sharp as it has been in other studies.This is because at the time this study was conducted theinstitution happened to house relatively few inmates with lowscores on the Revised Psychopathy Checklist (PCL-R). As aresult, the nonpsychopathic group was not as "nonpsychopathic"as it might otherwise have been. Some indication of theconsequences this might have had for the results can be gleanedfrom the Affect Intensity Measure (AIM); the mean score of thenonpsychopaths was considerably lower than were the AIM scoresfrom more extreme groups of offenders used in another study(Day, 1991).Experiment 1 provided no evidence that psychopaths differfrom nonpsychopaths in autonomic responsivity to pleasant orunpleasant stimuli. These results differ from those obtained byseveral other researchers (see reviews by Hare, 1978; Siddle &Trasler, 1981). However, most previous studies focused onanticipatory autonomic activity to discrete stimuli (tones,electric shocks), whereas in the present study autonomicmeasures were recorded during exposure to affective visualstimuli. Using the same sorts of affective slides as thosepresented in Experiment 1, Patrick et al. (1991) also failed tofind significant differences between psychopathic andnonpsychopathic sex offenders in autonomic nervous systemactivity or self-report measures of emotion.The negative and positive films in Experiment 2 generallyelicited different patterns of heart rate activity, a finding-103-that is consistent with results from several other recentstudies (Ekman et al., 1983; Levenson et al. 1990: Levenson,Carstensen, Friesen, & Ekman, 1991). Although these findingssupport the argument that different emotions are associatedwith distinctive patterns of autonomic responsivity, it shouldbe noted that several researchers recently have failed toobserve different physiological patterns for different emotions(Leventhal & Tomarken, 1986; Tassinary, Cacioppo, & Geen, 1989;see also Rime, Philippot, & Cisamolo, 1990).Although the primary purpose of this research was to studyemotion in psychopaths, the data are also relevant to the issueof whether or not emotion-specific ANS activity occurs acrossdifferent modes of elicitation. To date, research has focusedon directed facial action tasks and relived emotion tasks tostudy the relation between specific emotions and specificpatterns of ANS activity. In this study, film clips were usedand the results provide support for the notion that ANSactivity is emotion specific and not task specific. Moreover,the similarity between the responses of college students andmale incarcerated offenders supports the generality of theautonomic specificity hypothesis.In Experiment 2 the magnitude of the heart rate response--particularly to the anger, fear, and sad films-- wasconsiderably smaller than is typically found in noncriminals(Levenson et al., 1990). It is possible that emotionalreactions are more difficult to elicit from criminals than from-104-noncriminals. The subjective self-reports and facialexpressions of the criminals in this study support thisinterpretation. Thus, the self-reports--particularly those forthe fear film--indicated that strong emotional reactions werenot elicited. Similarly, very few facial expressions ofsadness, anger, or fear occurred during exposure to any of thefilms. The stimuli used to elicit emotional reactionsapparently were less effective than they are with noncriminals.Overt facial actions provided an index of facial activity.However, this type of measurement requires an observable musclemovement for an expression to be scored. If criminals arereluctant to express emotions overtly, it would be necessary touse facial EMG activity as an index of covert muscle activity--activity at levels insufficient to produce actual changes infacial features. Facial EMG might be used in combination withovert facial action coding systems in future studies, in orderto allow examination of potential differences between covertand overt facial expressions in criminal populations.When films are used to elicit emotions there are periodsin which no or little emotion is present or when emotionsinconsistent with the target emotion occur. The best method toensure that the EEG is sampled during periods of the targetemotion is to use facial expressions to flag onset and offsetof relevant sampling periods (Davidson et al., 1990a). Thiscould not be done in the present study because the exact onsetand offset of specific facial expressions could not be-105-synchronized precisely with sampling of the EEG. If EEG hadbeen sampled at the exact time the facial expression had beendisplayed stronger cerebral asymmetries may have been found.It is not clear what parietal asymmetry (greater left thanright hemisphere activation) might reflect in theseexperiments. Other EEG asymmetry studies have found that verbalcognitive activity reliably increases left parietal activationmore than right parietal activation (e.g., Davidson et al.1990b; Ehrlichman & Weiner, 1979). It is therefore possiblethat by asking subjects to rate the films, the affectiveresponse was accompanied by heightened cognitive activity. Infuture studies it would be useful to assess cognitive activityin addition to experienced emotion in order to evaluate thispossibility.Although the results from numerous studies, including thisone, seem to support the view that both frontal lobes areinvolved in the processing of emotion, the relative roles ofeach of the hemispheres in relation to positive and negativeaffect remains controversial. While some studies have reportedhemispheric asymmetry of the frontal regions in the processingof emotion (e.g., Ahern & Schwartz, 1985; Robinson et al.,1984; see Davidson & Tomarken, 1989 for review), others havenot found this to be so (Davidson & Fox, 1982; Fox, in press;Meyers & Smith, 1987). For example, Davidson et al. (1990a)found that although facial expressions of disgust wereassociated with significant increases in right frontal-106-activation--relative to baseline conditions--frontal asymmetryduring facial expressions of enjoyment failed to be associatedwith significant increases in left frontal activation.Similarly, although Tomarken et al. (1990) found that restingfrontal asymmetry predicted negative affective responses tofilm clips, it failed to predict positive affective response.Kinsbourne (1978) was the first to theorize that frontaland anterior temporal cortical regions are lateralized along anapproach-withdrawal continuum, with approach being associatedwith left hemisphere activation and withdrawal with righthemisphere activation. Davidson (1984; Davidson & Tomarken,1989) has further argued that to the degree that withdrawal andapproach are components of different emotions, such emotionsshould differentiately activate the two cerebral hemispheres.Thus, relative right frontal activation is associated withincreased withdrawal tendencies and negative emotions thatinclude behavioral components of withdrawal, such as fear anddisgust. Relative left frontal activation is associated withapproach tendencies and positive emotions such as amusement.Strong left frontal activation is particularly evident whenpositive affect is associated with unambiguous approachbehavior, such as when babies reach for their mothers (Fox &Davidson, 1987, 1988; for review see Fox, in press).In the present study, the amusing film was not associatedwith left frontal activation, a finding that was contrary toexpectation. However, asymmetry in frontal activation may-107-depend on the extent with which the emotion is associated withapproach-withdrawal tendencies. Thus, even though the primaryself-reported emotion elicited by the amusing film was indeedamusement, it is possible that the film did not elicit atendency to approach. Davidson et al. (1990a) have suggestedthat not all types of amusement will activate approachbehavior. In order to evaluate this hypothesis, EEG asymmetrywould have to be measured during positive affect states thatvaried in approach tendency (e.g., as assessed by self-reports).Davidson et al. (1990a) have suggested that certainemotions are hard-wired to include withdrawal or approachaction tendencies. However, they also have indicated thatlearned components may alter inherent approach and withdrawalcomponents. In addition, emotions such as anger or sadness mayreflect either approach or withdrawal, depending on thestimulus context and the subject's appraisal of the context.Psychopaths may differ from others in terms of the extent towhich certain emotions elicit approach or withdrawalcomponents. For example, the psychopath (as a stimulation-seeker) may respond to fear with approach rather thanwithdrawal. It would be useful to include self-report measuresof approach and withdrawal in future research in order todetermine to what extent the emotions being elicited areassociated with these components in psychopaths andnonpsychopaths.-108-The disgust film produced the expected asymmetry (relativeright frontal activation) in nonpsychopaths, but not inpsychopaths. The finding for the nonpsychopaths is consistentwith previous electrophysiological data from both infants(e.g., Davidson & Fox, 1982; Fox & Davidson, 1988) andnoncriminal adults (Ahern & Schwartz, 1985; Davidson et al.,1990a); that is, negative emotions that elicit withdrawal areassociated with relative right frontal activation.The failure of psychopaths to show relative right frontalactivation during exposure to the disgust film is of interest,given recent attempts to explain psychopathy in terms oflateralized cerebral dysfunction (Day, 1991; Gillstrom & Hare,1988; Hare & McPherson, 1984; Hare et al., 1988). This lack offrontal asymmetry in psychopaths apparently was not due tofailure of the film to elicit the appropriate emotion; theirratings of disgust were just as high as were those of thenonpsychopaths.Although the EEG data are consistent with the view thatthe right hemisphere of psychopaths has difficulty inprocessing negative affect, the small number of subjects (13)and the possibility of inflated Type I errors associated withthe calculation of many statistical tests reduce confidencethat the finding is reliable. Clearly, more research iswarranted. But meanwhile, there are other lines of evidencethat seem to implicate right cerebral dysfunction inpsychopathy.-109-The consequences of brain damage to a large extent dependon which hemisphere suffers the damage (Gainotti, 1972; Sackeimet al., 1982). For example, Gainotti (1972) reported that aprimary feature right frontal damage is "indifference,"characterized by denial or minimization of symptoms and byjocularity. In contrast, left frontal damage is associated with"catastrophic" reactions such as tears, swearing, andexpressions of fear and anger. More recent research hasconfirmed these findings (Robinson et al., 1984). There is alsoevidence that damage to the right hemisphere results in aspecific affective deficit. For example, Benowitz et al. (1983)reported that patients with right hemisphere damage showeddeficits in perceiving emotion in faces and in the tone ofvoice. In addition, Cicone, Wagner, & Gardner (1980) found thatsuch patients lack sensitivity to the polarity of emotions.The recent finding (Patrick et al., 1991) that psychopathsfail to show probe startle modulation to aversive stimuliappears to be compatible with a right hemisphere deficit viewof psychopathy. In Patrick et al.'s (1990) study, the startledeficit was restricted to aversive stimuli, with psychopathsshowing no startle abnormalities in response to pleasantstimuli. Lang, Bradley, & Cuthbert (1990) have theorized thatthe startle reflex is a component of the defensive withdrawalreflex. Given the association between the right hemisphere andwithdrawal tendencies (Davidson & Tomarken, 1989), a specificstartle deficit to aversive stimuli may be a reflection of a-110-right hemisphere deficit in the processing of affect.Recent evidence suggests that the affect-startle effectmay be lateralized (Bradley, Cuthbert, & Lang, 1991). Whenmonaural acoustic probes are used left-ear probes showed thestrongest relationship with affective valence, presumablybecause of the right-hemisphere's specialization in themediation of negative affect. If future research demonstratesthat left-ear probes show facilitation of the startle-effect(enhanced response) during aversive stimulation (i.e., duringviewing of unpleasant slides) and right-ear probes showenhanced inhibition of the startle-effect during pleasantstimulation, this could have significant implications for thestudy of emotion in psychopaths. Presentation of lateralizedprobes combined with recorded lateralized EEG may provide asensitive noninvasive method to explore brain lateralization ofemotion in psychopaths.Although the present study was designed to investigate EEGasymmetry across several different emotions, facial expressionsof happy and disgust were elicited only by the amusing anddisgust films; EEG analyses therefore were limited to these twoemotions. The degree to which EEG asymmetry in psychopathsdiffers from that of nonpsychopaths across different emotionsis unknown. However, we would predict that to the extent thatan emotion is associated with withdrawal (e.g., fear, disgust,some types of anger and sadness), there would be less right-sided activation in psychopaths than in nonpsychopaths.On the other hand, a right frontal deficit view ofpsychopathy has difficulty in accounting for the evidence thatpsychopaths show performance deficits when processing affectivelinguistic material but not when processing affectivenonlinguistic material. Williamson et al. (1990) reported thatwhen psychopaths performed an affect-matching task they tendedto confuse emotions of opposite polarity when integratinglinguistic information but not when integrating pictorialinformation. In a study of emotional words and faces describedearlier, Day (1991) found that both psychopaths andnonpsychopaths showed a right hemisphere advantage whenprocessing the emotional content of faces, but that only thenonpsychopaths showed a right hemisphere advantage whenprocessing emotional words. It is possible that psychopaths andnonpsychopaths used different strategies to evaluate theemotional content of the faces. For example, Day (1991)suggested that psychopaths may have made use of visuospatialinformation to evaluate the faces (mouth curves upwards,therefore happy), whereas nonpsychopaths may have directlyextracted emotional information from the faces. Since bothstrategies are mediated by the right hemisphere, no between-group differences would have been detected.Although the EEG results in this study are suggestive itis possible that more fine-grained analysis of cerebralfunctioning will be required in order to properly studyaffective processes in psychopaths. One of the serious-112-limitations of scalp-recorded EEG is the poor spatialresolution obtained (Davidson, 1988). The location of thebrain's electrical activity may be distorted by the skull andthus it is inappropriate to assume that there is a one-to-onerelationship between electrode location and the underlyingsource generator. Studies using measures with better spatialresolution than EEG (e.g., regional cerebral blood flow) maymore precisely specify the locus and extent of abnormalfunctional asymmetries in psychopaths.In conclusion, results from the current study add to thegrowing body of evidence that psychopaths demonstrateabnormalities in functional cerebral asymmetries. Futureresearch on lateral processing of emotion using both verbal andnonverbal stimuli, across a range of distinct emotions, andusing multiple measures of responses, are needed to provideconvergent information on affective abnormalities inpsychopaths.-113-ReferencesAhern, G.L. & Schwartz, G.E. (1985). Differentiallateralization for positive and negative emotion in thehuman brain: EEG spectral analysis. Neuropsychologia, 23,745-756.American Psychiatric Association. (1987). Diagnostic andStatistical Manual of Mental Disorders (3rd ed., revised).Washington, D.C.: Author.Annett, M.A. (1970). A classification of hand preferences byassociation analysis. British Journal of Psychology, 61,303-321.Ax, A.F. (1953). The physiological differentiation between fearand anger in humans. Psychosomatic Medicine, 15, 433-442.Ax, A.F. (1962). Psychophysiological methodology for the studyof schizophrenia. In R. Ressler & N. Greenfield (Eds.),Psychophysiological correlates of psychological disorder Map. 29-44). Madison: University of Wisconsin Press.Bartko, J.J. (1976). On various intraclass correlationreliability coefficients. Psychological Bulletin, 83, 762-765.Beck, A.T., Ward, C.H., Mendelson, M., Mock, J., & Erbaugh,J.K. (1961). An inventory for measuring depression. Archives of General Psychiatry, 4, 561-571.Benowitz, L.I., Bear, D.M., Rosenthal, R., Mesulam, M.M.,Zaidel, E., & Sperry, R. (1983). Hemispheric specializationin nonverbal communication. Cortex, 19, 5-11.Bradley, M.M., Cuthbert, B.N., & Lang, P.J. (1990). Startlereflex modification: Emotion and attention?Psychophysiology, 27, 513-522.Bradley, M.M., Cuthbert, B.N., & Lang, P.J. (1991). Startle andemotion: Lateral acoustic stimuli and the bilateral blink.Psycohphysiology, 28, 285-296.Cacioppo, J.T. & Petty, R.E. (1981). Electromyograms asmeasures of extent and affectivity of informationprocessing. American Psychologist, 36, 441-456.Cannon, W.B. (1927). The James-Lange theory of emotions: Acritical examination and an alternative theory. AmericanJournal of Psychology, 39, 106-124.-114-Cicone, M., Wapner, W., & Gardner, H. (1980). Sensitivity toemotional expressions and situations in organic patients.Cortex, 16, 145-158.Cleckley, H.M. (1976). The mask of sanity. (5th ed.) St. Louis,MO: Mosby.Craft, M.J. (1965). Ten studies into psychopathic personality.Bristol: Wright.Darwin, C. (1872). The expression of emotion in man andanimals. London: Murray.Davidson, R.J. (1984). Hemisphere asymmetry and emotion. In K.Scherer & P. Ekman (Eds.), Approaches to emotion (pp. 39-57). Hillsdale, NJ: Erlbaum.Davidson, R.J. (1987). Cerebral asymmetry and the nature ofemotion: Implications for the study of individualdifferences and psychopathology. In R. Takahashi, P. Flor-Henry, J. Gruzelier, & S. Niwa (Eds.), Cerebral dynamics, laterality, and psychopathology (pp. 71-83). New York:Elsevier.Davidson, R.J. (1988). EEG measures of cerebral asymmetry:Conceptual and methodological issues. International Journal of Neuroscience, 39, 71-89.Davidson, R.J., Chapman, J.P., Chapman, L.P., & Henriques, J.B.(1990b). Asymmetrical brain electrical activitydiscriminates between psychometrically matched verbal andspatial cognitive tasks. Psychophysiology, 27, 528-543.Davidson, R.J., Ekman, P., Saron, C.E., Senulis, J.A., &Friesen, W.V. (1990a). Approach-withdrawal and cerebralasymmetry: Emotional expression and brain physiology I.Journal of Personality and Social Psychology, 58, 330-341Davidson, R.J. & Fox, N.A. (1982). Asymmetrical brain activitydiscriminates between positive versus negative affectivestimuli in human infants. Science, 218, 1235-1237.Davidson, R.J., Schaffer, C.E., & Saron, C. (1985). Effects oflateralized presentations of faces on self-reports ofemotion and EEG asymmetry in depressed and non-depressedsubjects. Psychophysiology_, 22, 353-364.Davidson, R.J. & Tomarken, A.J. (1989). Laterality and emotion:An electrophysiological approach. In F. Boller & J. Grafman(Eds.), Handbook of neuropsychological (pp. 419-441).-115-Amsterdam: Elsevier.Davis, M.H. (1983). Measuring individual differences inempathy: Evidence for a multidimensional approach. Journal of Personality and Social Psychology, 44, 113-126.Day, R. (1991). Psychopathy and emotion. Unpublished masterthesis, University of Saskatchewan, Saskatoon, Canada.Dimberg, U. (1982) Facial reactions to facial expressions.Psychophysiology, 19, 643-647.Dimberg, U. (1986). Facial reactions to fear-relevant and fear-irrelevant stimuli. Biological Psychology, 23, 153-161.Duchenne, B. (1990). The mechanism of human facial expressionor an electro-physiological analysis of the expression of the emotions (A. Cuthbertson, Trans.). New York: CambridgeUniversity Press. (Original work published 1862).Duclos, S.E., Laird, J.D., Schneider, E., Sexter, M., Stern,L., & Van Lighten, 0. (1989). Emotion-specific effects offacial expressions and postures on emotional experience.Journal of Personality and Social Psychology, 57, 100-108.Ehrlichman, H. & Weiner, M.S. (1979). Consistency of task-related EEG asymmetries. Psychophysiology, 16, 247-252.Eisenberg, N., Fabes, R.A., Bustamant, D., Mathy, R.M., Miller,P. A., & Lindholm, E. (1988). Differentiation of vicariouslyinduced emotional reactions in children. Developmental Psychology, 24, 237-246.Ekman, P. (1985). Telling lies. New York: W.W. Norton.Ekman, P., Davidson, R.J., & Friesen, W.V. (1990). The Duchennesmile: Emotional expression and brain physiology II. Journal of Personality and Social Psychology, 58, 342-353.Ekman, P. & Friesen, W.V. (1978). Facial Action Coding System(FAGS): a technique for the measurement of facial action.Palo Alto, CA: Consulting Psychologists Press.Ekman, P., Friesen, W.V., & Ancoli, S. (1980). Facial signs ofemotional experience. Journal of Personality and Social Psychology, 39, 1125-1134.Ekman, P., Friesen, W.V., & Ellsworth, P.C. (1972). Emotion inthe human face. New York: Pergamon Press.Ekman, P., Friesen, W.V., & O'Sullivan, M. (1988). Smiles when-116-lying. Journal of Personality and Social Psychology, 54,414-420.Ekman, P., Levenson, R.W., & Friesen, W.V. (1983). Autonomicnervous system activity distinguishes among emotions.Science, 221, 1208-1210.Fitzgibbons, L. & Simons, R.F. (1990). A three-system analysisof emotional response in normal and anhedonic subjects.Psychophysiology, 27, 31.Fowles, D.C. (1980). The three arousal model: Implications ofGray's two-factor learning theory for heart rate,electrodermal activity, and psychopathy. Psychophysiology,17, 87-104.Fox, N. (in press). Frontal brain asymmetry and vulnerabilityto stress: Individual differences in infant temperament. InT. Field, P. McCabe, & N. Schneiderman (Eds.), Stress andcoping.Fox, N., & Davidson, R.J. (1988). Patterns of brain electricalactivity during facial signs of emotion in 10-month-oldinfants. Developmental Psychology, 24, 230-236.Fridlund, A.J., & Izard, C. (1983). Electromyographic studiesof facial expressions of emotions and patterns of emotions.In N.A. Fox & R.J. Davidson (Eds.), The psychobiology ofaffective development (pp. 243-386). Hillsdale, NJ: Erlbaum.Fridlund, A.J., Schwartz, G.E., & Fowler, S.G. (1984). Patternrecognition of self-reported emotional state from multiple-site facial EMG activity during affective imagery.Psychophysiolgy, 21, 622-637.Friedman, B.H. & Thayer, J.F. (1989). Pattern classification ofEEG responses to induced emotion: Site and frequencyspecificities. Psychophysiology, 26, S27.Friesen, W.V. & Ekman. P. (1984). EMFACS-7: Emotional Facial Action Coding System, Unpublished manuscipt.Gainotti, G. (1972). Emotional behavior and hemispheric site oflesion. Cortex, 8, 41-55.Gillstrom, B.J. & Hare, R.D. (1988). Language-related handgestures in psychopaths. Journal of Personality Disorders,2, 21-27.Gorenstein, E.E. & Newman, J.P. (1980). Disinhibitorypsychopathology: A new perspective and a model for research.-117-Psychological Review, 87, 301-315.Gray, J.A. (1975). Elements of a two-process theory of learning. New York: Academic Press.Greenwald, M.K., Cook, E.W., & Lang, P.J. (1989). Affectivejudgment and psychophysiological response: Dimensionalcovariation in the evaluation of pictorial stimuli. Journal of Psychophysiology_, 3, 51-64.Hare, R.D. (1970). Psychopathy: Theory and research. New York:Wiley.Hare, R.D. (1972). Cardiovascular components of orienting anddefensive responses. Psychophysiology, 9, 606-614.Hare, R.D. (1978). Electrodermal and cardiovascular correlatesof psychopathy. In R.D. Hare & D. Schalling (Eds.),Psychopathic behavior: Approaches to research (pp. 107-142).Chichester, England: Wiley.Hare, R.D. (1982). Psychopathy and physiological activityduring anticipation of an aversive stimulus in a distractionparadigm. Psychophysiology, 19, 266-271.Hare, R.D. (1985). Comparison of procedures for the assessmentof psychopathy. Journal of Consulting and Clinical Psychology, 53, 7-16.Hare, R.D. (1991). Manual for the Revised PsychopathyChecklist. Toronto, Ontario: Multi-Health Systems.Hare, R.D. & Craigen, D. (1974). Psychopathy and physiologicalactivity in a mixed-motive game situation. Psychophysiology,11, 197-206.Hare, R.D. Forth, A.E., & Hart, S.D. (1989). The psychopath asa clinical prototype for lying and deception. In J. Yuille(Ed.), Credibility assessment (pp. 25-49). Dordrecht,Netherlands: Martinus Nijhoff.Hare, R.D., Frazelle, J., & Cox, D. (1978). Psychopathy andphysiological responses to the threat of an aversivestimulus, Psychophysiology, 15, 165-172.Hare, R.D. & Gillstrom, B.J. (1990). Hand gestures and speechencoding difficulties in psychopaths. Unpublishedmanuscript, University of British Columbia, Vancouver,Canada.Hare, R.D., Harpur, T.J., Hakstian, A.R., Forth, A.E., Hart,-118-S.D., & Newman, F.P. (1990). The Revised PsychopathyChecklist: Reliability and factor structure. Psychological Assessment: A Journal of Consulting and Clinical Psychology,2, 338-341.Hare, R.D. & Jutai, J.W. (1988). Psychopathy and cerebralasymmetry in semantic processing. Personality and IndividualDifferences, 9, 329-337.Hare, R.D. & McPherson, L.M. (1984). Psychopathy and perceptualasymmetry during verbal dichotic listening. Journal of Abnormal Psychology, 93, 141-149.Hare, R.D., Williamson, S.H., & Harpur, T.J. (1988).Psychopathy and language. In T.E. Moffitt & S.A. Mednick(Eds.), Biological contributions to crime causation (pp. 68-92). Dordrecht, The Netherlands: Marinus Nijhoff.Hare, R.D. & Quinn, M.J. (1971). Psychopathy and autonomicconditioning. Journal of Abnormal Psychology, 77, 223-235.Harpur, T.J., Hakstian, A.R,. & Hare, R.D. (1988). Factorstructure of the Psychopathy Checklist. Journal of Consulting and Clinical Psychology, 56, 741-747.Harpur, T.J., Hare, R.D., & Hakstian, A.R. (1989). Two-factorconceptualization of psychopathy: Construct validation andassessment implications. Psychological Assessment: A Journal of Consulting and Clinical Psychology, 1 6-17.Hart, S.D., Hare, R.D., & Harpur, T.J. (in press). ThePsychopathy Checklist: An overview for researchers andclinicians. In P. McReynolds & J. Rosen (Eds.), Advances inpsychological assessment (Vol. 8). New York: Plenum.Hemphill, J.F., Hart, S.D., & Hare, R.D. (1990). Substance use by criminal psychopaths. Manuscript submitted forpublication.Izard, C.E. (1971). The face of emotion. New York: Appleton-Century-Crofts.Izard, C.E. (1977). Human emotions. New York: Plenum Press.Izard, C.E., Kagan, J., & Zajonc, R.B. (1984). Emotions, cognition and behavior. Cambridge: Cambridge UniversityPressJames, W. (1884). What is an emotion? Mind, 9, 188-205.Jutai, J.W., Hare, R.D., & Connolly, J.F. (1987). Psychopathy-119-and event-related brain potentials (ERPs) associated withattention to speech stimuli. Personality and Individual Differences, 8, 175-184.Kappas, A., McHugo, G.J., & Lanzetta, J.T. (1989). Componentialcontrol and the modulation of emotional experience.Psychophysiology_, 26. 37.Kinsbourne, M. (1978). Evolution of language in relation tolateral action. In M. Kinsbourne (Ed.), Asymmetrical function of the brain (pp. 553-556). Cambridge, England:Cambridge University Press.Klorman, R., Weissberg, R., & Wiesenfeld, A. (1977). Individualdifferences in fear and autonomic reactions to affectivestimulation. Psychophysiology, 14, 45-51.Kolb, B. & Milner, B. (1981). Observations on spontaneousfacial expression after focal cerebral excisions and afterintracarotid injection of sodium amytal. Neuropsychologia,19, 505-514.Kosson, D.S., Smith, S.S., & Newman, J.P. (1990). Evaluatingthe construct validity of psychopathy on Black and Whitemale inmates: Three preliminary studies. Journal of Abnormal Psychology, 99, 250-259.Krause, R., Steimer, E., Sanger-Alt, C., & Wagner, G. (1989).Facial expression of schizophrenic patients and theirinteraction partners. Psychiatry, 52, 1-12.Lacey, J. (1967). Somatic patterning and stress: Some revisionsof activation theory. In N.H. Appley & R. Trumbell (Eds.),Psychological stress: Issues in research (pp. 14-42). NewYork: Appleton-Century Crofts.Lacey, J., & Lacey, R. (1974). Studies of heart rate and otherbodily processes in sensorimotor behavior. In R. Obrist, A.Black, & L. Dicara (Eds.), Cardiovascular psychophysiology(pp. 538-564). Chicago: Aldine.Laird, J.D. (1974). Self-attribution of emotion: The effects ofexpressive behavior on the quality of emotional experience.Journal of Personality and Social Psychology, 29, 475-486.Laird, J.D. (1984). The real role of facial response in theexperience of emotion: A reply to Tourangeau and Ellsworth,and others. Journal of Personality and Social Psychology,47, 909-917.Lang, P.J. (1978). Anxiety: Toward a psychophysiological-120-definition. In H.S. Akiskal & W.L. Webb (Eds.), Psychiatricdiagnosis: Explanation of biological predictors (pp. 365-389). New York: Spectrum.Lang, P.J. (1984). Cognition in emotion: Concept and action. InC. Izard, J. Kagan, & R. Zajonc (Eds.), Emotion, cognitionand behavior (pp. 196-226). NY: Cambridge University Press.Lang, P.J. (1985). The cognitive psychophysiology of emotion:Fear and anxiety. In A.H. Tuma & J.D. Maser (Eds.), Anxietyand the anxiety disorders (pp.131-170). Hillsdale, NJ:Erlbaum.Lang, P.J., Bradley, M.M., & Cuthbert, B.N. (1990). Emotion,attention, and the startle reflex. Psychological Review, 97,377-398.Lang, P.J., Ohman, A., & Vaitl, D. (1988). The International Affective Picture System [photographic slides]. Gainesville,FL: Center for Research in Psychophysiology, University ofFlorida.Larsen, R.J., Diener, E., & Emmons, R.A. (1986). Affectintensity and reactions to daily life events. Journal of Personality and Social Psychology,  51, 803-814.Lee, G.P., Loring, D.W., Meador, K.J., & Flanigin, H.F. (1988,January). Emotional reactions and behavioral complications following unilateral intracarotid sodium amytal injections.Paper presented at the annual meeting of the InternationalNeuropsychological Society, New Orleans.Leventhal, H. & Tomarken, A.J. (1986). Emotion: Today'sproblems. In M.R. Rosenzweig & L.Y. Porter (Eds.), Annual review of psychologv_, (Vol. 37, pp. 565-610). Palo Alto, CA:Annual Reviews, Inc.Levenson, R.W. (1988). Emotion and the autonomic nervoussystem: A prospectus for research on autonomic specificity.In H. Wagner (Ed.), Social psychophysiology: Theory andclinical applications (pp. 17-42). London: Wiley.Levenson, R.W., Cartensen, L.L., Friesen, W.V., & Ekman, P.(1991). Emotion, physiology, and expression in old age.Psychology and Aging, 6, 28-35.Levenson, R.W., Ekman, P., & Friesen, W.V. (1990). Voluntaryfacial action generates emotion-specific autonomic nervoussystem activity. Psychophysiology, 27, 363-384.Lindsley, D.B. & Wicke, J.D. (1974). The electroencephalogram:-121-Autonomous electrical activity in man and animals. In R.Thompson & M.N. Patterson (Eds.), Bioelectric recordingtechniques (pp.3-79). New York: Academic Press.Lykken, D.T. (1957). A study of anxiety in the sociopathicpersonality. Journal of Abnormal and Social Psychology, 55,6-10.Lykken, D.T. & Venables, P.H. (1971). Direct measurement ofskin conductance: A proposal for standardization.Psychophysiology_, 8, 656-672.Mandler, G. (1975). Mind and emotion. New York: Wiley.Manning, S.K. & Melchiori, M.P. (1974). Words that upset urbancollege students: Measured with GSRs and rating scales.Journal of Social Psychology, 94, 305-306.Matsumoto, D. (1987). The role of facial response in theexperience of emotion: More methodological problems and ameta-analysis. Journal of Personality and Social Psychology,52, 769-774.McCord, W. & McCord, J. (1964). The psychopath: An essay on thecriminal mind. Princeton, NJ: Van Nostrand.Meyers, M.B. & Smith, B.D. (1987). Cerebral processing ofnonverbal affective stimuli: Differential effects ofcognitive and affective sets on hemispheric asymmetry.Biological Psychology, 24, 67-84.Nauta. W.J.H. (1964). The problem of the frontal lobe: Areinterpretation. Journal of Psychiatric Research, 8, 167-187.Ogloff, J. & Wong, S. (1990). Electrodermal and cardiovascularevidence of a coping response in psychopaths. Criminal Justice and Behavior, 17, 231-245.Ouimette, P.C., Klein, D.N., Clark, D.C., & Margolis, E.T.(1992). Personality traits in offspring of parents withunipolar affective disorder: An exploratory study. Journal of Personality Disorders, 6, 91-98.Patrick, C.J., Bradley, M., & Lang, P.J. (1991). Emotion in thecriminal psychopath: Startle reflex modulation. Manuscriptsubmitted for publication.Patrick, C.J., Cuthbert, B.N., & Lang, P.J. (1990). Emotion inthe criminal psychopath: Fear imagery. Psychophysiology, 27,55.-122-Pennebaker, J.W. (1982). The psychology of physical symptoms.New York: Springer-Verlag.Pinel P. (1809). Traite medico-phyiosophlique sur l'alienationmentale. Paris: J. Ant. Brosson.Plutchik, R. (1980). Emotion: A psychoevoluntionary synthesis.New York: Harper and Row.Prichard, J.C. (1835). A treatise on insanity and other disorders affecting the mind. London: Sherwood, Gilbert-Piper.Raine, A., O'Brien, M., Smiley, N., Sacerbo, A., & Chan, C.(1990). Reduced lateralization in verbal dichotic listeningin adolescent population. Journal of Abnormal Psychology,99, 272-277.Ray, W.J. & Cole, H.W. (1985). EEG alpha activity reflectsattentional demands, and beta activity reflects emotionaland cognitive processes. Science, 228, 750-752Rime, B., Philippot, P., & Cisamolo, D. (1990). Social schemataof peripheral changes in emotion. Journal of Personality andSocial Psychology, 59, 38-49.Rinn, W.E. (1984). The neuropsychology of facial expression: Areview of the neurological and psychological mechanisms forproducing facial expressions. Psychological Bulletin, 95,52-77.Roberts, R.J. & Weerts, T.C. (1982). Cardiovascular respondingduring anger and fear imagery. Psychological Reports, 50,219-230.Robinson, R.G., Kubos, K.L., Starr, L.B., Rao, K., & Price,T.R. (1984). Mood disorders in stroke patients: Importanceof location of lesion. Brain, 107, 81-93.Rush, B. (1812). Medical enquiries and observations upon diseases of the mind. Philadelphia.Russell, R.A. (1980). A circumplex model of affect. Journal of Personality and Social Psychology_, 39, 1161-1178.Russell, J.A. (1989). Measures of emotion. Emotion: Theory, research, and experience, 4, 83-111.Russell, J.A., & Bullock, M. (1986). Fuzzy concepts and theperception of emotion in facial expressions. Social -123-Cognition, 4, 309-341.Russell, J.A. & Mehrabian, A. (1977). Evidence for a three-factor theory of emotions. Journal of Research inPersonality, 11, 273-294.Russell, J.A., Weiss, A., & Mendelsohn, G.A. (1989). Affectgrid: A single-item scale of pleasure and arousal. Journal of Personality and Social Psychology, 57, 493-502.Sackeim, H.A., Greenberg, M.S., Weiman, A.L., Gur, R.,Hungerbuhler, J.P., & Geschwind, N. (1982). Hemisphericasymmetry in the expression of positive and negativeemotions. Archives of Neurology, 39, 210-218.Sackeim, H.A., & Gur, R.C. (1983). Facial asymmetry and thecommunication of emotion. In J.T. Cacioppo & R.E. Petty(Eds.), Social psychophysiology (pp. 307-352). New York:Guilford Press.Sartory, G. (1983). The orienting response and psychopathology:Anxiety and phobias. In D. Siddle (Ed.), Orienting andhabituation: Perspectives in human research. Chichester:Wiley.Schachter, J. (1957). Pain, fear, and anger in hypertensivesand normotensives: A psychophysiological study.Psychosomatic Medicine, 19, 17-29.Schachter, J. & Singer, J.E. (1962). Cognitive, social, anddeterminants of emotional states. Psychological Review, 69,379-399.Schlosberg, H. (1954). Three dimensions of emotion.Psychological Review, 61, 81-88.Schmauk, F.J. (1970). Punishment, arousal and avoidancelearning in sociopaths. Journal of Abnormal Psychology, 76,325-335.Schwartz, G.E., Weinberger, D.A., & Singer, J.A. (1981).Cardiovascular differentiation of happiness, sadness, anger,and fear following imagery and exercise. Psychosomatic Medicine, 43, 343-364.Shaver, P., Schwartz, J., Kirson, D., & O'Connor, L. (1987).Emotion knowledge: Further exploration of a prototypeapproach. Journal of Personality and Social Psychology, 52,1061-1086.Siddle, D. A. & Trasler, G. B. (1981). The psychophysiology of-124-psychopathic behavior. In M.J. Christie & P.G. Mellett(Eds.), Foundations of psychosomatics (pp. 283-303). London:Wiley.Silberman, E.K., & Weingartner, H. (1986). Hemisphericlateralization of functions related to emotion. Brain and Cognition, 5, 322-353.Smith, C.A. & Ellsworth, P.C. (1985). Patterns of cognitiveappraisal in emotion. Journal of Personality and Social Psychology, 48, 813-838.Spielberger, C.D., Gorsuch, R.L., & Lushene, R.E. (1970).Manual for the State-Trait Anxiety Inventory (STAI). PaloAlto: Consulting Psychologist Press.Springer, S.P. & Deutsch, G. (1989). Left brain, right brain(3rd ed.). New York: Freeman and Co.Stemmler, G. (1989). The autonomic differentiation of emotionsrevisited: Convergent and discriminant validation.Psychophysiology_, 26, 617-632.Tassinary, L.G., Cacioppo, J.T., & Geen, T.R., (1989). Apsychometric study of surface electrode placements forfacial electromyographic recording: I. The corrugatorsupercilii and zygomaticus major muscle regions.Psychophysiology, 26, 1-16.Tomarken, A.J., Davidson, R.J., & Henriques, J.B. (1990).Resting frontal brain asymmetry predicts affective responsesto films. Journal of Personality and Social Psychology, 59,791-801.Tomkins, S.S. (1962a). The role of facial response in theexperience of emotion: A reply to Tourangeau and Ellsworth.Journal of Personality and Social Psychology, 40, 355-357.Tomkins, S.S. (1962b). Affect, imagery, consciousness: Thepositive affects. (Vol. 1), New York: Springer.Tucker, D.M. & Frederick, S.L. (1989). Emotion and brainlaterlization. In H. Wagner & T. Manstead (Eds.), Handbook of psychophysiology: Emotion and social behavior, (pp. 27-70). London: Wiley.Uijtdehaage, S.H.J. & Thayer, J.F. (1990). Patternclassification of emotions: Autonomic, somatic, and self-report measures. Psychophysiology, 27.Voglmaier, M.M. & Hakerem, G. (1989). Facial electromyography-125-(EMG) in response to facial expressions: Relation tosubjective emotional experience and trait affect.Psychophysiology, 26, 64.Vrana, S.R., Cuthbert, B.N., & Lang, P.J. (1989). Processingfearful and neutral sentences: Memory and heart rate change.Cognition and Emotion, 3, 179-195.Williamson, S., Crockett, D., Hurwitz, T., & Remick, R. (inpress). The processing of emotional information by depressedand organic patients. Archives of Clinical Neuropsychology.Williamson, S., Harpur, T.J., & Hare, R.D. (1991). Abnormalprocessing of affective words by psychopaths.Psychophysiology, 28, 260-273.Williamson, S., Harpur, T.J., & Hare, R.D. (August, 1990).Sensitivity to emotional polarity in psychopaths. Paperpresented at the meeting of the American PsychologicalAssociation, Boston.Winton, W.M., Putnam, L.E., & Krauss, R.M. (1984). Facial andautonomic manifestations of the dimensional structure ofemotion. Journal of Experimental Social Psychology, 20, 195-216.Wong, S. (1984). Criminal and institutional behaviors ofpsychopaths. Programs Branch Users Report. Ottawa, Ontario,Canada: Ministry of the Solicitor-General of Canada.Zajonc, R.B. (1984). On the primacy of affect. AmericanPsychologist, 39, 117-123.Zajonc, R.B. (1985). Emotion and facial efference: A theoryreclaimed. Science, 228, 15-21.-126-Appendix AEmotion: AmusementFilm: "Robin Williams at the Met"Scene: Skit about being stopped by police and drug useEmotion: HappyFilm: "An Officer and A Gentleman"Scene: This clip begins as Zac (an officer from nearby navybase), in his dress whites, is striding thought thepaper factory where his girlfriend Paula works. Heattracts the attention of different factory workers,including Paula's mother, who follow him. Zac comes upfrom behind Paula and gives her a kiss on the back ofthe neck. Startled, Paula turns around, sees that it isZac and they embrace. Zac sweeps Paula up in his armsand carries her out of the factory. Paula takes Zac'scap and places it on her head and the film freezes onthat moment and the clip ends.Emotion: SadFilm: "Kramer vs. Kramer"Scene: The clip begins with Kramer walking through a park withhis young son. He tells his son that the court hasawarded custody to the boy's mother, and the boy mustmove to his mother's house. The boy looks confused andbegins to cry, telling his father that he does not wantto go. The clip ends with Kramer hugging his son and asboy cries.Emotion: FearFilm: "Platoon"Scene: The clip begins with an American soldier left on lookout duty in the midst of a Vietnam jungle. It is raininglightly and dark and the other soldiers are all asleep.The soldier looks up and sees the Viet Kong approaching.There is a close up of the soldier face as his eyes dartbetween his machine gun lying beside him, his sleepingcompanions, and the Viet Kong as they approach. The clip-127-ends just before the soldier grabs his gun.Emotion: AngerFilm: "Witness"Scene: John Book has been in town with the Amish making phonecalls. The clip begins as he rides with the old Amishfarmer back to the farm. The Amish farmers in the wagonahead of them are harassed by a group of localteenagers. This angers Book and he goes to the defenseof the Amish farmers. The clip ends as one teenagerinsults Book.Emotion: DisgustFilm: "Pink Flamingos"Scene: The clip beings as an overweight transvestite (Devine)is walking along a New York sidewalk. Devine sees apoodle defecating and goes over and picks up theexcrement. There is a close up of Devine's mouth as heeats the excrement. The clip ends with him spitting someof it out.Appendix BRATING SCALEFilm #:Subject #:-128-Please rate your emotional state while you watched the film.Make sure that you complete each of the rating scales. Pleasecircle the number that represents how you felt during the film.HAPPY:O 1^2^3^4^5^6^7^8^91^ 1^1^ 1^ 1^ 1^ 1^ 1^ 1^ 1^Did not feel at all Felt very stronglySAD:O 1^2^3^4^5^6^7^8^91^1^ 1^ 1^1^ 1^ 1^ 1^ 1 1Did not feel at all Felt very stronglyCONTENT:O 1^2^3^4^5^6^7^8^91^1^1^1^1^1^1^1 1^1Did not feel at all Felt very stronglySURPRISED:O 1^2^3^4^5^6^7^8^91^ 1^ 1^ 1^ 1^ 1^1^ 1^ 1^ 1Did not feel at all Felt very stronglyAFRAID:O 1^2^3^4^5^6^7^8^91^ 1^ 1^ 1^ 1^1^ 1^ 1^ 1^ 1Did not feel at all Felt very stronglyDISGUSTED:O 1^2^3^4^5^6^7^8^91^ 1^ 1^1^1^ 1^ 1^ 1^ 1^ 1Did not feel at all Felt very stronglyINTERESTED:O 1^2^3^4^5^6^7^8^91^ 1^ 1^ 1^ 1^ 1^ 1^ 1^ 1^1Did not feel at all Felt very stronglyAMUSED:O 1^2^3^4^5^6^7^8^91^ 1^ 1^ 1^ 1^ 1^ 1^ 1^ 1^ 1Did not feel at all Felt very stronglyANGRY:O 1^2^3^4^5^6^7^8^91^ 1^ 1^ 1^ 1^ 1^ 1^ 1^ 1^ 1Did not feel at all Felt very strongly


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items