Download - Temporal lobe abnormalities in semantic processing by criminal psychopaths as revealed by functional magnetic resonance imaging

Psychiatry Research: Neuroimaging 130(2004) 297–312

0925-4927/04/$ - see front matter� 2004 Elsevier Ireland Ltd. All rights reserved.doi:10.1016/j.pscychresns.2004.02.002

Temporal lobe abnormalities in semantic processing by criminalpsychopaths as revealed by functional magnetic resonance imaging

Kent A. Kiehl *, Andra M. Smith , Adrianna Mendrek , Bruce B. Forster ,a,b, c d e

Robert D. Hare , Peter F. Liddlef g

Institute of Living, 200 Retreat Ave, Hartford, CT 06106, USAa

Department of Psychiatry, Yale University, School of Medicine, New Haven, CT, USAb

Department of Diagnostic Imaging, The Ottawa Hospital, Ottawa, ON, Canadac

Department of Psychiatry, University of Montreal, Montreal, QC, Canadad

Department of Radiology, University of British Columbia, Vancouver, BC, Canadae

Department of Psychology, University of British Columbia, Vancouver, BC, Canadaf

Division of Psychiatry, School of Community Health Sciences, University of Nottingham, Nottingham, UKg

Abstract

We tested the hypothesis that psychopathy is associated with abnormalities in semantic processing of linguisticinformation. Functional magnetic resonance imaging(fMRI) was used to elucidate and characterize the neuralarchitecture underlying lexico-semantic processes in criminal psychopathic individuals and in a group of matchedcontrol participants. Participants performed a lexical decision task in which blocks of linguistic stimuli alternatedwith a resting baseline condition. In each lexical decision block, the stimuli were either concrete words andpseudowords or abstract words and pseudowords. Consistent with our hypothesis, psychopathic individuals, relativeto controls, showed poorer behavioral performance for processing abstract words. Analysis of the fMRI data for bothgroups indicated that processing of word stimuli, compared with the resting baseline condition, was associated withneural activation in bilateral fusiform gyrus, anterior cingulate, left middle temporal gyrus, right posterior superiortemporal gyrus, and left and right inferior frontal gyrus. Analyses confirmed our prediction that psychopathicindividuals would fail to show the appropriate neural differentiation between abstract and concrete stimuli in the rightanterior temporal gyrus and surrounding cortex. The results are consistent with other studies of semantic processingin psychopathy and support the theory that psychopathy is associated with right hemisphere abnormalities forprocessing conceptually abstract material.� 2004 Elsevier Ireland Ltd. All rights reserved.

Keywords: Psychopathy; Antisocial personality; Neuroimaging; Limbic system; fMRI; Lexical decision; Language; Concretewords; Abstract words

*Corresponding author. Tel.:q1-860-545-7385; fax:q1-860-545-7066.E-mail address: [email protected](K.A. Kiehl).

298 K.A. Kiehl et al. / Psychiatry Research: Neuroimaging 130 (2004) 297–312

1. Introduction

Psychopathy is a complex personality disorderof unknown etiology. Psychopathic individuals areglib, superficial, impulsive, callous, and lackempathy, guilt and remorse for their depredations.For many years, research on psychopathy hasfocused on elucidating and characterizing the roleemotional processes play in the disorder. In gen-eral, these studies have shown that psychopathicindividuals have difficulty accessing and under-standing the connotative aspects of affective stim-uli (Day and Wong, 1996; Intrator et al., 1997;Kiehl et al., 1999a; Louth et al., 1998; Williamsonet al., 1991). These difficulties are particularlyevident during language tasks(Williamson et al.,1991). However, recent developments in theoryand research suggest that the cognitive deficitsobserved in psychopathic individuals are not lim-ited to the affective domain(Kiehl et al., 1999a;Newman et al., 1997). One cognitive realm inwhich psychopathic individuals appear to showabnormalities for both emotional and neutral stim-uli is in semantic processing(Gillstrom, 1994;Hare and Jutai, 1988; Kiehl et al., 1999a; William-son et al., 1991). In particular, psychopathic indi-viduals appear to have difficulty integratingcontextual information regarding conceptuallyabstract material(Hare and Jutai, 1988; Kiehl etal., 1999a).Early empirical research sought to elucidate

cognitive impairments in psychopathy by examin-ing the relationship between psychopathy and hem-ispheric lateralization for language stimuli(Dayand Wong, 1996; Hare, 1979; Hare and Jutai,1988; Hare and McPherson, 1984; Jutai et al.,1987; Raine et al., 1990). The impetus for explor-ing this relationship arose from clinical observa-tions of psychopathic individuals. Numerousclinicians noted that the actual behavior of psycho-pathic individuals is often strikingly inconsistentwith their verbalized reports(Cleckley, 1976;McCord and McCord, 1964), leading some tospeculate that psychopathy is associated with lan-guage abnormalities(Flor-Henry, 1972). Subse-quent research found that abnormalities inlanguage processes are most prevalent when psy-chopathic individuals are required to perform tasks

involving semantic processing(Hare, 1979; Hareand Forth, 1985; Hare and McPherson, 1984).More recently, Kiehl et al.(1999a) observed

that psychopathic individuals performed morepoorly than control participants during a task(Task2 in their study) that required classifying wordstimuli as either concrete(e.g. table) or abstract(e.g. justice). Previous studies have demonstratedthat healthy subjects respond more quickly andaccurately to concrete words than to abstract wordsin lexical decision and concreteyabstract discrimi-nations tasks(Day, 1977; Holcomb et al., 1999;James, 1975; Kounios and Holcomb, 1994; Krolland Merves, 1986). These data led to theories thatthe cognitive operations and, by inference, theneural systems involved in processing concreteand abstract words are disassociated(Holcomb etal., 1999; Kiehl et al., 1999b; Paivio, 1986, 1991;Schwanenflugel et al., 1988; Schwanenflugel andStowe, 1989). Consistent with the hypothesis thatpsychopathic individuals have difficulty processingabstract words, psychopathic individuals mademore errors when they had to classify word stimulias abstract. Kiehl et al.(1999a) also recordedevent-related potential(ERP) data during theirprocedure and observed that psychopathic individ-uals failed to show the normal ERP differentiationbetween concrete and abstract words(Tasks 1 and2). In non-criminals and in criminal non-psycho-pathic individuals, concrete words elicit greaterERP negativity in the 300–800 ms window thando abstract words(Kounios and Holcomb, 1994;Paller et al., 1987). This latter effect is strongestat fronto–temporal electrode sites, suggesting fron-tal–temporal involvement in the differentiation ofconcrete and abstract words. Given that the ERPdifferentiation between concrete and abstract wordsappears to be most robust 300–500 ms post-stimulus, it has been argued that this effect is dueto modulations of semantic generators, an effectthat is also believed to contribute the N400 poten-tial typically observed in semantic word and sen-tence processing tasks(Kutas and Hillyard, 1980,1983, 1984). Recent evidence suggests that theamplitude of the N400 may reflect processes relat-ed to the integration of a word within an ongoingcognitive context(Holcomb, 1993). On the basisof this interpretation, it would appear that psycho-

299K.A. Kiehl et al. / Psychiatry Research: Neuroimaging 130 (2004) 297–312

paths differ from others in the degree and extentof cognitive processes required to perform lan-guage tasks.In addition to the failure to show appropriate

ERP differentiation between concrete and abstractwords, the psychopathic individuals’ ERPs to allword stimuli contained a large fronto–central neg-ativity in the 300–500 ms(N350) post-stimulustime window. The N350 of the psychopathic indi-viduals was present during a concreteyabstractlexical decision task(Task 1), concreteyabstractdiscrimination task(Task 2) and an emotionalpolarity discrimination task(Task 3). Kiehl et al.(1999a) suggested that the reduced behavioral andERP differentiation between concrete and abstractwords and the psychopathic individuals’ abnormalN350 may reflect abnormal semantic processing.Evidence for abnormalities in language process-

ing also comes from analyses of the speech ofpsychopathic individuals. Gillstrom and Hare(1988) found that psychopathic individuals usemore ‘beats’, defined as language-related handgestures that do not reflect the semantic content ofspeech, than do control participants. This findingwas interpreted as evidence that psychopathic indi-viduals compartmentalize their speech into smaller,more discrete units than others. Psychopathic indi-viduals also do not differentiate, in voice analyses,between affective and neutral words(Louth et al.,1998).Thus, on balance, there appears to be strong

evidence for abnormalities in semantic processingin psychopathy. In particular, these abnormalitiesappear to be strongest when accessing right hem-isphere resources to process conceptually abstractinformation regardless of whether the stimuli haveemotional connotations or are affectively neutral.

1.1. Role of the right hemisphere in languageprocessing

Since the classic work of Broca and Wernicke,language functions have generally been assumedto reside in left hemisphere neural systems. How-ever, accumulating evidence from convergingmodalities suggests that the right hemisphere mayplay an important role in language processing. Forexample, patients with right hemisphere brain dam-

age show deficits in verbal reasoning ability(Car-amazza et al., 1976), interpretation of verbalhumor(Brownell et al., 1983), understanding pros-ody of speech(Ross, 1981) and comprehensionand production of the connotative meanings ofwords and figures of speech(Gardner and Denes,1973; Winner and Gardner, 1977).More recent evidence from neuroimaging stud-

ies has further implicated the right hemisphere inlanguage functions. Beauregard et al.(1997), usingPET, observed that passive viewing of abstractwords relative to baseline produced neuronal acti-vation in the right inferior frontal gyrus. It isimportant to note that this right hemisphere acti-vation was found for abstract word processing butnot for concrete words minus baseline comparisonsor for emotional words minus baseline compari-sons(Beauregard et al., 1997). D’Esposito et al.(1997) also found activation of the right superiorfrontal gyrus during passive viewing of abstractwords, when compared with active processing ofconcrete words. Similarly, several areas in the righthemisphere, including prefrontal cortex and middletemporal gyrus, have been found to be activatedduring comprehension of metaphors(Bottini et al.,1994). More recently, Kiehl et al.(1999b) foundthat a region in the right anterior superior temporalgyrus extending into the inferior frontal gyrus wasmore strongly activated for processing abstractstimuli than for concrete stimuli during a lexicaldecision task. Thus, converging evidence suggeststhat the right hemisphere may play a special rolein interpreting the abstract representations of lan-guage confirming the argument put forth by Bee-man et al.(1994), who suggested that the linguisticstrength of the right hemisphere is its ability tobring together semantic associations.The purpose of the present study was to use

functional magnetic resonance imaging(fMRI) toelucidate the neural architecture underlying lexico-semantic processing in criminal psychopathic indi-viduals during performance of a concreteyabstractlexical decision task. Given that psychopathic indi-viduals have difficulty processing abstract infor-mation(Hare and Jutai, 1988; Kiehl et al., 1999a),we hypothesized that we would observe reducedneural differentiation between abstract and con-crete stimuli in the right hemisphere for psycho-


Table 1Demographic data for the criminal psychopaths and control participants

Group Age Years of formal NART score Quick test Hollingshead parentalMean(S.D.) education Mean(S.D.) Mean(S.D.) social position index

Mean(S.D.) Mean(S.D.)

Controls 27.9(5.0) 12.4 (0.74) 111.82(7.0) 104.75(5.4) 4.25 (3.4)Psychopaths 33.9(7.6) 9.9 (3.5) 111.19(7.5) 102.75(9.9) 4.25 (1.4)

pathic individuals relative to control participants.We specifically hypothesized that this effect wouldbe present in the right anterior superior temporalgyrus (Kiehl et al., 1999b). We also expectedpsychopathic individuals to be slower and lessaccurate than control individuals in processingabstract words(Hare and Jutai, 1988; Kiehl et al.,1999a).

2. Methods

2.1. Participants

Criminal psychopathic individuals(ns8; allmale) were inmates from a maximum-securityprison located in Abbotsford, British Columbia,Canada. Psychopathic inmates were transported tothe University of British Columbia Hospital’s MRIunit by the Correctional Services of CanadaRegional Escort Team. Healthy control participants(ns8; all male) were recruited from the generalpopulation. All participants were free from anyhistory of head injury or psychotic illness(in selfand first-degree relatives), were right-handed(Annett, 1970), and spoke English as their firstlanguage. All participants had normal or correctedto normal vision. No participants met the criteriafor substance abuse according to DSM-IV criteria(American Psychiatric Association, 1994) withinthe last 6 months. There were no group differencesin age, parental socioeconomic status, educationlevel, or IQ, measured with the National AdultReading Test(NART) (Sharpe and O’Carroll,1991) and Quick Test(Ammons and Ammons,1962; Ammons and Ammons 1979a; Ammons andAmmons, 1979b). These data are summarized inTable 1.Two clinicians used the Hare Psychopathy

Checklist-Revised(PCL-R) to assess psychopathy

(Hare, 1991). The PCL-R is a reliable and validmeasure of psychopathy(Fulero, 1996; Hare,1980, 1991; Hare et al., 1990a,b, 1991; Harpur etal., 1988; Harpur and Hare, 1994; Harpur et al.,1989; Hart and Hare, 1989; Hart et al., 1992).Each of the 20 items on the PCL-R is scored on a3-point scale(0–2) according to the extent towhich it applies to the inmate. All inmates had aPCL-R score above 28(scores range from 0 to40) on the PCL-R(mean 32.8, S.D. 2.9), whichis above the mean score on the PCL-R(23.6, S.D.7.9) listed in the test manual for normative dataof 1192 prison inmates(Hare, 1991). None of thecontrol participants had a criminal history.

2.2. Materials

Stimulus words(3 to 8 letters in length) wereselected from the word norms of Toglia and Battig(1978) and were either concrete or abstract. Wordsrated as more than 0.75 standard deviations aboveor below the mean concreteness rating containedin the word norms were defined as concrete andabstract, respectively. The word lists for each taskdid not differ in word frequency or length(Francisand Kucera, 1982). Furthermore, only affectivelyneutral words(at or within one standard deviationof the mean pleasantness rating given in Togliaand Battig, 1978) were selected in order to elimi-nate any confound of emotionality. We developedsets of pronounceable pseudowords by selectivelyaltering one letter of each of the concrete andabstract words.

2.3. Procedure

Stimuli were presented to the participant by acomputer-controlled projection system that deliv-ered a visual stimulus to a rear-projection screen


located at the entrance to the magnet bore. Theparticipant viewed this screen through a mirrorsystem attached to the top of the head coil. Thescanning room and magnet bore were darkened toallow easy visualization of the experimentalstimuli.Two stimulus runs were presented, each consist-

ing of a series of four 30-s lexical decision blocksalternating with a baseline session. Each run wasprefaced by a 10-s rest session that was collectedto allow T effects to stabilize. These images were1

not included in any subsequent analyses. Duringthe lexical decision blocks, 15 letter stimuli(350-ms duration; 1650-ms interstimulus interval) wererandomly presented. All stimuli were presented incapital letters. During the baseline period, thecharacters ‘*****’ were continuously displayedfor 29.5 s(500-ms interstimulus interval). Stimu-lus runs were balanced such that equal proportionsof word and pseudoword stimuli were presented.Lexical decision blocks consisted of either con-crete words and associated pseudowords or abstractwords and associated pseudowords. The word andits associated pseudoword did not appear duringthe same run. Concrete and abstract lexical deci-sion blocks were presented in random order. Theparticipant was not informed of the concreteyabstract manipulation. Participants were instructedto respond with one hand each time the letterstimuli presented formed a real English speakingword and to respond with their other hand if theletter stimuli did not form an English speakingword. The hand used to make the response wascounterbalanced across participants. Reaction timeand accuracy were equally stressed. A commer-cially available MRI-compatible fiber-opticresponse device(Lightwave Medical, Vancouver,BC) was used to acquire behavioral responses. Acustom visual(and auditory) presentation package(VAPP; http:yynilab.psychiatry.ubc.cayvapp) wasused to control the timing of the experimentalstimuli and recording of all behavioral data. Beforeentry into the scanning room, each participantperformed a practice block of lexical decisions,repeated twice, to ensure he understood the instruc-tions. None of the stimuli used in the practiceblocks were used in the fMRI session.

Reaction times were computed on trials forwhich the participant responded correctly within1500 ms post-stimulus. Errors included incorrectresponses within 1500 ms post-stimulus or anyresponse with a latency of greater than 1500 msfollowing the onset of the target stimulus. Weperformed 2 Group(psychopath vs. control)=2Word (concrete=abstract)=2 Lexical (realword=pseudoword) repeated measures analysesof variance(ANOVAs) on the reaction time andaccuracy data. Planned comparisons were thenperformed to assess whether psychopathic individ-uals would respond slower and be less accuratethan control participants for the lexical decisionsfor abstract words.

2.4. Image acquisition

Functional data were collected using a clinicalGE 1.5-T whole body system fitted with a Horizonecho-speed upgrade. The participant’s head wasfirmly secured using a custom head holder, andexternal references were used to position the ante-rior commissure-posterior commissure(AC-PC)line at right angles to the slice-select gradient.Conventional spin echoT weighted sagittal local-1

izers were acquired to confirm external land-marking. Functional image volumes were collectedwith a gradient-echo sequence(TRyTE 2500y50ms, flip angle 908, FOV 24=24 cm, 64=64matrix, 62.5 kHz bandwidth, 3.75=3.75 mm inplane resolution, 4-mm slice thickness, 23 slices)effectively covering the entire brain(except forthe inferior cerebellum).

2.5. Image processing

Functional images were reconstructed offline,and the two runs were separately realigned usingthe procedure of Friston et al.(1995b) as imple-mented in the Statistical Parametric Mapping(SPM96, Wellcome Department of Cognitive Neu-rology). Translation and rotation corrections didnot exceed 2.5 mm and 2.58, respectively, for anyof the participants. A mean functional image vol-ume was constructed for each participant for eachrun from the realigned image volumes. This meanimage volume was then used to determine para-


Table 2Behavioral data for the criminal psychopaths and control participants for the concreteyabstract lexical decision task

Reaction times(milliseconds) Control participants PsychopathsMean(S.D.) Mean(S.D.)

Concrete words 608.5(65.7) 678.6(52.0)Abstract words 640.6(80.2) 714.5(59.2)Pseudoconcrete stimuli 709.2(81.2) 845.6(113.9)Pseudoabstract stimuli 700.5(86.3) 833.9(93.1)

Percentage correctConcrete words 94.3(3.5) 96.2 (3.4)Abstract words 89.5(3.6) 88.5 (11.9)Pseudoconcrete stimuli 95.0(3.3) 91.3 (5.7)Pseudoabstract stimuli 93.6(3.7) 85.0 (9.9)

meters for spatial normalization into the modifiedTalairach space employed in SPM96 using bothaffine and non-linear components(Friston et al.,1995a). In this space, coordinates are expressedrelative to a rectangular coordinate frame with theorigin at the midpoint of the anterior commissureand they-axis passing through the posterior andanterior commissures. The normalization parame-ters determined for the mean functional volumewere then applied to the corresponding functionalimage volumes for each participant.Adjusted mean functional images were created

for the lexical decision blocks and rest session. Inthe computation of these adjusted mean images, atemporal delay of 6 s was incorporated to accountfor the relatively slow onset of the hemodynamicresponse. These adjusted mean images were thensmoothed with a 10=10=10 mm Gaussian kernel.The smoothed adjusted mean images were enteredinto a two-stage analysis.(Note: It is only thesecond stage that treats participant as a randomeffect.) In the first stage, we performed a confirm-atory fixed effect analysis in the control partici-pants by comparing the concrete and abstractstimuli vs. the rest condition. These latter analyseswere performed to determine whether we couldreplicate the results of our previous study(Kiehlet al., 1999b) in this new sample of healthyparticipants. We also performed an identical fixedeffect analysis in the psychopathic group.In the second stage we performed an independ-

ent samples 2 Group(psychopathy vs. control) t-test (14 d.f.) on the difference image of the

abstract stimuli minus the concrete stimuli. Herewe tested our hypothesis that control participantswould show greater abstract than concrete differ-entiation than psychopathic individuals(i.e.Group=Condition interaction).

3. Results

3.1. Behavioral data

Consistent with our hypothesis, psychopathicindividuals were significantly slower than controlparticipants for processing abstract wordswplannedcomparison,F(1,14) 4.40,P-0.05x. Control par-ticipants were, in general, faster to respond thanwere the psychopathic individualswmain effect ofGroup, F(1,14)s9.17, P-0.009x. Post hoc testsalso revealed that psychopaths were slower torespond to concrete words than control partici-pants. There were no group differences in accuracyfor real word stimuli. However, psychopathic indi-viduals were less accurate than control participantsfor the pseudoword stimuliwGroup=Lexical inter-action,F(1,14)s8.99,P-0.0096x. Summary sta-tistics of the behavioral data are presented in Table2.Overall, responses to real word stimuli were

faster than those to pseudoword stimuliwmaineffect of Lexical, F(1,14)s32.62, P-0.0001x.Concrete words were responded to more quicklythan abstract wordswLexical=Word interaction,F(1,14)s10.59, P-0.0058x. Concrete stimuli(words and pseudowords) were classified more


accurately than abstract stimuliwmain effect ofWord, F(1,14)s7.94,P-0.0137x.

3.2. Imaging data

Illustrations of the areas of activation for theconcrete stimuli vs. baseline and abstract stimulivs. baseline comparisons for the control partici-pants and psychopathic individuals are illustratedin Fig. 1A and B, respectively.The psychopathic individuals showed a very

similar pattern of activation to that of the controlparticipants for two comparisons of the wordstimuli vs. baseline(see Tables 3 and 4). For thepsychopathic individuals, activation for both com-parisons was observed bilaterally in the superiorparietal lobules, anterior cingulate, inferior frontalgyrus, insula, precuneus, bilateral fusiform gyrus,left middle temporal gyrus, and right posteriorsuperior temporal gyrus. However, no significantactivation was observed in the right anterior supe-rior temporal gyrus for the psychopathic individ-uals for the abstract vs. baseline comparison.In the control participants, the pattern of acti-

vation for the concrete stimuli vs. baseline andabstract stimuli vs. baseline largely confirmedprevious results using this task(Kiehl et al.,1999b). In both studies, concrete stimuli andabstract stimuli(relative to the baseline condition)elicited activation in the bilateral superior parietallobules, anterior cingulate, inferior frontal gyrus,insula, precuneus, bilateral fusiform gyrus, leftmiddle temporal gyrus and right posterior superiortemporal gyrus(see Tables 3 and 4). In addition,significant activation was found in bilateral pre-motor cortex consistent with the fact that thelexical decision was indicated with a button pressusing either the right or the left hand. In thecontrol participants, we observed significant acti-vation bilaterally in the thalamus for both of theselatter comparisons. In our previous study, activa-tion of the thalamus only reached trend levels(seeTables 3 and 4). Comparison of the abstract stimulivs. baseline revealed a very similar pattern ofactivation to the concrete stimuli vs. baselinecomparison. As in our previous study, activationin the right superior temporal gyrus was observedfor the abstract stimuli vs. baseline comparison

that was not observed for the concrete stimuli vs.baseline comparison.In the second stage analyses, as predicted, con-

trol participants produced greater activation forabstract compared to concrete stimuli than did thepsychopathic individuals in the right anterior supe-rior temporal gyrus (Talairach coordinates,52,15,y10; z-score 3.35,P-0.001; see Figs. 2and 3). There was also a trend for greater activa-tion during the processing of abstract words rela-tive to that for concrete words for controlparticipants than psychopathic individuals in theright lateral frontal cortex(Talairach coordinates,34,11,32,z-score 4.28,P-0.0001, uncorrected formultiple comparisons).

4. Discussion

This study was designed to elucidate the abnor-mal functional neural architecture underlying lex-ico-semantic processing in psychopathy. Consistentwith previous research, psychopathic individualsperformed more poorly, manifested as slower reac-tion times, than control participants, when process-ing abstract word stimuli(Hare and Jutai, 1988;Kiehl et al., 1999a). Psychopaths also respondedless accurately than control participants in classi-fying pseudoword stimuli.In general, the pattern of neural activation asso-

ciated with processing concrete and abstract stimuliwas similar in the psychopathic individuals andcontrol participants. However, psychopathic indi-viduals showed clear deficits in activating the rightanterior superior temporal gyrus and surroundingcortex for processing abstract stimuli, and theyfailed to show the appropriate neural differentiationin this region for abstract and concrete stimulirelative to control participants.These data support the hypothesis that there is

an abnormality in the function of the right anteriorsuperior temporal gyrus in psychopathy. Accumu-lating behavioral and ERP data suggest that psy-chopathic individuals have difficulty processingabstract words(current data; Kiehl et al., 1999a),performing abstract categorization tasks(Hare andJutai, 1988), understanding and interpreting meta-phors (Gillstrom, 1994), processing affectivelyvalenced word(Day and Wong, 1996; Intrator et


Fig. 1. (A) Cortical surface rendering of the areas in which control participants(left) and criminal psychopaths(right) showsignificantly greater activation for processing of concrete stimuli relative to baseline. Locations and labels of areas of activation arelisted in Table 1. The displayed area of activation is thresholded at az-score of 3.0 or greater.(B) Cortical surface rendering of theareas in which control participants(left) and criminal psychopaths(right) show significantly greater activation for processing ofabstract stimuli relative to baseline. Locations and labels of areas of activation are listed in Table 2. The displayed area of activationis thresholded at az-score of 3.0 or greater.

al., 1997; Kiehl et al., 1999a; Williamson et al.,1991) and speech(Louth et al., 1998; Williamsonet al., 1990) stimuli, and resolving action in spokennarratives (Brinkley et al., 1999). Converging

evidence from patients with brain damage andneuroimaging studies suggest that the right hemi-sphere is involved with processing and categoriz-ing abstract stimuli (Beauregard et al., 1997;


Table 3Summary of the significant areas of activation for the comparison of the concrete stimuli vs. the baseline condition

Region Talairach Kiehl et al., 1999b Control participants Psychopathiccoordinates z-score z-score(Talairach individuals

x y zcoordinates:x,y,z) z-score(Talairach

coordinates:x,y,z)

Frontal lobe1. R insula 41 26 0 4.26a 5.69*** (34,26,0) 5.13*** (34,22,0)2. L inferior frontal gyrus y49 8 32 7.68*** 6.88***(y45,4,32) 6.84*** (y49,15,24)3. R inferior frontal gyrus 52 4 32 7.68*** 6.21***(44,8,24) 6.98*** (45,11,24)4. Cingulate gyrus 0 11 40 7.65*** 6.84***(4,8,48) 7.06*** (y4,11,50)5. L middle frontal gyrus y34 0 60 7.10*** 5.72***(y34,y4,64) 6.34*** (y44,y4,50)6. R middle frontal gyrus 38 4 52 6.91*** 4.99**(34,y4,56) 6.97*** (41,y4,48)7. L inferior frontal gyrus y30 0 44 5.80*** 6.24***(y34,y8,48) 6.34*** (y45,y4,48)8. L insula y30 34 4 5.36** 6.36**(y34,24,4) 6.51*** (y38,22,0)9. L Insula y34 22 8 4.81* 6.36***(y34,24,4) 6.51*** (y38,22,0)

Parietal lobe10. R superior parietal lobule 30 y60 48 7.00*** 5.11**(34,y56,40) 5.74*** (34,y52,44)11. L inferior parietal lobule y30 y52 40 7.58*** 5.52***(y26,y56,48) 5.19**(y25,y50,40)12. L inferior parietal lobule y45 y34 48 4.55** 4.64*(y45,y34,40) 4.75*(y45,y34,44)

Temporal lobe13. L middle temporal gyrus y52 y52 4 5.46** 5.49***(y41,y71,0) 4.64*(56,y49,4)14. R superior temporal gyrus y56 y38 16 5.47** 3.99(ns)(60,y38,20) ns

Occipital lobe15. R fusiform gyrus 40 y74 y12 5.92*** 4.57*(49,y71,y4) 4.70*(y38,y75,y4)16. L fusiform gyrus y41 y60 y12 7.97*** 5.49*** (y41,y71,5) 4.69*(y38,y56,y4)

Talairach coordinates andz-scores are given from our previous study(Kiehl et al., 1999b). Control participants and psychopathicindividuals’ z-scores and Talairach coordinates are listed for the same anatomical locations as in Kiehl et al., 1999b(at or withinon smoothing element, 12=12=12 mm). Lsleft; Rsright. denotes az-score in Kiehl et al.(1999b) that was not reported in thea

original study because the significance level only reached trend levels.(P-0.15).Note: *** P-0.001, **P-0.01, *P-0.05, nssnon-significant.

Beeman et al., 1994; D’Esposito et al., 1997),accessing abstract representations of metaphors(Bottini et al., 1994; Gardner and Denes, 1973;Winner and Gardner, 1977), and processing affec-tive valence, intonation, and prosody of speech(Ross, 1981). Taken together, these data providesupport for the hypothesis that psychopathy isassociated with right hemisphere abnormalities ingeneral, anterior superior temporal gyrus in partic-ular. We note, however, that the present data donot address whether psychopathy is associated withabnormalities in the left hemisphere. The hypoth-eses in the present experiment are only concernedwith the function of the right anterior superiortemporal gyrus. Future research is needed beforeany strong conclusions can be reached regardingunilateral deficits in psychopathy.

It is important to note that the abnormalitiesobserved in the psychopathic individuals do notappear to be due to any gross structural brainpathology. High-resolution structural MRI scanswere collected in all participants, and none hadany overt brain pathology. It is possible that moredetailed analyses of the structural MRI data mayreveal subtle structural brain abnormalities in thepsychopathic offenders, but at this time there is noevidence to support the view that the cognitiveabnormalities observed in the present study aredue to structural brain damage. In other words, itappears that the abnormalities observed in thepsychopathic individuals in the present study arefunctional, rather than structural, in nature.Currently, abnormalities in processing abstract

material have only been characterized in samples


Table 4Summary of the significant areas of activation for the comparison of the abstract stimuli vs. baseline condition

Region Talairach Kiehl et al., 1999b Control participants Psychopathiccoordinates z-score z-score(Talairach individuals

x y zcoordinates:x,y,z) z-score(Talairach

coordinates:x,y,z)

Frontal lobe1. L inferior frontal gyrus y49 8 32 7.93*** 7.27***(y49,8,32) 6.93*** (y49,15,24)2. R inferior frontal gyrus 52 8 28 7.72*** 7.16***(45,8,28) 7.21*** (45,11,24)3. Cingulate gyrus 0 11 40 7.97*** 7.16***(y8,8,48) 7.14*** (y4,11,52)4. R middle frontal gyrus 38 4 52 7.26*** 6.03***(34,y4,52) 7.09*** (41,y4,52)5. L middle frontal gyrus y34 0 56 7.05*** 6.94***(y34,y8,52) 5.28**(y38,y11,52)6. L inferior frontal gyrus y30 0 44 6.83*** 6.94***(y34,y8,52) 5.28**(y38,y11,52)7. L inferior frontal gyrus y56 19 16 6.35*** 7.13***(y52,11,24) 4.91**(y64,15,12)8. L insula y30 y34 4 6.04*** 6.95***(y38,25,0) 5.84*** (34,24,0)9. R inferior frontal gyrus 41 26 0 4.64* 6.68***(34,22,0) 5.84*** (34,22,0)

Parietal lobe10. L parietal lobe y41 y34 44 6.23*** 5.73***(y45,y34,36) 4.86**(y45,y34,40)11. L inferior parietal lobule y30 y52 40 7.66*** 6.28***(y26,y52,44) 5.10**(y25,y52,44)12. R superior parietal lobule 30y56 48 7.59*** 6.06***(34,y52,40) 4.96**(34,y52,40)

Temporal lobe13. R superior temporal gyrus 56y38 16 7.06*** 5.01***(60,y38,20) Non-significant14. L middle temporal gyrus y52 y52 4 6.66*** 5.05**(y49,y52,0) 5.22**(y56,y49,4)15. R superior temporal gyrus 56 11 0 5.00* 5.77***(49,20,10) Non-significant

Occipital lobe16. L fusiform gyrus y41 y60 y12 8.02*** 5.70*** (y49,y70,5) Non-significant17. R fusiform gyrus 38 y74 y12 5.59** 5.45**(45,y64,y4) 4.72**(38,y84,0)

Deep grey18. L thalamus y8 y19 y8 4.30a 5.29*** (y4,y22,y4) 6.08*** (y12,y20,y4)19. R thalamus 4 y22 y8 4.37a 5.30*** (4,y22,y8) 5.89*** (4,y22,y12)

Talairach coordinates andz-scores are given from our previous study(Kiehl et al., 1999b). Control participants and psychopathicindividuals’ z-scores and Talairach coordinates are listed for the same anatomical locations as in the previous study(at or withinon smoothing element, 12=12=12 mm). Lsleft; Rsright. denotes az-score in Kiehl et al.(1999b) that was not reported in thea

original study because the significance level only reached trend levels(P-0.15).Note: *** P-0.001, **P-0.01, *P-0.05.

of adult psychopathic populations. To our knowl-edge, only one study has found evidence forlanguage abnormalities in adolescent psychopathicindividuals (Raine et al., 1990). This raises theissue of when in development these abnormalitiesmight originate. Unfortunately, very little is knownabout the onset and course of psychopathy, andeven less is known about the cognitive correlatesof the syndrome at young ages. Modern psychop-athy assessment techniques are currently investi-gating the identification of these individuals at avery early age(Forth and Burke, 1998; Frick,

1998). Clearly, research on the cognitive andneural processes of these individuals is needed.For the sake of argument, if we assume that

psychopathy is present at an early age and thatabnormalities in semantic processes related to con-ceptually abstract material are also present at thatage, then how might these abnormalities lead topsychopathic-like behavior? Perhaps psychopathicindividuals have difficulty engaging in cognitivefunctions that involve material that has no concreterealization in the external world. We might spec-ulate that complex social emotions such as love,


Fig. 2. Cortical surface rendering of the areas in which controlparticipants showed significantly greater activation for proc-essing of abstract stimuli than for concrete stimuli than didpsychopathic individuals. The region was located in the rightanterior superior temporal gyrus(Talairach coordinates,xs52,ys15, zsy10). The displayed area of activation is threshold-ed at az-score of 3.0 or greater.

Fig. 3. Illustration of the percent signal in the right anteriorsuperior temporal gyrus(Talairach coordinates,xs52, ys15,zsy10) where controls showed greater activation for proc-essing abstract words, relative to concrete words, than did non-psychopaths.

empathy, guilt and remorse may be a form of moreabstract functioning. Thus, difficulties in process-ing and integrating these conceptually abstractrepresentations to regulate or modulate behaviorwould be impaired in these individuals. Such aninterpretation is consistent with the developmentalliterature. Piaget(1926), for example, argued thatabstract thought was crucial to the developmentalstages of cognition. In particular, Piaget(1926)believed that abstract thought was necessary toreach a cognitive level termed ‘formal operations’.Formal operations were believed to take placeduring puberty and included the ability to reason‘in terms of verbally stated hypotheses and nolonger merely in terms of concrete objects andtheir manipulation’ (Piaget, 1926). Theories inregard to moral development also emphasizeabstract reasoning(Kohlberg, 1981). In neuropsy-chiatry, abstract reasoning is believed to be anability, rather than a developmental stage, that canbe lost in some pathological conditions(Goldstein,1939, 1959). Given the growing evidence support-ing abnormalities in processing conceptuallyabstract representations, these data suggest that

abstract processing deficits may be a fundamentalabnormality in psychopathy.Such an interpretation is not incompatible with

other theories of psychopathy. For example, thelow fear hypothesis of psychopathic behavior pos-its that psychopathic individuals suffer from achronic need for stimulation, compounded by arelative fearlessness of novel and dangerous situ-ations (Levenson, 1990, 1992; Levenson et al.,1995; Lykken, 1957, 1982; Patrick, 1994; Patricket al., 1993, 1994). If the semantic(and perhapslimbic) networks that normally engender learning,memory, and perhaps conditional relations betweenabstract materials and concept formations—alltools used to regulate behavior, then a failure toincorporate contextual cues related to emotional(including learned fearful stimuli) and otherabstract concepts(e.g. the association betweenpunishment and behavior) may lead to psycho-pathic behaviors. Moreover, theorists have arguedthat if these behaviors are present at an early age,they may be further compounded by rearing envi-ronments in which there are poor parental styles(Lykken, 1995). Individuals who have these defi-cits may be at particular risk in such environmentsfor the development of psychopathy.


Newman et al.(1997) have also shown thatpsychopathic individuals are less likely than non-psychopathic individuals to show ‘automatic’ inter-ference to contextual cues(at short interstimulusintervals). In clarifying their use of the term‘automatic’, Newman et al. indicated that ‘the termis used to indicate that psychopaths do not considerthemeaning wemphasis oursx of their words, theiractions, and situational cues in a spontaneous way’.Interestingly, patients with right hemisphere lesionsshow reduced interference effects for languagestimuli compared with left hemisphere brain-dam-aged patients or control participants(Doyon andMilner, 1991). It is plausible that the deficitsobserved in psychopaths for processing abstractstimuli may also be related to the contextualdeficits observed by Newman et al. We note thatby using ERPs, we have been able to characterizethe temporal structure of the processes in psycho-paths. These ERP data indicate that the observedabnormalities in semantic processes occur in psy-chopathy as early as 200 ms post-stimulus. Thus,deficits in accommodating contextual cues, thosethat are possibly semantic in nature, may lead toless interference, at least in part, because of arelative deficit (or advantage) in processingsemantic information. Studies have shown thatsemantic processes can be accessed automatically(e.g. in the absence of awareness) up to 600 mspost-stimulus(Luck et al., 1996). We are notarguing against the interpretation that there are‘response modulation’ deficits in psychopaths;rather we are suggesting that the behavioral abnor-malities observed in psychopathic individuals maybe due, at least in part, to abnormalities in semanticprocesses.Although these deficits appear to be present in

psychopathic individuals in various paradigms,there is some evidence that some forms of semantictasks are processed normally in psychopaths. Forexample, at least two studies have shown thatpsychopaths perform normally on semantic tasksrelated to non-verbal processing(Day and Wong,1996; Gillstrom, 1994).These data may have implications for the treat-

ment and management of psychopaths. If part ofthe psychopathic syndrome is due to failing toappropriately integrate semantically abstract infor-

mation because of a deficit or limitation in proc-essing the abstract representations of stimuli, thentraining and treatment protocols that specificallyaddress these deficits may be better suited foreffectively treating the condition. Moreover, per-haps treatment and management of psychopathicindividuals would be improved if these programsemphasized concepts in more concrete forms.There are a number of limitations in the present

study that should be addressed in future work. Ithad a small sample size, which raises the possibil-ity that some of the observed effects may besample-specific. However, previous research fromour laboratory, in a separate cohort of inmates, hasshown that psychopathy is associated with reducedbehavioral and ERP differentiation between con-crete and abstract words in a very similar task(Kiehl et al., 1999a). Thus, these data are sup-ported by the findings of Kiehl et al.(1999a).Moreover, as summarized above, numerous stud-ies, from our laboratory and others, have shownthat psychopathy is related to abnormalities inprocessing conceptually abstract information. Also,in the comparisons of interest, we treated partici-pants as a random effect; therefore, a significanteffect involving group can be expected to gener-alize to a new sample of similarly chosenparticipants.It is important to note that the psychopaths did

perform very well overall—with above 88% accu-racy for all word types. Psychopaths were not lessaccurate than controls on concrete word process-ing; indeed they preformed slightly better thancontrols, suggesting the slower reaction times maybe due, at least in part, to a reaction timeyaccuracytradeoff. Also, the pattern of hemodynamics asso-ciated with processing word and pseudoword stim-uli was similar between psychopaths and controlsin many brain regions known to be involved inlanguage processing. As predicted, only the rightanterior superior temporal gyrus showed groupdifferences in hemodynamic activity. That is, psy-chopaths activated nearly all brain regions believedto be associated with lexico-semantic processing,(see Fig. 1A,B). Thus, the preponderance of evi-dence suggests that these data illustrate a focal,rather than a global, deficit in psychopathy.


Psychopaths and non-psychopaths showed smallbut oppositely directed response biases during bothtypes of block. The psychopaths selected ‘word’in 52.45% of trials in concrete blocks and 51.75%of trials in abstract blocks, while the non-psycho-paths selected ‘word’ in 49.85% of trials in con-crete blocks and 47.95% of trials in abstract blocks.The overall difference between the groups inresponse bias is not likely to explain the increasedhemodynamic response observed in non-psycho-paths during the abstract blocks compared withconcrete blocks. However, we cannot exclude thepossibility that the small shift in response biasfavoring a ‘non-word’ response by non-psycho-paths during abstract blocks is related to theincrease in activation during abstract blocks. How-ever, even if the small shift in response bias innon-psychopaths did contribute to the increasedactivation during abstract blocks, it would supportthe conclusion that the observed differences inbrain activation reflect a greater responsiveness toabstract stimuli by non-psychopaths.In previous studies of psychopathy, the most

common diagnostic cutoff for psychopathy was 30and above(Hare, 1991). In the present study, twoinmates had PCL-R scores slightly below thiscutoff. We note, however, that the exact diagnosticcutoff for psychopathy is a matter of currentresearch and debate. Moreover, inclusion of indi-viduals with scores below the typical cutoff forpsychopathy would, if anything, lead to a conser-vative bias in any results.The third limitation of the present study is that

we did not employ an incarcerated non-psycho-pathic control group. This raises the issue thatsome of the observed effects in the psychopathicgroup may be due to the effects of incarcerationper se rather than psychopathy. Due to the logis-tical and financial cost of transporting inmates 100km from the prison to the UBC Hospital by theCorrectional Services of Canada’s Regional Escortteam, we opted not to recruit an incarcerated non-psychopath control group. This decision also wasmade in consideration of the results from ourprevious study(Kiehl et al., 1999a) examiningconcrete and abstract word processing in psycho-paths and incarcerated non-psychopaths in whichwe observed that the performance and event-

related potentials of the non-psychopathic groupclosely resembled those of non-criminals(Kouniosand Holcomb, 1994; Paller et al., 1987). Thecontrol group employed in the present study wasmatched to the psychopathic group on age, edu-cation, IQ measures, socio-economic status andhandedness.In summary, we have shown psychopathy is

associated with abnormalities in semantic process-ing of conceptually abstract information. Theseabnormalities appear to be localized to the rightanterior superior temporal gyrus and surroundingcortex. Evidence from converging sources is alsoconsistent with the hypothesis that psychopathy isrelated to right hemisphere abnormalities, particu-larly in the temporal lobe.

Acknowledgments

We thank Drs Alex MacKay and Ken Whittallfor their assistance. We also thank MR technolo-gists Trudy Shaw, Karen Smith and Sylvia Ren-neberg. This research was supported in part bygrants from the Medical Research Council(MRC)of Canada, the British Columbia Health Services,the British Columbia Medical Services Foundationand funds from the Schizophrenia Division,Department of Psychiatry, University of BritishColumbia. The first author was supported by theMichael Smith Graduate Scholarship, MedicalResearch Council of Canada. The second authorwas supported by a Killam Graduate Scholarship.The third author was supported by a UniversityGraduate Fellowship.

References

Ammons, R.B., Ammons, C.H., 1962. The Quick Test(QT):provisional manual. Psychological Reports 11, 111–161.

Ammons, R.B., Ammons, C.H., 1979a. Use and evaluation ofthe Quick Test(QT): partial summary through October,1979: I. published papers. Psychological Reports 45, 943–946.

Ammons, R.B., Ammons, C.H., 1979b. Use and evaluation ofthe Quick Test(QT): partial summary through October,1979: II. Reviews, theses, unpublished reports and papers.Psychological Reports 45, 953–954.

Annett, M., 1970. A classification of hand preference byassociation analysis. British Journal of Psychology 61,303–321.


Beauregard, M., Chertkow, H., Bub, D., Murtha, S., 1997. Theneural substrate for concrete, abstract, and emotional wordlexica: a positron emission tomography study. Journal ofCognitive Neuroscience 9, 441–461.

Beeman, M., Friedman, R.B., Grafman, J., Perez, E., Diamond,S., Lindsay, M.B., 1994. Summation priming and coarsesemantic coding in the right hemisphere. Journal of Cogni-tive Neuroscience 6, 25–45.

Bottini, G., Corcoran, R., Sterzi, R., Paulesu, E., Schenone, P.,Scarpa, P., Frackowiak, R.S., Frith, C.D., 1994. The role ofthe right hemisphere in the interpretation of figurativeaspects of language. A positron emission tomography acti-vation study. Brain 117, 1241–1253.

Brinkley, C.A., Bernstein, A., Newman, J.P., 1999. Coherencein the narratives of psychopathic and nonpsychopathiccriminal offenders. Personality and Individual Differences27, 519–530.

Brownell, H.H., Michel, D., Powelson, J., Gardner, H., 1983.Surprise but not coherence: sensitivity to verbal humor inright-hemisphere patients. Brain and Language 18, 20–27.

Caramazza, A., Gordon, J., Zurif, E.B., DeLuca, D., 1976.Right-hemispheric damage and verbal problem solvingbehavior. Brain and Language 3, 41–46.

Cleckley, H., 1976. The Mask of Sanity. 5th ed. Mosby, St.Louis, MO.

Day, J., 1977. Right-hemisphere language processing in normalright-handers. Journal of Experimental Psychology: HumanPerception and Performance 3, 518–528.

Day, R., Wong, S., 1996. Anomalous perceptual asymmetriesfor negative emotional stimuli in the psychopath. Journal ofAbnormal Psychology 105, 648–652.

D’Esposito, M., Detre, J.A., Aguirre, G.K., Stallcup, M., 1997.A functional MRI study of mental image generation. Neu-ropsychologia 35, 725–730.

Doyon, J., Milner, B., 1991. Right temporal-lobe contributionto global visual processing. Neuropsychologia 29, 343–360.

Flor-Henry, P., 1972. Psychiatric syndromes considered asmanifestations of lateralized temporal-limbic dysfunction.Paper presented at the Third International Congress ofPsychosurgery.

Forth, A.E., Burke, H.C., 1998. Psychopathy in adolescence:assessment, violence, and developmental precursors. In:Cooke, D.J., Forth, A.E., Hare, R.D.(Eds.), Psychopathy:Theory, Research, and Implications for Society. KluwerAcademic Publishers, Dordrecht, The Netherlands, pp.205–229.

Francis, W.N., Kucera, H., 1982. Frequency Analysis ofEnglish Usage. Houghton Mifflin, Boston.

Frick, P.J., 1998. Callous-unemotional traits and conduct prob-lems: applying the two-factor model of psychopathy tochildren. In: Cooke, D.J., Forth, A.E., Hare, R.D.(Eds.),Psychopathy: Theory, Research, and Implications for Soci-ety. Kluwer Academic Publishers, Dordrecht, The Nether-lands, pp. 161–188.

Friston, K.J., Ashburner, J., Frith, C.D., Poline, J.-B., Heather,J.D., Frackowiak, R.S.J., 1995a. Spatial registration andnormalization of images. Human Brain Mapping 2, 165–189.

Friston, K.J., Holmes, A.P., Worsley, K.J., Poline, J.-P., Frith,C.D., Frackowiak, R.S.J., 1995b. Statistical parametric mapsin functional imaging; a general linear approach. HumanBrain Mapping 2, 189–210.

Fulero, S., 1996. Review of the Hare Psychopathy Checklist-Revised. In: Conoley, J.C., Impara, J.C.(Eds.), 12th MentalMeasurements Yearbook. Buros Institue, Lincoln, pp.453–454.

Gardner, H., Denes, G., 1973. Connotative judgments byaphasic patients on a pictorial adaptation of the semanticdifferential. Cortex 9, 183–196.

Gillstrom, B.J., 1994. Abstract thinking in criminal psycho-paths. Unpublished Dissertation, University of BritishColumbia, Vancouver.

Goldstein, K., 1939. The significance of special mental testsfor diagnosis and prognosis in schizophrenia. AmericanJournal of Psychiatry 96, 575–588.

Goldstein, K., 1959. Concerning the concreteness in schizo-phrenia. Journal of Abnormal and Social Psychology 59,146–148.

Hare, R.D., 1979. Psychopathy and laterality of cerebralfunction. Journal of Abnormal Psychology 88, 605–610.

Hare, R.D., 1980. A research scale for the assessment ofpsychopathy in criminal populations. Personality and Indi-vidual Differences 1, 111–119.

Hare, R.D., 1991. Manual for the Hare Psychopathy Checklist-Revised. Multi-Health Systems, Toronto.

Hare, R.D., Forth, A.E., 1985. Psychopathy and lateral pref-erence. Journal of Abnormal Psychology 94, 541–546.

Hare, R.D., Harpur, T.J., Hakstian, A.R., Forth, A.E., 1990a.The revised psychopathy checklist: reliability and factorstructure. Psychological Assessment 2, 338–341.

Hare, R.D., Harpur, T.J., Hakstian, A.R., Forth, A.E., Hart,S.D., Newman, J.P., 1990b. The revised psychopathy checkl-ist: description statistics, reliability, and factor structure.Psychological Assessment 2, 338–341.

Hare, R.D., Hart, S.D., Harpur, T.J., 1991. Psychopthy and theDSM-IV criteria for antisocial personality disorder. Journalof Abnormal Psychology 100, 391–398.

Hare, R.D., Jutai, J.W., 1988. Psychopathy and cerebral asym-metry in semantic processing. Personality and IndividualDifferences 9, 329–337.

Hare, R.D., McPherson, L.M., 1984. Psychopathy and percep-tual asymmetry during verbal dichotic listening. Journal ofAbnormal Psychology 93, 141–149.

Harpur, T.J., Hakstian, A.R., Hare, R.D., 1988. The factorstructure of the psychopathy checklist. Journal of Consultingand Clinical Psychology 56, 741–747.

Harpur, T.J., Hare, R.D., 1994. Assessment of psychopathy asa function of age. Journal of Abnormal Psychology 103,604–609.

Harpur, T.J., Hare, R.D., Hakstian, A.R., 1989. Two-factorconceptualization of psychopathy: construct validity andassessment implications. Psychological Assessment 1, 6–17.

Hart, S.D., Hare, R.D., 1989. Discriminant validity of thepsychopathy checklist in a forensic psychiatric population.Psychological Assessment 1, 211–218.


Hart, S.D., Hare, R.D., Harpur, T.J., 1992. The PsychopathyChecklist-Revised(PCL-R): an overview for researchersand clinicians. In: James, P.M., Rosen, C.(Eds.), Advancesin Psychological Assessment, Vol. 8. Plenum Press, NewYork, pp. 103–130.

Gillstrom, B.J., Hare, R.D., 1988. Language-related handgestures in psychopaths. Journal of Personality Disorders 2,21–27.

Holcomb, P.J., 1993. Semantic priming and stimulus degrada-tion: implications for the role of the N400 in languageprocessing. Psychophysiology 30, 47–61.

Holcomb, P.J., Kounios, J., Anderson, J.E., West, W.C., 1999.Dual-coding, context-availability, and concreteness effectsin sentence comprehension: an electrophysiological investi-gation. Journal of Experimental Psychology: Learning,Memory and Cognition 25, 721–742.

Intrator, J., Hare, R.D., Stritzke, P., Brichtswein, K., Dorfman,D., Harpur, T., Bernstein, D., Handelsman, L., Schaefer, C.,Keilp, J., Rosen, J., Machac, J., 1997. A brain imaging(single photon emission computerized tomography) studyof semantic and affective processing in psychopaths. Bio-logical Psychiatry 42, 96–103.

James, C., 1975. The role of semantic information in lexicaldecisions. Journal of Experimental Psychology: HumanPerception and Performance 14, 130–136.

Jutai, J.W., Hare, R.D., Connolly, J.F., 1987. Psychopathy andevent-related brain potentials(ERPs) associated with atten-tion to speech stimuli. Personality and Individual Differenc-es 8, 175–184.

Kiehl, K.A., Hare, R.D., McDonald, J.J., Brink, J., 1999a.Semantic and affective processing in psychopaths: an event-related potential study. Psychophysiology 36, 765–774.

Kiehl, K.A., Liddle, P.F., Smith, A.S., Mendrek, A., Forster,B.B., Hare, R.D., 1999b. Neural pathways involved in theprocessing of concrete and abstract words. Human BrainMapping 7, 225–233.

Kohlberg, L., 1981. Essays on Moral Development. 1st edn.Harper and Row, San Francisco.

Kounios, J., Holcomb, P.J., 1994. Concreteness effects insemantic priming: ERP evidence supporting dual-codingtheory. Journal of Experimental Psychology: Learning,Memory and Cognition 20, 804–823.

Kroll, J.F., Merves, J.S., 1986. Lexical access for concrete andabstract words. Journal of Experimental Psychology: Learn-ing, Memory and Cognition 12, 92–107.

Kutas, M., Hillyard, S.A., 1980. Reading senseless sentences:brain potentials reflect semantic incongruity. Science 207,203–205.

Kutas, M., Hillyard, S.A., 1983. Event-related brain potentialsto grammatical errors and semantic anomalies. Memory andCognition 11, 539–550.

Kutas, M., Hillyard, S.A., 1984. Brain potentials during readingreflect word expectancy and semantic association. Nature307, 161–163.

Levenson, M.R., 1990. Risk taking and personality. Journal ofPersonality and Social Psychology 58, 1073–1080.

Levenson, M.R., 1992. Rethinking psychopathy. Theory andPsychology 2, 51–71.

Levenson, M.R., Kiehl, K.A., Fitzpatrick, C.M., 1995. Assess-ing psychopathic attributes in a non-institutionalized popu-lation. Journal of Personality and Social Psychology 68,151–158.

Louth, S.M., Williamson, S., Alpert, M., Pouget, E.R., Hare,R.D., 1998. Acoustic distinctions in the speech of malepsychopaths. Journal of Psycholinguistic Research 27,375–384.

Luck, S.J., Vogel, E.K., Shapiro, K.L., 1996. Word meaningscan be accessed but not reported during the attentional blink.Nature 383, 616–618.

Lykken, D.T., 1957. A study of anxiety in the sociopathicpersonality. Journal of Abnormal and Social Psychology 55,6–10.

Lykken, D.T., 1982. Fearlessness: its carefree charm anddeadly risks. Psychology Today 16, 20–28.

Lykken, D.T., 1995. The Antisocial Personalities. LawrenceErlbaum Associates, Hillsdale, NJ.

McCord, W.M., McCord, J., 1964. The Psychopath: An Essayon the Criminal Mind. Van Nostrand, Princeton, NJ.

Newman, J.P., Schmitt, W.A., Voss, W.D., 1997. The impactof motivationally neutral cues on psychopathic individuals:assessing the generality of the response modulation hypoth-esis. Journal of Abnormal Psychology 106, 563–575.

Paivio, A., 1986. Mental Representaions: A Dual CodingApproach. Oxford University Press, London.

Paivio, A., 1991. Dual coding theory: retrospect and currentstatus. Canadian Journal of Psychology 45, 255–287.

Paller, K.A., Kutas, M., Shimamura, A.P., Squire, L.R., 1987.Brain responses to concrete and abstract words reflectprocesses that correlate with later performance on test ofrecall and stem-completion priming. In: Johnson, R., Rohr-baugh, J., Parasuraman, R.(Eds.), Current Trends in BrainPotential Research. Elsevier, Amsterdam, pp. 360–365.

Patrick, C.J., 1994. Emotion and psychopathy: startling newinsights. Psychophysiology 31, 319–330.

Patrick, C.J., Bradley, M.M., Lang, P.J., 1993. Emotion in thecriminal psychopath: startle reflex modulation. Journal ofAbnormal Psychology 102, 82–92.

Patrick, C.J., Cuthbert, B.N., Lang, P.J., 1994. Emotion in thecriminal psychopath: fear image processing. Journal ofAbnormal Psychology 103, 523–534.

Piaget, J., 1926. The Language and Thought of the Child.Harcourt, Brace, New York.

Raine, A., O’Brien, M., Smiley, N., Scerbo, A., Chan, C.J.,1990. Reduced lateralization in verbal dichotic listening inadolescent psychopaths. Journal of Abnormal Psychology99, 272–277.

Ross, E.D., 1981. The aprosodias. Functional-anatomic organ-ization of the affective components of language in the righthemisphere. Archives of Neurology 38, 561–569.

Schwanenflugel, P.J., Harnishfeger, K.K., Stowe, R.W., 1988.Context availability and lexical decisions for abstract andconcrete words. Journal of Memory and Language 27,499–520.


Schwanenflugel, P.J., Stowe, R.W., 1989. Context availabilityand the processing of abstract and concrete words insentences. Reading Research Quarterly 24, 114–126.

Sharpe, K., O’Carroll, R., 1991. Estimating premorbid intel-lectual level in dementia using the National Adult ReadingTest: a Canadian study. British Journal of Clinical Psychol-ogy 30, 381–384.

Toglia, M.P., Battig, W.F., 1978. Handbook of Semantic WordNorms. Lawrence Erlbaum, Hillsdale, NJ.

Williamson, S., Harpur, T.J., Hare, R.D., 1990. Sensitivity toemotional valence in psychopaths. Paper presented at theAnnual meeting of the American Psychological Association,Boston.

Williamson, S., Harpur, T.J., Hare, R.D., 1991. Abnormalprocessing of affective words by psychopaths. Psychophys-iology 28, 260–273.

Winner, E., Gardner, H., 1977. The comprehension of metaphorin brain-damaged patients. Brain 100, 717–729.