A, Midsagittal slice of the corpus callosum illustrating measurement of the length, thickness, and area of the genu and splenium. B, Coronal slice through the genu and the rostrum of the corpus callosum illustrating the volume estimate of the body of the corpus callosum and the corona radiata extending into each hemisphere.
Raine A, Lencz T, Taylor K, Hellige JB, Bihrle S, Lacasse L, Lee M, Ishikawa S, Colletti P. Corpus Callosum Abnormalities in Psychopathic Antisocial Individuals. Arch Gen Psychiatry. 2003;60(11):1134-1142. doi:10.1001/archpsyc.60.11.1134
Copyright 2003 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2003
Psychopathic antisocial individuals have previously been characterized by abnormal interhemispheric processing and callosal functioning, but there have been no studies on the structural characteristics of the corpus callosum in this group.
To assess whether (1) psychopathic individuals with antisocial personality disorder show structural and functional impairments in the corpus callosum,(2) group differences are mirrored by correlations between dimensional measures of callosal structure and psychopathy, (3) callosal abnormalities are associated with affective deficits, and (4) callosal abnormalities are independent of psychosocial deficits.
Fifteen men with antisocial personality disorder and high psychopathy scores and 25 matched controls, all from a larger sample of 83 community volunteers.
Main Outcome Measures
Structural magnetic resonance imaging measures of the corpus callosum (volume estimate of callosal white matter, thickness, length, and genu and splenium area), functional callosal measures (2 divided visual field tasks), electrodermal and cardiovascular activity during a social stressor, personality measures of affective and interpersonal deficits, and verbal and spatial ability.
Psychopathic antisocial individuals compared with controls showed a 22.6% increase in estimated callosal white matter volume (P<.001), a 6.9% increase in callosal length (P = .002), a 15.3% reduction in callosal thickness (P = .04), and increased functional interhemispheric connectivity (P = .02). Correlational analyses in the larger unselected sample confirmed the association between antisocial personality and callosal structural abnormalities. Larger callosal volumes were associated with affective and interpersonal deficits, low autonomic stress reactivity, and low spatial ability. Callosal abnormalities were independent of psychosocial deficits.
Corpus callosum abnormalities in psychopathic antisocial individuals may reflect atypical neurodevelopmental processes involving an arrest of early axonal pruning or increased white matter myelination. These findings may help explain affective deficits and previous findings of abnormal interhemispheric transfer in psychopathic individuals.
AN INCREASING body of knowledge from brain imaging research is implicating brain abnormalities in the etiology of psychopathic and antisocial behavior, including abnormalities of the prefrontal cortex,1- 3 temporal cortex,3,4 hippocampus,5 parahippocampal gyrus,6 angular gyrus,7 cingulate,8 basal ganglia,9 and amygdala.6,10 The neurophysiologic basis of antisocial and psychopathic behavior is probably complex, with many interconnected brain structures likely to be involved in the regulation of impulsivity, emotional arousal, affect, and aggressive feelings. Interruption of the normal connectivity of these brain circuits is therefore likely to interfere with the normal brain processing that is critical to the regulation of affect.
Perhaps the single most important brain structure involved in connectivity is the corpus callosum. This structure orchestrates the complex interhemispheric regulation of attention, arousal, and emotion.11,12 Callosal abnormality could have profound implications not only for cognition, affect, and emotion regulation but also for psychopathologic status. Increased callosal area or thickness has been found in several disorders that are thought to be neurodevelopmental or genetic, including schizophrenia,13- 16 childhood-onset schizophrenia,17 schizotypal personality disorder,18 developmental language disorder,19 dyslexia,20 Cohen syndrome,21 velocardiofacial syndrome,22 and neurofibromatosis.23
To date, to our knowledge, there has been no research on the structural characteristics of the corpus callosum in any antisocial or psychopathic group. However, some studies suggest that antisocial and violent offenders may have functional abnormalities of the corpus callosum and its associated white matter radiations as indicated by glucose metabolism measured by positron emission tomography,7 increased nonspecific white matter abnormalities as assessed by magnetic resonance imaging,24 and increased interhemispheric coherence as measured by electroencephalography.25 Furthermore, structural callosal abnormalities could contribute to the reduced interhemispheric asymmetries of function previously found in psychopathic and antisocial groups as assessed by psychophysiologic and neuropsychologic tasks.26- 30
The key question addressed by this study concerns whether psychopathic antisocial individuals in the community show structural abnormalities in the corpus callosum and its associated white matter compared with controls drawn from the same socioeconomic stratum. In addition to linear and area measures of the corpus callosum, the volume of the callosal body and its white matter interhemispheric radiations was also estimated. Because the single previous functional imaging study7 in offenders observed reduced callosal glucose metabolism, it may be that antisocial individuals may have reductions in callosal size. Alternatively, reduced callosal glucose metabolism has been associated with increased callosal thickness and area,31 and, consequently, increased callosal size may also be hypothesized. A second question concerns whether any structural callosal abnormalities are paralleled by functional callosal abnormalities as measured by divided visual field tasks. Third, are abnormalities in callosal structure paralleled by dimensional analyses on a larger unselected sample? Fourth, emotion deficits hallmarked by lack of emotional depth, poor interpersonal relations, and reduced autonomic stress reactivity are key features of psychopathy32- 34; consequently, are callosal abnormalities particularly associated with these features? Fifth, are any observed callosal abnormalities independent of psychosocial risk factors for antisocial behavior?
Eighty-three men were recruited from 5 temporary employment agencies.2 Participants were unselected, except for the following exclusion criteria: age younger than 21 years or older than 45 years, nonfluency in English, history of epilepsy, claustrophobia, pacemaker, and metal implants. One man was excluded a priori because brain scanning revealed major atrophy of the right superior temporal gyrus.2 All individuals who read a description of the study and wanted to participate were included. Full written informed consent was obtained from all participants in accordance with institutional review board procedures at the University of Southern California. This community recruitment strategy is novel but has the advantages of (1) sampling individuals at socioeconomical high risk of psychopathy and antisocial personality disorder and (2) allowing recruitment of a control group that is demographically matched to the experimental group.
From this sample, individuals meeting the criteria for DSM-IV antisocial personality disorder and scoring in the top third(23-40) of the distribution of scores on the revised Psychopathy Checklist35 were assigned to the psychopathic antisocial group(n = 15). Controls neither met DSM-IV criteria for antisocial personality disorder or conduct disorder nor scored in the bottom third (0-14) of revised Psychopathy Checklist scores (n = 25). Full demographic, cognitive, physical, psychiatric, and antisocial measures in the 2 groups, and statistical comparisons, are given in Table 1.
All diagnoses were made using DSM-IV criteria36 and were ascertained using the Structured Clinical Interview for Axis I DSM-IV Disorders (SCID)37 and the Structured Clinical Interview for DSM-IV Axis II Personality Disorders (SCID-II).38 Diagnoses were made by research assistants who had undergone a standardized training and quality assurance program for diagnostic assessment.39 Schizophrenia-spectrum disorder was defined as schizotypal and paranoid personality disorders. Participants also completed an alcohol use questionnaire to assess the number of times alcohol was used in the past week and in the past month.2
Subtests of the Wechsler Adult Intelligence Scale–Revised40 were used to estimate verbal IQ (vocabulary, arithmetic, and digit span), performance IQ (digit symbol and block design), and total IQ. Degree of right vs left hand preference was assessed using the abbreviated Oldfield Inventory,41 with higher scores indicating greater right-handedness. History of head injury was defined as head trauma resulting in hospitalization. Ten demographic and psychosocial measures were derived from a structured psychosocial interview with the participant,42 with social class measured using the Hollingshead classification system.43 A physical examination was conducted after psychophysiologic testing to derive measures of height, weight, and head circumference.
Psychopathy was assessed using the revised Psychopathy Checklist35 supplemented by 5 sources of collateral data. Ratings were made by a PhD clinical graduate student trained and supervised by one of us (A.R.). A total score and scores on 3 subfactors of psychopathy—arrogant/deceptive, deficient affect, and impulsive/irresponsible—were derived.44 Internal reliability (Cronbach α) was .90. The 5 collateral data sources for assessing psychopathy were (1) the Interpersonal Measure of Psychopathy,45 which provides an interviewer's ratings of an individual's interpersonal behaviors, has demonstrated construct validity with the revised Psychopathy Checklist in a prison sample, and has been validated for use with nonincarcerated samples (ie, college students45); (2) self-reported crime as assessed by an adult extension2 of the National Youth Survey self-report delinquency measure46; (3) criminal history transcripts obtained from the Department of Justice; and (4 and 5) data derived from, and behavioral observations made during, the SCID and the SCID-II.
To help minimize denial of self-reported crime by truly criminal offenders, a certificate of confidentiality was obtained from the Secretary of Health and Human Services. Under section 303 (a) of Public Health Act 42, the research investigators were protected from being subpoenaed by any federal, state, or local court in the United States to release the self-reported crime data.
Lack of emotional and social depth was assessed using 3 self-report personality questionnaires: (1) the social closeness subscale of the Multidimensional Personality Questionnaire,47 (2) the no close friends subscale of the Schizotypal Personality Questionnaire,48 and(3) the blunted affect subscale of the Schizotypal Personality Questionnaire.48 In addition, lack of remorse, the only item (scored on a 3-point scale) from the SCID-II definition of antisocial personality disorder that reflects lack of emotional or social depth, was also used as an indicator.
Autonomic activity (skin conductance and heart rate) were assessed during a social stressor while the participant gave a speech about his worst faults.2 These measures have been previously reported in methodological detail in this sample.2
Degree of functional interhemispheric connectivity was assessed using a consonant-vowel-consonant trigram identification task49,50 and a letter-matching task.51 Full methodological details as used in the larger sample are provided in the article by Hellige et al.52 Briefly, in the consonant-vowel-consonant task, participants were presented with a nonsense consonant-vowel-consonant stimulus in either the left or right visual field or in both fields simultaneously (bilateral condition). The callosal integration measure was the difference in accuracy between the bilateral condition and the better of the 2 individual visual field accuracy scores, with higher scores indicating greater integration.49 In the letter-matching task, uppercase and lowercase letters that either matched (eg, A-a) or did not match (eg, A-b) were presented either within or between visuals fields. The callosal integration measure was defined as errors on within-field trials minus errors on between-field trials, with higher scores indicating increased integration.51
Imaging procedures have been previously reported.2 Briefly, structural magnetic resonance imaging was conducted on a 1.5-T scanner (model S15/ACS; Philips, Shelton, Conn). After an initial alignment sequence of 1 midsagittal and 4 parasagittal scans (spin-echo T1-weighted image acquisition: repetition time, 600 milliseconds; echo time, 20 milliseconds) to identify the anterior commissure–posterior commissure plane, 128 3-dimensional T1-weighted gradient-echo coronal images (repetition time, 34 milliseconds; echo time, 12.4 milliseconds; flip angle, 35°; overcontiguous slices, 1.7 mm; matrix, 256 × 256; and field of view, 23 cm) were taken directly orthogonal to the anterior commissure–posterior commissure line.
Three-dimensional brain images were reconstructed using a SPARC workstation (Sun Microsystems Inc, Santa Clara, Calif) and semiautomated software (CAMRA S200 ALLEGRO; Cedar Software Corp, Mississauga, Ontario, Canada) for gray matter–white matter–cerebrospinal fluid segmentation. Segmentation was performed using a thresholding algorithm, with the operator (masked to group membership and other nonbrain measures) applying a cutoff value to the signal-intensity histogram to optimally differentiate white matter from gray matter, areas of which were defined using an automated seeding algorithm on each slice. Callosal and whole-brain volume measures were derived from the original coronal slices, whereas callosal area and linear measures were derived from the reformatted midsagittal slice orthogonal to the coronal acquisition plane and identified using the fourth ventricle and the aqueduct of Sylvius as landmarks.
Callosal area and linear measurement procedures were adopted from previous studies.53- 55 A line was drawn between anterior and posterior points of the corpus callosum54 to define callosal length. Perpendicular lines were then drawn to define the area of the genu as the anterior fourth and the area of the splenium as the posterior fifth (Figure 1, A). Callosal thickness was defined as the thickness of the corpus callosum at the midpoint of the length of the corpus callosum (Figure 1, A). Using coronal slices, an estimate of callosal and pericallosal white matter volume was obtained (Figure 1, B). On every coronal slice showing the body of the corpus callosum, the area of the corona radiata extending into each hemisphere, together with the body of the corpus callosum, was calculated by seeding white matter (see first paragraph of this section). Only white matter adjacent to and contiguous with the body of the corpus callosum was measured. Callosal volume was expressed as a function of whole-brain volume, whereas the genu and splenium areas were expressed as a function of total callosal area.
Whole-brain volume was defined as all cerebral cortex excluding the ventricles, pons, and cerebellum. The pons was excluded by drawing a straight line between the 2 innermost points that form the superior border. Colliculi were excluded when they were no longer attached to the cerebral hemispheres. For volume measures, areas on each slice were multiplied by slice thickness(1.7 mm) and summated to provide volumes. Interrater reliabilities (intraclass correlation coefficients) based on 27 scans (raters masked to each other's ratings and to group membership) were as follows: total brain volume, 0.99; estimated callosal and white matter volume, 0.72; genu, 0.95; splenium, 0.97; thickness, 0.55; and length, 0.99.
All analyses were conducted using statistical software (SPSS Inc, Chicago, Ill). Between-group t tests and χ2 tests were used to assess group differences on antisocial, demographic, magnetic resonance imaging, personality, cognitive, and psychophysiologic variables. All tests of significance are 2-tailed. Effect sizes were calculated using η2 (% variance accounted for) and Cohen d.56 To obtain a single overarching measure of callosal structure, a principal component analysis was conducted to extract a first component with significant loadings from all variables and a factor score computed using the regression method (mean ± SD, 0 ± 1). To assess the overall relationship between corpus callosum structure and group, all callosal measures were first entered into a multivariate analysis of variance (MANOVA). Follow-up univariate F tests were then conducted to assess which specific callosal measures differentiated the groups. To control for confounds, multivariate analysis of covariance was used. Logistic regression was conducted to predict group membership from the callosal measures over and above psychosocial measures using the enter procedure, a classification cutoff value of 0.5, and with variance estimation assessed using the Nagelkerke statistic. Categorical group analyses were followed up with dimensional, correlational analyses conducted on the entire sample(N = 83).
A principal component analysis of the 5 callosal measures produced a first principal component accounting for 32.4% of the variance, with significant loadings from volume (0.54), length (0.58), thickness (−0.48), genu area (−0.40), and splenium area (0.78). High factor scores indicated increased volume, increased length, increased splenium area, reduced genu area, and reduced thickness. The psychopathic antisocial group (mean ± SD, 0.59 ± 0.98) scored significantly higher than controls (mean ± SD, −0.31 ± 0.97) on this factor (t =2.8, d = 0.93; P = .009).
An omnibus MANOVA on callosal thickness, length, genu and whole-brain volume, splenium and whole-brain volume, and callosal white matter and whole-brain volume indicated an overall group difference (F5,32 = 6.8, η2 = 0.52; P<.001). Follow-up univariate F tests indicated that psychopathic antisocial individuals compared with controls had significantly increased callosal white matter and whole-brain volumes, increased callosal length, and reduced callosal thickness, but only a trend for a larger splenium (Table 2).
Groups differed most on estimated callosal volume (Table 2). It is possible that the increased length of the corpus callosum in the psychopathic antisocial group entirely accounts for the increase in white matter volume. After entry of callosal length as a covariate in an analysis of covariance, group difference in estimated callosal volume remained significant (F1,36 = 7.7; P = .009), indicating that increased callosal volume was not a function of increased callosal length.
Logistic regression using callosal measures to predict group membership was significant (χ25 = 33.8; P<.001) and correctly classified 92% of the cases. Of the 2 controls who were misclassified as being psychopathic antisocial, one had the highest score in the control group on psychopathy (score, 14) and the other had the lowest skin conductance level during the stressor (2.19 microseconds). The one psychopathic antisocial individual misclassified as a control had the highest heart rate during the stressor (88 bpm) in their group.
Although psychopathic antisocial individuals scored significantly lower than controls on self-reported social closeness, the effect of higher scores on blunted affect was only marginally significant, and the higher scores on no close friends was nonsignificant (Table 3). Consistent with the way groups were composed, psychopathic antisocial individuals were lacking remorse. As expected from previous findings,2 psychopathic antisocial individuals showed significantly reduced skin conductance and heart rate activity during the social stressor(Table 3).
The previous categorical findings were confirmed in a dimensional, correlational analysis using the entire sample. The strongest findings were found for the overall callosal factor score, with positive correlations for antisocial measures ranging from 0.35 to 0.51 (Table 4). In addition, increased callosal volume was significantly associated with blunted affect, lack of remorse, no close friends, lack of social closeness, reduced skin conductance and heart rate activity during the stressor, and, to a more limited extent, reduced spatial ability. Of the 5 contributing callosal measures, estimated callosal volume provided the most substantive contribution, with increased volume significantly associated with increased psychopathy, increased interpersonal deficits, reduced autonomic activity, and low spatial IQ. The association with psychopathy factors was most pronounced for deficient affect and least pronounced for arrogant/deceptive.
A MANOVA indicated group differences on the 2 interhemispheric communication measures (F2,37 = 4.3, η2 = 0.19; P = .02). The higher scores on both measures in the psychopathic antisocial group indicated greater hemispheric connectivity in this group compared with controls. However, in the enlarged sample (N = 83), these functional measures correlated neither with structural callosal measures (r = −0.07 to 0.19; P>.07) nor with antisocial, personality, and psychophysiologic measures (r =−0.15 to 0.18; P>.10).
Because the psychopathic antisocial group had significantly higher rates of alcohol and substance dependence than controls (Table 1), these variables were entered together as covariates in a MANOVA containing structural and functional callosal measures. The main effect for group remained significant (F7,28 = 3.9, η2 = 0.51; P<.004). Because alcohol use is of particular concern given its previous links to callosal abnormalities,57 alcohol use measures were entered as covariates alongside alcohol dependence. The main effect for group remained significant (F7,28 = 5.6, η2 = 0.59; P<.001).
Similarly, because increased white matter volume has been observed in cannabis users,58 cannabis could account for the observed relationships. However, after entry of cannabis abuse/dependence as a covariate, the main group effect remained significant (F7,28 =5.3, η2 = 0.57; P = .001).
Because schizophrenia-spectrum disorders show callosal abnormalities,14,18 and because psychopathic antisocial individuals had a higher rate of such disorders, they could be confounds. Consequently, diagnoses of schizotypal and paranoid personality disorder were simultaneously entered as covariates in the MANOVA. The main effect of group on callosal measures remained significant (F7,28 = 4.1, η2 = 0.52; P = .003).
To assess whether effects were independent of head circumference and IQ, these variables were entered as covariates. The main group effect remained significant (F5,27 = 8.0, η2 = 0.60; P<.001).
Psychosocial risk factors for antisocial behavior could account for the link between psychopathy and callosal structure and function. After entry of the 10 psychosocial risk factors as covariates in the MANOVA (physical abuse, sexual abuse, raised in an institution, raised by foster parents, parental criminality, early parental divorce, parental physical fights, parental verbal arguments, large family size, and low social class), the main effect remained significant (F7,20 = 3.9, η2 = 0.59; P = .009), indicating that the group findings were not attributable to these psychosocial processes.
Callosal structure and function measures also added to the prediction of psychopathic antisocial vs control group membership over and above psychosocial measures. In logistic regression, the 10 psychosocial risk factors accounted for 32.7% of the variance in group membership on block 1. After entry of the callosal measures on block 2, variance accounted for increased significance to 81.5% (χ22 = 23.9; P<.001).
To our knowledge, this study establishes for the first time the existence of a structural abnormality in the corpus callosum of psychopathic antisocial individuals. This group had a 22.6% increase in the volume of the corpus callosum and corona radiata compared with controls, a large effect size corresponding to d = 1.8.56 Callosi were 15.3% thinner (d= 0.7) and 6.9% longer (d = 1.1) in psychopathic antisocial individuals. These group differences were mirrored by dimensional, correlational findings in the larger sample of 83 men, with larger estimated callosal volumes associated with higher antisocial and psychopathy scores, especially with respect to affective psychopathic features and affective-social personality characteristics. Psychopathic antisocial individuals also showed significantly increased functional connectivity between the 2 hemispheres, but in this case categorical differences were not reflected in dimensional analyses in the entire sample. These structural and functional callosal abnormalities could not be attributed to alcohol and substance use, psychosocial deficits, head injury, schizophrenia-spectrum disorder, or whole-brain volume, and neither could callosal volume effects be attributed to increased callosal length.
Callosal structural abnormalities, particularly increased length and volume, suggest the possibility that psychopathic antisocial personality may be partly neurodevelopmental. Research with monkeys,59 cats,60 and hamsters61 has shown that approximately two thirds of callosal axons are eliminated postnatally through adulthood, with most of this pruning being to excitatory rather than inhibitory fibers.62 Early arrest of this normal process of axonal pruning, therefore, could contribute to the increased callosal white matter volume and the functional overconnectedness of the hemispheres observed in the psychopathic antisocial group. Despite extensive postnatal axonal pruning in healthy individuals, recent brain imaging research63,64 on children has documented major, rapid growth in the corpus callosum from early childhood to early adolescence presumably due to increased myelination of fibers that survive early elimination. Consequently, an abnormality in the myelination process involving overutilization of oligodendrocytes in the production of the myelin sheath could also partly explain callosal abnormalities in psychopathic antisocial individuals. Together, reduced retraction of inhibitory callosal fibers and increased myelination of axons could facilitate excitatory nerve conduction, which, in turn, could facilitate interhemispheric transfer and be consistent with the functional callosal findings of increased interhemispheric connectivity in psychopathic antisocial individuals.
A neurodevelopmental perspective of adult psychopathic antisocial personality is consistent with the facts that such behavior has its roots early in life,65,66 unfolds relatively consistently during childhood and adolescence,67,68 is relatively impervious to conventional treatments,69,70 and is in part genetically determined.71 In addition, psychosocial, demographic, and head injury measures could not account for structural callosal impairments in this group. As noted herein, a variety of other disorders that are neurodevelopmental in nature (eg, schizophrenia, schizotypal personality disorder, velocardiofacial syndrome, developmental language disorder, and dyslexia) are also characterized by increased callosal size. The fact that morphologic changes to the corpus callosum were complex, involving thinning and lengthening as well as increased white matter volume, tends to dictate against simple, nondevelopmental processes such as discrete trauma or degenerative disease processes.
The callosal structural abnormalities observed during this study may help account for the abnormal interhemispheric processing in psychopathic individuals repeatedly observed in multiple cross-laboratory studies.26- 29,72- 77 The most commonly reported abnormalities in psychopathic individuals are reduced lateralization in P300 amplitudes,29 verbal dichotic listening,26- 28 and visual event-related potentials.30 Reduced lateralization could be a consequence of greater interhemispheric connectivity. In line with this proposition, greater interhemispheric connectivity was found herein for psychopathic antisocial individuals as assessed by divided visual field tasks. Our group72 has previously hypothesized that prior findings of reduced lateralization may arise from a disturbance in the normal neurodevelopmental processes of hemispheric specialization. Atypical neurodevelopment of the corpus callosum could have wide-ranging effects on the interhemispheric regulatory processes that contribute to the affective, autonomic, cognitive, and antisocial characteristics that predispose to antisocial personality disorder and psychopathy.
A key feature of psychopathy is blunted affect, and low autonomic activity during emotional and social stressors is a well-replicated correlate of psychopathy.32- 34,78,79 Callosal white matter volume was significantly related to the deficient affect factor of psychopathy and, to a lesser extent, to the impulsive/irresponsible factor but not to the arrogant/deceptive factor. Similarly, autonomic and personality measures reflecting blunted affect, lack of social closeness, and no close friends were related to callosal abnormalities. Individuals who experience neurodevelopmental failure of the corpus callosum do not evidence gross psychiatric symptoms but do show deficits in social insight and self-perception,80 deficits that also characterize psychopathic individuals.35,81 As such, abnormal interhemispheric connectivity may partly account for the social, insight, autonomic, and emotion deficits observed in psychopathic individuals.
Low spatial IQ was observed in the psychopathic antisocial group and was additionally associated with increased estimated callosal volume. These associations are of interest for 3 reasons. First, agenesis of the corpus callosum and split-brain surgery have been associated with poor spatial ability.82,83 Second, low spatial but not verbal ability early in life has been found to characterize lifelong antisocial individuals.84 Third, the same correlation between increased estimated callosal volume and low spatial IQ has also been found in neurofibromatosis.85 These findings, in turn, suggest that callosal abnormalities may account for the spatial deficits in psychopathic antisocial individuals and that increased callosal volume relative to normal is disadvantageous rather than beneficial.
Limitations of this study include the fact that because only men were studied, findings cannot currently be generalized to antisocial women. Group sizes were modest, although expanded analyses on the larger sample of 83 men indicate that the key results are reliable. Although group differences in callosal functioning were found in parallel with structural callosal deficits, these measures were not intercorrelated in the expanded sample, indicating that structural and functional deficits are found only at the level of pathological grouping. The volume estimate of callosal white matter includes pericallosal white matter, and so findings may not be entirely specific to callosal integrity. Alternatively, this volume measure loaded substantially (0.54) on the callosal factor, indicating that results were not driven exclusively or primarily by pericallosal white matter. Reliability for some callosal measures was modest, and, consequently, effect sizes may be underestimates of true effect sizes. Personality measures of affect, although supported by autonomic measures, also have their limitations.
Finally, only an association has been shown between callosal deficits and psychopathic antisocial behavior, and causality cannot be assumed. Specifically, it is unlikely that callosal abnormalities directly cause antisocial psychopathic behavior in a one-to-one fashion and more likely they contribute to a breakdown in a wider network of regulatory interhemispheric cortical systems that regulate and control behavior. Thus, a critical question to be addressed in the future concerns how these individual brain deficits are networked and conspire to produce psychopathic antisocial behavior. Atypical neurodevelopment of the corpus callosum and consequent "faulty wiring" of the brain may be one of the contributory mechanisms to the development of psychopathic behavior, but alternative processes and the interplay with early psychosocial development influences should not be discounted. Future studies using diffusion tensor imaging to measure white matter integrity could help confirm and extend the present findings.
Corresponding author: Adrian Raine, DPhil, Department of Psychology, University of Southern California, Los Angeles, CA 90089-1061 (e-mail: firstname.lastname@example.org).
Submitted for publication May 7, 2002; final revision received May 9, 2003; accepted May 12, 2003.
This study was supported by Research Scientist Development Award K02 MH01114-01, Independent Scientist Award K02 MH01114-08, and grant 5 RO3 MH50940-02 from the National Institute of Mental Health, Bethesda, Md (Dr Raine), and by the Wacker Foundation, Dallas, Tex (Dr Raine).
We thank Jennifer Bobier, BA, Nicole Diamond, BA, Kevin Ho, BA, Blane Horvath, BSc, Rina Kadakia, Shari Mills, PhD, and Pauline Yaralian, PhD, for assistance with data collection and scoring.