حافظه برای رویدادهای عاطفی در مجرمان خشونت طلب با اختلال شخصیت ضد اجتماعی

Abstract Eight-eight violent offenders meeting the criteria for DSM-IV antisocial personality disorder (ASPD) were assessed using the Psychopathy Checklist-Screening Version (PCL-SV) and compared with 20 healthy controls on an emotional memory task. All participants showed enhanced memory for the emotional phase of the task. On the free and cued recall components of the task high psychopathy scorers showed most impairment on the emotional phase compared with healthy controls. Analyses of psychopathy dimensions indicated no evidence of a dimensional relationship with emotional memory, but subjects categorised as scoring on the extremes of the psychopathy scale particularly those with high emotional detachment (Factor 1) and antisocial behaviour (Factor 2) had impairments in free recall an effect that was attenuated when subjects received cues to assist recall. The results suggest that emotional memory impairments in antisocial populations may be related to a variety of emotional and attentional processes linked with the prefrontal-limbic neural circuitry.

Introduction The antisocial personality disorders (antisocial personality disorder and psychopathy) are a group of overlapping disorders of personality that are associated with significant intra and interpersonal dysfunction. Antisocial Personality Disorder (ASPD) as defined in DSM-IV is more common than the dimensionally defined higher order construct of psychopathy as described by Cleckley (1976) and more recently by Hare (1991). Rates of these disorders are significantly higher in prison and forensic samples than in the general population (Hare, 1991). Half of prison inmates meet DSM-IV criteria for ASPD in the UK (Singleton, Meltzer, & Gatward, 1998). Coid (1992) reported a prevalence rate for psychopathy of 38% in maximum-security psychiatric hospital patients. Psychopathy assessed using Hare’s (1991) criteria is estimated to occur in less than 20% of prisoners with a diagnosis of ASPD (Hare, 1998). Although all these disorders are known to be a major cause of social distress—disruption to families, criminality and violence (Robins, 1990), very little is known about causation. Social factors are associated with antisocial personality disorders (Robins, 1990). However, they also occur in those with apparently normal family backgrounds and there is evidence of genetic liability (McGuffin & Thapar, 1992) suggesting a neurobiological basis for these disorders. There are a number of theories relating to the development of antisocial behaviour, the most prominent of which are the punishment/low fear theories (see Lykken, 1995); the Response Modulation deficit hypothesis (Patterson & Newman, 1993; see Newman, 1998, for a review) and more recently the “Violence Inhibition Mechanism”(VIM) deficit proposed by Blair (1995) in which psychopathic behaviours and low empathy are perceived to be related to the failure of basic emotions (e.g. fear) to result in autonomic arousal and the inhibition of ongoing behaviour. Many of these theories focus on the behavioural components of antisocial behaviour and few focus on the interpersonal (low empathy, callousness) components of the disorder. Blair and Frith (2000), however, suggest that the amygdala may be a core component of the neural circuit that mediates the VIM and they have proposed that early amygdala dysfunction may result in the development of core psychopathic (affective–interpersonal) traits. Amygdala lesions in humans reduce the ability to acquire conditioned autonomic responses (Bechara et al., 1995) and impair the capacity to recall emotional material (Cahill, Babinsky, Markowitsch, & McGaugh, 1995). Functional imaging studies also confirm the notion that the amygdala is activated by affectively loaded visual stimuli (Blair et al., 1999, Breiter et al., 1996, Morris et al., 1996 and Phillips et al., 1997). There is now fairly extensive empirical evidence that psychopaths have attenuated electro-dermal responses in anticipation of aversive stimuli (see Hare, 1998, for review) and attenuated startle responses to unpleasant slides (Lang et al., 1990, Patrick et al., 1993 and Levenston et al., 2000). The latter findings are consistent with a hypothesised deficit in neural circuits involved in the processing of negative affect e.g. amygdala (Davis, 1989). To date most of the research literature on emotional information processing in antisocial samples has focussed on physiological responses to emotional visual stimuli and there has only been one study looking at memory for details of emotional events in psychopathic and non-psychopathic offenders (Christianson et al., 1996). In the latter study, psychopathic offenders did not show the normal memory bias for affectively laden material compared with non-psychopathic offenders. We tested the hypothesis that violent offenders meeting the criteria for ASPD would have impairments in emotional memory compared with healthy controls. We also examined emotional memory in relation to scores on the Psychopathy Checklist-Screening Version (PCL: SV, Hart, Cox, & Hare, 1995) in those rated as High or Low on Factor 1 (emotional detachment) as this measures the core affective traits that may be related to emotional information processing deficits (Patrick, 1994).

. Results 3.1. Characteristics of the sample The mean PCL: SV score for the offender sample was 15.82 (SD 3.12). Factor 1 (interpersonal/affective items) mean score was 7.37 (SD 2.45). Factor 2 (social deviance) score was 8.42 (SD 2.01), see Table 1 for group mean scores. There were no significant differences between controls and the antisocial offenders on the NART, see Table 1. The NART score correlated significantly with Cahill total free recall (r = 0.26, n = 106, p < 0.01) and with recognition score (r = 0.43, n = 105, p < 0.001) and was used as a covariate in the analysis. There were no significant correlations between PCL: SV total and sub-factor scores and performance on the Cahill emotional memory test. 3.2. Performance on the Cahill emotional memory task 3.2.1. Recall test Overall, all participants showed a significant phase effect on the free recall task (F[2, 214] = 109.79, p < 0.001). Pair-wise comparisons revealed significant differences between phase 1 and 2 (mean difference −0.60, p < 0.001), phase 2 and 3 (mean difference −1.61, p < 0.001) and phase 1 and 3 (mean difference −1.01, p < 0.001). In general, subjects tended to remember phase 2, i.e. the emotional component of the task more readily. Analyses of each group separately indicated that there were significant differences (paired t test) in recall of phase 1 and 2 [controls (t = −3.03, df 19, p < 0.01); Low psychopathy (t = −1.84, df 26, 0.07); Medium psychopathy (t = −3.55, df 37, p < 0.001) but not in the High psychopathy group (t = −1.69, df 20, p = 0.10). A similar comparison of phase 2 and 3 indicated all 4 groups had better recall for phase 2 than 3 at the p < 0.001 level. A repeated measures ANOVA revealed a significant effect of group (F[3, 102] = 3.6, p < 0.01), phase (F[2, 204] = 105.6, p < 0.001) but no significant phase × group interaction (F[6, 204] = 0.86, p = 0.52). Posthoc testing indicated that the significant group differences were primarily between the healthy control comparison group and the High scoring psychopathic group (mean difference 0.71, p < 0.01). Table 2 shows the mean scores for the free recall component of the task. Table 2. Cahill emotional memory task Control Low PCL:SV Medium PCL:SV High PCL:SV F Significance Free recall 20 27 38 21 Phase 1 2.84 (0.68) 2.37 (1.1) 2.23 (1.10) 2.19 (0.98) 1.49 N/S Phase 2 3.47 (0.77) 2.81 (1.14) 3.00 (1.06) 2.66 (0.86)⁎ 2.26 0.08 Phase 3 1.68 (0.67) 1.55 (1.12) 1.23 (0.88) 0.95 (0.74)⁎ 3.9 0.05 Recognition Phase 1 19.47 (2.98) 16.70 (4.2) 15.78 (4.46)⁎ 16.00 (3.33)⁎ 3.7 0.05 Phase 2 21.10 (2.13) 18.17 (3.9) 17.78 (3.23)⁎ 17.85 (2.86)⁎ 5.1 0.01 Phase 3 11.47 (2.73) 10.33 (2.86) 10.78 (2.95) 11.00 (2.84) 0.28 N/S df = (4, 105). ⁎ Different from controls, p < 0.05. Table options Analyses of covariance controlling for IQ indicated that there was a trend towards a significant group difference in phase 2 (the free recall of the emotional phase) of the task (p = 0.08) and a significant difference in phase 3 (p < 0.05). There were no significant group differences in phase 1 of the task, see Table 2. Key differences were between High Psychopathy scorers and healthy controls for phase 2. 3.2.2. Recognition On the recognition task a similar effect of phase was seen in the whole sample (F[2, 208] = 261, p < 0.001) with higher rates of recognition in phase 2 compared with phase 1 (paired t test t = −5.4, df 105, p < 0.001) and between phase 2 and 3 (paired t test, t = 22.3, df 104, p < 0.001). Analyses of each of the groups separately indicated that there were significant differences in recognition between phase 1 and 2 in healthy controls (t = −2.1, df 19, p < 0.05), the Low psychopathy group (t = −2.8, df 26, p < 0.01), Medium psychopathy group (t = −2.8, df 37, p < 0.01) and High psychopathy groups (t = −4.4, df 20, p < 0.001). There were significant differences in the recognition of phase 2 and 3 in all groups at the level of p < 0.001. A repeated measures ANOVA with phase and group as factors indicated a significant effect of phase (F[2, 202] = 264.8, p < 0.001), a significant effect of group (F[3, 101] = 4.29, p = 0.007) and a significant group × phase interaction (F[6, 202] = 2.21, p = 0.04). Posthoc testing indicated there were significant differences between controls and the Medium and High scoring psychopathy groups. Group differences in the correct recognition of each phase are shown in Table 2. Analysis of covariance indicated that there were significant group differences in phase 1 (p < 0.05), phase 2 (p < 0.01) but not in phase 3, see Table 2. 3.3. Emotional detachment in violent criminals In a secondary analyses we categorised our antisocial sample into High and Low scorers on Factor 1 (emotional detachment). A repeated measures ANOVA on free recall indicated that there was a significant effect of group [healthy controls, High Factor 1, Low Factor I] (F[2, 103] = 4.22, p < 0.05), a significant effect of phase (F[2, 206] = 101.4, p < 0.001) but no significant phase × group interaction (F[4, 206] = 0.24, N/S). Posthoc testing indicated that the differences were between controls and High Factor 1 scorers and between Low and High Factor 1 scorers (p < 0.05) for phase 2 of the task. A similar analysis on the cued recognition task indicated a significant effect of group (F[2, 102] = 6.2, p < 0.01) a significant effect of phase (F[2, 206] = 256.5, p < 0.001) and a significant group × phase interaction (F[4, 206] = 3.9, p < 0.05). Posthoc testing indicated that both High and Low factor 1 subjects were significantly worse that the healthy control comparison group for both phase 1 and 2 of the task but not phase 3, see Table 3. Table 3. High/low emotional detachment Variable Controls Low Factor 1 High Factor 1 F Significance Recall (20) (44) (42) Phase 1 2.8 (0.69) 2.27 (1.08) 2.260 (1.08) 2.19 N/S Phase 2 3.50 (0.7) 2.81 (1.14)⁎ 2.90 (0.93)⁎ 3.39 0.03 Phase 3 1.68 (0.6) 1.25 (0.96) 1.28 (0.94) 1.39 N/S Recognition Phase 1 19.5 (2.9) 15.9 (3.9)⁎ 16.35 (4.3)⁎ 5.76 0.004 Phase 2 21.1 (2.1) 18.18 (3.3)⁎ 18.0 (3.4)⁎ 6.77 0.002 Phase 3 11.4 (2.7) 10.6 (2.69) 10.7 (3.0) 0.609 N/S df = (3, 105). ⁎ Significantly different from controls, p < 0.05. Table options 3.4. Social deviance A similar analysis on the free recall task based on High and Low Factor 2 score (antisocial behaviour) revealed a significant effect of group (F[2, 102] = 4.6, p < 0.05), significant effect of phase (F[2, 206] = 101.8, p < 0.001) but no group × phase interaction (F[4, 206] = 0.42, N/S). Posthoc testing indicated that the differences were between High and Low Factor 2 scorers and controls. However, group differences were not significant after covarying for IQ, see Table 4. Table 4. High and Low Factor 2 Variable Controls Low Factor 2 High Factor 2 F Significance Recall 19 43 43 Phase 1 2.8 (0.68) 2.27 (1.18) 2.25 (0.97) 2.1 N/S Phase 2 3.47 (0.77) 2.93 (1.07) 2.7 (1.01)⁎ 2.5 0.08 Phase 3 1.68 (0.67) 1.39 (1.04) 1.13 (0.83)⁎ 2.6 0.07 Recognition Phase 1 19.47 (2.9) 16.2 (4.7)⁎ 16.0 (3.4)⁎ 5.2 0.007 Phase 2 21.10 (2.1) 18.53 (3.9)⁎ 17.6 (2.6)⁎ 7.6 0.001 Phase 3 11.47 (2.7) 10.37 (3.2) 11.0 (2.4) 0.74 N/S df = (3, 105). ⁎ Significantly different from controls, p < 0.05. Table options On the cued recall task there was a significant effect of group (F[2, 102] = 6.28, p < 0.01) a significant effect of phase (F[2, 206] = 6.28, p < 0.01) and a significant group × phase interaction (F[4, 206] = 3.40, p < 0.01). Both High and Low factor 2 scorers had lower performance than the healthy comparison group. These differences remained significant for phase I and 2 of the task after controlling for IQ differences, see Table 4.