صفات عاطفه و عاری از احساس و مرزی در نوجوانان غیربالینی: پروفایل های شخصیت و روابط برای رفتارهای ضد اجتماعی

Abstract Callous-unemotional and borderline traits, antisocial behaviors, depressive symptoms and social anxiety were assessed in 972 high-school students. Cluster analysis yielded four groups: a low traits group, a second group that was above the mean in callous-unemotional traits and below the mean in borderline traits (moderate callous-unemotional traits cluster), a third group that was well above the mean in borderline traits but low on callous-unemotional traits (borderline traits cluster), and a fourth group that was high on both traits (high traits cluster). The high traits cluster was characterized by the highest level of antisocial behaviors and shared with the borderline traits cluster a high level of depressive symptoms and social anxiety. The moderate callous-unemotional traits cluster and the borderline traits cluster had similar levels of antisocial behaviors that were higher than those of the low traits cluster. This study suggests that there is a significant minority of non-clinical adolescents characterized by the presence of both callous-unemotional and borderline traits, and higher levels of antisocial behaviors and emotional distress.

Introduction Psychopathic traits, mainly characterized by callousness and unemotionality, and borderline personality traits, mainly characterized by instability of affects, identity, behavior and relationships, have been reported to be frequent in community samples of adolescents and to be linked to antisocial behaviors (e.g., Andershed et al., 2007, Bernstein et al., 1993, Chabrol et al., 2001 and McManus et al., 1984). However, there are only few studies investigating the relations between both psychopathic and borderline traits and antisocial behavior (Chabrol, van Leeuwen, Rodgers, & Séjourné, 2009). Psychopathic and borderline personality disorder traits have in common some characteristics but differ on other ones, both at the symptoms and defense mechanisms levels. Both disorders share impulsivity and heightened irritability, anger, and aggressiveness. For example, Fossati et al. (2004) reported that borderline and antisocial personality disorders symptoms shared a common impulsivity dimension but were linked to different aggressiveness facets. Borderline and psychopathic personality disorders principally differ in emotional reactivity which is considered to be low in psychopathic traits and high in borderline traits. However, borderline and psychopathic personality disorders share poor skills in understanding and managing emotions (Bateman and Fonagy, 2008 and Gardner et al., 2010). At the defensive level, borderline and psychopathic personality disorders are both characterized by the predominance of immature or primitive defense mechanisms but differences in defense use have been reported. For example, Presniak, Olson, and Macgregor (2010) found that in borderline personality, the defenses may manifest themselves in interpersonal dependency and a tendency to direct aggression toward the self, whereas in antisocial personality, the defenses may be expressed through egocentricity, interpersonal exploitation, and a tendency to direct aggression toward others. There are conflicting theories about the relationships between psychopathic and borderline traits in adolescents. On the one hand, Frick (2001) considered that there are two opposite types of conduct disorder, first juvenile psychopathy, characterized by callousness and low emotional reactivity, and a second type distinguished on the contrary by high emotional reactivity and closely paralleled to borderline personality disorder. On the other hand, Kernberg (1978) hypothesized that adolescents with psychopathic personality disorder may present with an underlying borderline personality organization characterized. In support of this hypothesis, Chabrol and Leichsenring (2006) reported a significant association between psychopathic traits and indicators of borderline personality organization in a non-clinical sample of adolescents. Further, attachment theory also suggests an association between borderline personality disorder and psychopathic personality disorder that have been linked to childhood abuse, disorganized attachment, and internalization of the abuser as an alien self (Fonagy, 1999 and Fonagy, 2000). Borderline and psychopathic personality disorders differ in the way of using projection in attempts to eliminate the alien self. Individuals with borderline personality tend to unstably alternate projections of the alien self into their own bodies and into interpersonal relationships. The projection of the alien self into the body leads to self-harm and suicide attempts to destroy the alien self within. Psychopathic personality is characterized by the predominance of a rigid and stable projection of the alien self into interpersonal relationships, the individual becoming the abuser tormenting others. Comorbid psychopathic and borderline personalities associate both types of projection leading to violence towards both self and others (Bateman & Fonagy, 2008). However, there are few studies examining the co-occurrence or symptom overlap of psychopathic and borderline personality traits or disorders in community, clinical or forensic samples of adolescents. In adults, moderate to high rates of comorbidity of borderline personality disorder with antisocial personality disorder have been reported in community and clinical samples of adults (Barrachina et al., 2011, Becker et al., 2000 and Grant et al., 2004). There are only two typological studies based on psychopathic and borderline traits in adults. Among violent male offenders, cluster analysis identified a borderline psychopathic group which had more frequent histories of affective and anxiety disorders than the non-borderline psychopathic groups (Blackburn & Coid, 1999). Similarly, among incarcerated adults with psychopathy, cluster analysis yielded two groups, the second one being characterized by more borderline traits, fewer psychopathic traits and higher trait anxiety (Skeem, Johansson, Andershed, Kerr, & Louden, 2007). These studies are consistent with the differentiation between primary psychopathy characterized by fearlessness and secondary psychopathy characterized by high levels of negative emotionality, in particular anxiety and depression (e.g., Lykken, 1995). There is no typological study aimed at identifying homogeneous groups of adolescents regarding psychopathic traits, borderline traits and antisocial behaviors. However, cluster analysis is a useful tool for developing empirically derived diagnostic groups that could be more objective, reliable, and predictive of outcome from different treatment (Borgen and Barnett, 1987 and Skinner and Blashfield, 1982). In particular, clustering techniques based on personality traits or dimensions are an underutilized means of exploring clinical heterogeneity of psychopathological disorders, and yielding empirically based personality typology that provides an important heuristic for personality theory, as well as useful information for the prediction of clinical outcome and for need-adapted treatment (Martel et al., 2010 and Morizot and Le Blanc, 2005). Cluster analyses may be useful to determine if there are subgroups of adolescents with specific profiles of psychopathic and borderline traits. Also, these analyses are helpful in assessing whether these subgroups of adolescents are varying in the levels of antisocial behaviors and emotional distress. At the theoretical level, cluster analyses may contribute to the debate about the relationships between psychopathic and borderline traits in adolescents. At the level of clinical implications, a better knowledge of psychopathological profiles might contribute to the understanding of antisocial behavior and to the identification of specific needs concerning prevention or treatment for psychopathologically distinct groups of adolescents. The aim of this study was to identify a typology of adolescents based on psychopathic and borderline traits in a community sample. We expected to find clusters that would differ in levels of antisocial behaviors, anxiety and depressive symptoms.

Results 3.1. Descriptive statistics Males scored higher than females on callous-unemotional traits (M = 32.7 and 30.3, SD = 5.6 and 5, respectively, t = 7.0, p < .001) and antisocial behaviors (M = 12 and 6.3, SD = 10.7 and 6.5, respectively, t = 9.2, p < .001). Females scored higher than males on borderline traits (M = 23.5 and 22.3, SD = 6.5 and 5.9, respectively, t = −3.5, p = .002), depressive symptoms (M = 22.8 and 19.3, SD = 11.2 and 12, respectively, t = −4.5, p < .001) and social anxiety (M = 19.7 and 17.7, SD = 6.4 and 6.6, respectively, t = −4.7, p < .001). Callous-unemotional traits and borderline traits were weakly associated (r = .14, p < .05). 3.2. Cluster analysis Cluster analysis was conducted in two steps to generate psychopathological profiles based on callous-unemotional and borderline traits scores converted to z-scores. In the first step, a hierarchical cluster analysis was conducted (Ward’s method with squared Euclidean distance). Based on the dendrogram and the agglomeration schedule, a four-cluster solution was identified ( Fig. 1). In the second step, K-means clustering was used to assign individuals to one of the identified clusters. The cluster solution consisted in one group that was below the mean on both traits (low traits cluster; n = 297; girls, 47%; boys, 53%), a second group that was above the mean in callous-unemotional traits and below the mean in borderline traits (moderate callous-unemotional traits cluster; n = 300; girls, 26%; boys, 74%), a third group that was well above the mean in borderline traits but low on callous-unemotional traits (borderline traits cluster; n = 236 girls, 54%; boys, 46%), and a fourth group that was high on both traits (high traits cluster; n = 138; girls, 24%; boys, 76%). The fourth group was the smallest (n = 138; 14% of the sample; girls, 24%; boys, 76%) and was high on both traits (high traits cluster). Cluster solution based on z-scores for callous-unemotional (CU) traits and ... Fig. 1. Cluster solution based on z-scores for callous-unemotional (CU) traits and borderline personality traits. The Y-axis represents the z-scores. Figure options The proportion of males in both the moderate callous-unemotional and the high traits clusters did not differ (χ2 = .22, p = .64) and were higher than in the total sample (χ2 = 16.6, p < .001 and χ2 = 11.6, p < .001, respectively) and in the low traits and the borderline traits clusters (χ2 = 28.8, p < .001, χ2 = 43.2, p < .001, χ2 = 32.1, p < .001, χ2 = 31.8, p < .001, respectively). The proportion of males in both the low traits and the borderline traits clusters did not differ (χ2 = 2.3, p = .12) and were lower than in the total sample (χ2 = 6.4, p = .001 and χ2 = 17.4, p < .001, respectively). A discriminant analysis was performed using borderline and callous-unemotional traits as predictors of membership in the four clusters. It showed clear discriminations amongst the four clusters (Wilks’ λ = .117, approximate F(6, 1932) = 616.65, p < .0001) with 95.6% of the original grouped cases correctly classified. Differences in levels of callous-unemotional and borderline traits, antisocial behaviors, depressive symptoms, and social anxiety were tested using a 2 (gender) × 4 (cluster groups) MANOVA. The results showed that they were significant multivariate effects of gender (Wilks’ λ = .902, F(5, 959) = 20.80, p < .001) and cluster groups (Wilks’ λ = .140, F(15, 2647) = 183.11, p < .0001). Interaction effects of gender with cluster groups were not significant (Wilks’ λ = .986, F(15, 2647) = .88, p = .58). Further univariate F-tests indicated significant group differences for all dependent variables in both genders. Pairwise comparisons (using Tukey’s Honestly Significant Difference test) are shown in Table 1. The four clusters were clearly distinct on callous-unemotional and borderline traits as post-hoc test showed all differences were significant with the exception of borderline traits which did not differ between the low traits and the moderate callous-unemotional clusters. The high traits cluster had significantly higher levels of antisocial behaviors than all other clusters and the low traits cluster had the lowest level of antisocial behaviors. The moderate callous-unemotional traits and the borderline traits cluster had intermediate and similar levels of antisocial behaviors. The high traits cluster and the borderline traits cluster had higher levels of depressive symptoms than the low traits and the moderate callous-unemotional traits clusters which did not differ on this variable. Regarding social anxiety, the borderline cluster had the higher scores, the high traits cluster an intermediate score and the low traits and the moderate callous-unemotional traits clusters the lower scores. Table 1. Callous-unemotional (CU) traits, Borderline traits, Antisocial behaviors, Depressive symptoms, and Social anxiety in the Four Clusters. Variables Low traits Moderate CU t. Borderline t. High traits F M SD M SD M SD M SD CU trait 26.7a 2.6 34.3b 2.6 29.9c 3.1 40.5d 4.0 640.7⁎ Borderline traits 18.9a 3.5 19.1a 3.1 29.7b 4.2 27.1c 5.4 445.3⁎ Antisocial behaviors 16.9a 15.8 23.2b 17.9 22.47b 18.7 40.5d 27.8 27.3⁎ Depressive symptoms 16.5a 10.3 17.0a 8.6 27.2b 10.9 26.5b 14.6 59.6⁎ Social anxiety 17.3a 6.0 16.6a 5.7 21.9b 6.2 19.1c 7.5 34.3⁎ Note. A cluster mean is significantly different from another mean if they have different superscripts (Tukey’s post-hoc tests). ⁎ p < 0.001.