پیش بینی ویژگی های اختلال شخصیت مرزی و ضد اجتماعی در افراد غیر بالینی با استفاده از اقدامات تکانشگری و پرخاشگری

Abstract This study examines impulsivity and aggressiveness dimensions as predictors of borderline (BPD) and antisocial (ASPD) personality disorder symptoms in nonclinical subjects. A total of 747 undergraduate university students were administered the Personality Diagnostic Questionnaire-4+, the Barratt Impulsiveness Scale-11, and the Buss-Durkee Hostility Inventory. Hierarchical regression analysis showed that motor impulsiveness, irritability, resentment, and guilt predicted BPD symptoms among university students after controlling for the effect of ASPD and depressive symptoms. ASPD symptoms were predicted by motor impulsiveness, physical aggression, indirect aggression, and negativism. These results indicate that in nonclinical subjects BPD and ASPD symptoms share a common impulsivity dimension but are linked to different aggressiveness facets.

. Introduction Although impulsivity and aggressiveness are personality characteristics relevant to various aspects of psychopathology (Lish et al., 1996 and Moeller et al., 2001), it is noted in DSM-IV (American Psychiatric Association, 1994) as well as in recent studies that these personality dimensions are key features of borderline (BPD) and antisocial (ASPD) personality disorders. It has been shown that impulsivity measures significantly predict the temporal stability of BPD diagnosis (Links et al., 1999) and discriminate BPD subjects from controls (Dougherty et al., 1999). Moreover, studies on sociality in BPD subjects showed that impulsive aggression is an important factor in suicide attempts in subjects with a BPD diagnosis (Soloff et al., 1994, Mann et al., 1999 and Soloff et al., 2000). Other studies reported significant associations between measures of impulsivity and ASPD (Eysenck, 1993 and Barratt et al., 1997), suggesting that impulsivity could link ASPD to BPD; however, the degree of impulsivity showed some variability among ASPD subjects (Moeller et al., 2001). Interestingly, even though impulsivity is currently considered a multidimensional construct (Eysenck, 1993, Patton et al., 1994 and Moeller et al., 2001), few if any data are currently available on the relative relationships between specific impulsivity components and BPD and ASPD. There are empirical data showing that subjects with ASPD manifest increased anger and aggression (Angst and Clayton, 1986, Bland and Orn, 1986, Haertzen et al., 1990, Dinwiddie, 1992, Hart et al., 1993 and Dinwiddie and Bucholz, 1993). Research data indicating an association between BPD and aggressiveness are sparser than those for ASPD (Lish et al., 1996). Although an association between BPD and actual violent acts has been reported (Raine, 1993), other studies showed that BPD is associated with the emotional component of aggressiveness (i.e. irritability) rather than with physical aggression (Gardner et al., 1991). As a whole, the results of these studies show a general link of both BPD and ASPD to impulsivity and aggressiveness, but give no definitive answer to whether BPD and ASPD subjects share the same or different impulsive and aggressive features. Starting from the above observations, the aim of this study was to evaluate the predictive role of different components of impulsivity and aggressiveness on BPD and ASPD features in a large sample of nonclinical subjects. The existing literature indicates that many of the thresholds for categorical personality disorder diagnoses are fairly arbitrary (Widiger, 1992), and that meaningful individual differences can be observed beyond the simple presence or absence of a categorical personality disorder diagnosis (Klein, 1993 and Widiger and Frances, 1985). Hence, in this study dimensional (i.e. number of symptoms) BPD and ASPD diagnoses were retained for further analyses. Since depressive symptoms frequently co-occur with BPD features (Elliott and Gunderson, 1988) and are reported to influence personality disorder measures, particularly self-report questionnaires (Loranger et al., 1991 and Zimmerman, 1994), the confounding role of depressive symptoms in the relationships of impulsivity and aggressiveness components to BPD and ASPD features was also analyzed in this study.

3. Results 3.1. Descriptive statistics As would be expected in a non-clinical sample, the average number of PDQ-4+BPD (mean=3.08, S.D.=1.95) and ASPD (mean=1.32, S.D.=1.45) symptoms observed in this sample of Italian university students was relatively small. The K-R 20 coefficient values of the PDQ-4+BPD and ASPD scales were 0.61 and 0.57, respectively. Female subjects scored significantly higher than males on the PDQ-4+BPD scale (t745=3.20, P<0.005). In contrast, the scores of female subjects on the ASPD scale were significantly lower than the scores of male subjects on the PDQ-4+ASPD scale (t745=−3.94, P<0.001). The descriptive statistics for the BIS-11 and BDHI subscales are listed in Table 1. No significant effect of subjects’ gender was observed on the BIS-11 subscale scores or on the BIS-11 total score (mean=64.11, S.D.=10.08, α=0.79). One-way MANOVA showed a significant multivariate difference between male and female subjects on the BDHI scale score, but Bonferroni-corrected multiple F tests evidenced that only three BDHI subscales (namely, Physical Aggression, Irritability, and Guilt) actually significantly discriminated male subjects from female subjects. In particular, male subjects scored significantly higher than female subjects on physical aggression. In contrast, female subjects scored significantly, but moderately, higher than male subjects on Irritability and Guilt subscales. Table 1. Barratt Impulsiveness Scale-11 and Buss-Durkee Hostility Inventory descriptive statistics N of items Whole sample (N=747) Male subjects (n=265) Female subjects (n=482) Mean S.D. α Mean S.D. α Mean S.D. α Barratt Impulsiveness Scale-11 subscales Motor impulsivenss 11 16.80 3.40 0.64 16.63a 3.55 0.63 16.89a 3.31 0.64 Attention 10 20.93 4.31 0.52 20.72a 4.23 0.57 21.05a 4.36 0.50 Lack of planning 9 26.37 4.64 0.55 26.25a 4.75 0.53 26.43a 4.58 0.56 Buss-Durkee Hostility Inventory subscales Physical aggression 10 3.22 2.48 0.74 3.82a 2.58 0.74 2.90b 2.37 0.73 Indirect aggression 9 4.58 1.84 0.55 4.53a 2.02 0.59 4.61a 1.74 0.53 Irritability 11 5.94 2.57 0.66 5.31a 2.48 0.63 6.28b 2.55 0.66 Negativism 5 2.69 1.43 0.52 2.67a 1.38 0.44 2.70a 1.46 0.56 Resentment 8 3.05 2.01 0.64 2.89a 2.01 0.63 3.15a 2.01 0.64 Suspiciousness 10 3.72 2.28 0.64 3.54a 2.40 0.68 3.86a 2.21 0.61 Verbal aggression 13 8.10 2.30 0.56 8.02a 2.30 0.56 8.14a 2.31 0.56 Guilt 9 3.74 2.08 0.61 3.40a 1.91 0.52 3.92b 2.15 0.64 α=Cronbach α/Kuder-Richardson 20 reliability coefficients. Multivariate mean comparisons: 1. Barratt Impulsiveness Scale-11 scores: Pillai V=0.00, F3,743=0.46, P>0.70; 2. Buss-Durkee Hostility Inventory scores: Pillai V=0.11, F8,738=11.51, P<0.001. Means in the same row that do not share the same superscripts differ at Bonferroni-corrected nominal P-level (i.e. P<0.0045) in the Bonferroni-corrected multiple F tests. Table options The mean SCL-90 Depression scale score observed in this sample was only 9.84 (S.D.=8.38); the Cronbach α value for this scale was 0.87. A significant effect of subjects’ gender on the intensity of self-report depressive symptoms was observed, with female subjects scoring significantly higher than male subjects on the SCL-90 Depression scale (t745=4.79, P<0.001). 3.2. Correlation and regression analyses As shown in Table 2, the PDQ-4+BPD and ASPD symptoms were not independent. The PDQ-4+BPD and ASPD symptoms showed significant positive correlations with all BIS-11 and BDHI subscale scores, as well as with the BIS-11 total score (rs=0.37 and 0.36, respectively, all Ps<0.001). Depressive symptoms were significantly and substantially correlated with BPD symptoms, whereas the small correlation between ASPD and depressive symptoms became nonsignificant once the effect of BPD symptoms was controlled for (partial r=0.02, P>0.50). Interestingly, the SCL-90 Depression scale correlated significantly with several BIS-11 and BDHI scales, suggesting that depressive symptoms can potentially influence the subjects’ responses to these scales and influence the relationships of BPD and, to a lesser extent, ASPD with aggressive and impulsive features, respectively. The BIS-11 subscales were all significantly intercorrelated. With the exception of the correlations between guilt, and physical aggression, negativism, and verbal aggression, the other correlations observed among the BDHI subscales were moderate and significant. As expected, several significant correlations were observed between the BIS-11 and the BDHI scales (median r=0.21). Table 2. Intercorrelations between borderline and antisocial personality disorder symptoms, depressive symptoms, Barratt Impulsiveness Scale-11 scores, and Buss-Durkee Hostility Inventory scores in university students (N=747) Scales 1 2 3 4 5 6 7 8 9 10 11 12 13 14 1. Borderline personality disorder (PDQ-4+) – 2. Antisocial personality disorder (PDQ-4+) 0.37* – 3. Depression (SCL-90) 0.54* 0.22* – 4. Motor impulsivenss (BIS-11) 0.36* 0.35* 0.21* – 5. Attention (BIS-11) 0.34* 0.32* 0.30* 0.49* – 6. Lack of planning (BIS-11) 0.21* 0.23* 0.14* 0.56* 0.43* – 7. Physical aggression (BDHI) 0.33* 0.40* 0.10 .26* 0.25* 0.15* – 8. Indirect aggression (BDHI) 0.38* 0.34* 0.19* 0.30* 0.25* 0.17* 0.40* – 9. Irritability (BDHI) 0.53* 0.31* 0.38* 0.29* 0.35* 0.16* 0.39* 0.50* – 10. Negativism (BDHI) 0.21* 0.30* 0.08 0.17* 0.19* 0.09 0.34* 0.26* 0.32* – 11. Resentment (BDHI) 0.52* 0.25* 0.44* 0.25* 0.28* 0.12* 0.31* 0.33* 0.48* 0.30* – 12. Suspiciousness (BDHI) 0.38* 0.25* 0.32* 0.13* 0.21* 0.00 0.29* 0.26* 0.42* 0.27* 0.54* – 13. Verbal aggression (BDHI) 0.30* 0.28* 0.09 0.26* 0.20* 0.15* 0.45* 0.42* 0.46* 0.32* 0.19* 0.19* – 14. Guilt (BDHI) 0.38* 0.12* 0.38* 0.13* 0.21* 0.02 0.05 0.16* 0.29* 0.11 0.36* 0.36* 0.05 – PDQ-4+=Personality Diagnostic Questionnaire-4+; SCL-90=Symptom Checklist-90; BIS-11=Barratt Impulsiveness Scale-11; BDHI=Buss-Durkee Hostility Inventory. Considering the large number of comparisons, the significance level for the correlation coefficients was set to P<0.001. *P<0.001. Table options Both the SCL-90 Depression scale and PDQ-4+ASPD scale scores significantly predicted the PDQ-4+BPD scale scores (R2=0.36, P<0.001). In order to evaluate whether the overall level of impulsivity and aggressiveness significantly predicted the number of BPD symptoms, the BIS-11 total score and the BDHI composite score (α=0.79), which was obtained summing all BDHI scales with the exception of the Guilt scale, were entered in the regression equation. The change in R2 value was substantial (ΔR2=0.12, R2=0.48, adjusted R2=0.48, all Ps<0.001); both the BIS-11 total score (β=0.11, P<0.001) and the BDHI composite score (β=0.37, P<0.001) significantly predicted the PDQ-4+BPD scale score. In the first step of the hierarchical regression analysis, only the PDQ-4+BPD scale score (β=0.36, P<0.001) significantly predicted the number of ASPD symptoms (R2=0.14, P<0.001). When the BIS-11 total score and the BDHI composite score were entered as predictors in the regression equation, the increase in the proportion of PDQ-4+ASPD scale variance explained by the regression model was significant (ΔR2=0.12, R2=0.26, adjusted R2=0.26, all Ps<0.001). It should be observed that the amount of variance explained by the final regression model was smaller than that reported for BPD symptoms. However, the relatively low frequency of ASPD symptoms observed in this sample could at least partially explain this finding. Both the BIS-11 total score (β=0.21, P<0.001) and the BDHI composite score (β=0.31, P<0.001) significantly predicted the number of ASPD symptoms measured by the PDQ-4+. As a whole, these results showed a general link of both BPD and ASPD symptoms with the overall levels of impulsivity and aggressiveness. To clarify these broad relationships, more detailed analyses involving specific components of impulsivity and aggressiveness were performed. The results of hierarchical regression analyses for BIS-11 and BDHI scales predicting BPD symptoms are listed in Table 3. Although the BIS-11 and BDHI scales were significantly intercorrelated, tolerance, VIF and CI values did not evidence the presence of collinear variables. While the variables entered into Step 1 explained roughly 36% of the PDQ-4+BPD score variance, entering the BIS-11 and BDHI subscales in Step 2 substantially increased the amount of explained variance (ΔR2=0.16). After controlling for the effect of ASPD and depressive symptoms, as well as for the effect of the BIS-11 and BDHI subscale intercorrelations, the irritability, resentment, guilt, and motor impulsiveness subscales were the only variables entered in Step 2 that significantly predicted BPD symptoms. The Chow test was nonsignificant, indicating that the predictor slopes computed in male and female subjects were not significantly different. Table 3. Summary of hierarchical regression analysis for variables predicting borderline personality disorder symptoms in university students (N=747) Variables B SE β Step 1 Antisocial personality disorder (PDQ-4+) 0.36 0.04 0.27* Depressive symptoms (SCL-90) 0.11 0.01 0.48* Step 2 Motor impulsivenss (BIS-11) 0.05 0.02 0.11** Attention (BIS-11) −0.00 0.02 −0.01 Lack of planning (BIS-11) 0.01 0.01 0.02 Physical aggression (BDHI) 0.05 0.03 0.07 Indirect aggression (BDHI) 0.05 0.03 0.05 Irritability (BDHI) 0.13 0.03 0.18** Negativism (BDHI) −0.05 0.04 −0.04 Resentment (BDHI) 0.17 0.03 0.18** Suspiciousness (BDHI) 0.00 0.03 0.00 Verbal aggression (BDHI) 0.04 0.03 0.05 Guilt (BDHI) 0.11 0.03 0.12** PDQ-4+=Personality Diagnostic Questionnaire-4+; SCL-90=Symptom Checklist-90; BIS-11=Barratt Impulsiveness Scale-11; BDHI=Buss-Durkee Hostility Inventory. R2=0.36 for Step 1 (adjusted R2=0.36), P<0.001; R2=0.52 for Step 2 (ΔR2=0.16, adjusted R2=0.51), Ps<0.001. The Bonferroni-corrected nominal significance level for the individual regression coefficients was set to P<0.025 for Step 1, and to P<0.0045 for Step 2. *P<0.025. **P<0.0045. Table options The hierarchical regression analysis results for variables predicting ASPD symptoms are listed in Table 4. As a whole, the final model explained approximately 30% of the ASPD symptom variance. Motor impulsiveness was the only BIS-11 subscale significantly predicting the number of PDQ-4+ASPD symptoms, but differently from what was observed for BPD symptoms, physical aggression, indirect aggression, and negativism were the BDHI subscales significantly predicting PDQ-4+ASPD symptoms. No significant effect of SCL-90 Depression scale scores on the PDQ-4+ASPD symptoms was observed. The Chow test indicated that the regression slopes for the variables predicting ASPD symptoms were significantly different in male and female subjects (F13,721=3.67, P<0.001), even though univariate homogeneity tests showed that only motor impulsiveness (t743=2.41, P<0.05) and resentment (t743=2.40, P<0.05) regression coefficients were significantly different in these sub-samples. Table 4. Summary of hierarchical regression analysis for variables predicting antisocial personality disorder symptoms in university students (N=747) Variables B SE β Step 1 Borderline personality disorder (PDQ-4+) 0.27 0.03 0.36* Depressive symptoms (SCL-90) 0.00 0.03 0.03 Step 2 Motor impulsivenss (BIS-11) 0.04 0.01 0.13** Attention (BIS-11) 0.04 0.02 0.08 Lack of planning (BIS-11) 0.01 0.01 0.04 Physical aggression (BDHI) 0.12 0.02 0.21** Indirect aggression (BDHI) 0.09 0.03 0.11** Irritability (BDHI) −0.03 0.03 −0.05 Negativism (BDHI) 0.14 0.04 0.14** Resentment (BDHI) −0.06 0.03 −0.08 Suspiciousness (BDHI) 0.05 0.03 0.09 Verbal aggression (BDHI) 0.01 0.02 0.01 Guilt (BDHI) −0.02 0.03 −0.04 PDQ-4+=Personality Diagnostic Questionnaire-4+; SCL-90=Symptom Checklist-90; BIS-11=Barratt Impulsiveness Scale-11; BDHI=Buss-Durkee Hostility Inventory. R2=0.14 for Step 1 (adjusted R2=0.14), P<0.001; R2=0.30 for Step 2 (ΔR2=0.16, adjusted R2=0.29), Ps<0.001. The Bonferroni-corrected nominal significance level for the individual regression coefficients was set to P<0.025 for Step 1, and to P<0.0045 for Step 2. *P<0.025. **P<0.0045. Table options Since the association between irritability and BPD, as well as the association between physical aggressiveness and ASPD, might be an artifact due to the consistency in answering to similar items on the BDHI and the PDQ-4+, the data were re-analyzed excluding the items assessing irritability and physical aggressiveness from BPD and ASPD scale scores, respectively. The model fitted adequately for BPD, although the R2 values were somewhat smaller than the corresponding R2 values listed in Table 3 [R2=0 36 for Step 1 (adjusted R2=0.36), P<0.001; R2=0.48 for Step 2 (ΔR2=0.12, adjusted R2=0.47), Ps<0.001]. Interestingly, irritability still remained a significant predictor of BPD symptoms, although its predictive role was slightly attenuated (β=0.13, P<0.0045). Similar considerations also hold for ASPD. The model fit statistics were as follows: R2=0 13 for Step 1 (adjusted R2=0.13), P<0.001; R2=0.30 for Step 2 (ΔR2=0.17, adjusted R2=0.29), Ps<0.001. Physical aggressiveness remained a significant predictor of ASPD symptoms (β=0.19, P<0.0045).