اختلال تئوری ذهن در اختلال اضطراب اجتماعی

Abstract Social anxiety disorder (SAD) is a common psychiatric disorder characterized by a persistent, excessive fear and avoidance of social and performance situations. Research on cognitive biases indicates individuals with SAD may lack an accurate view of how they are perceived by others, especially in social situations when they allocate important attentional resources to monitoring their own actions as well as external threat. In the present study, we explored whether socially anxious individuals also have impairments in theory of mind (ToM), or the ability to comprehend others’ mental states, including emotions, beliefs, and intentions. Forty socially anxious and 40 non-socially-anxious comparison participants completed two ToM tasks: the Reading the Mind in the Eyes and the Movie for the Assessment of Social Cognition. Participants with SAD performed worse on ToM tasks than did non-socially-anxious participants. Relative to comparison participants, those with SAD were more likely to attribute more intense emotions and greater meaning to what others were thinking and feeling. These group differences were not due to interpretation bias. The ToM impairments in people with SAD are in the opposite direction of those in people with autism spectrum conditions whose inferences about the mental states of other people are absent or very limited. This association between SAD and ToM may have important implications for our understanding of both the maintenance and treatment of social anxiety disorder.

Results Preliminary Results The groups did not differ significantly in number of men and women, χ(1) = 3.38, p = .07, or in cognitive ability ( Table 2). Relative to non-socially-anxious comparison participants, those with SAD were older, t(78) = 3.34, p = .001, r = .35, and had higher levels of social anxiety and depression (see Table 2). The SAD group had an average LSAS score of 72.48 (SD = 22.35), thereby scoring above the clinical threshold ( Rytwinski et al., 2009). Fourteen of the 40 nonanxious comparison participants and 32 of the 40 participants with SAD met criteria for Axis I disorders on the MINI ( Table 3). The socially anxious group had more Axis I diagnoses (other than SAD; M = 1.55, SD = 1.13) than did the non-socially-anxious group (M = .48, SD = .75), t(78) = 1.08, p < .001. Table 2. Group Characteristics and Cognitive Ability SAD M (SD) Non-SAD M (SD) t(df) p r LSAS (anxiety severity) 72.48 (22.35) 26.03 (16.37) t(78) = 10.60 < .001* .77 CESD (depression severity) 20.70 (13.88) 6.35 (6.63) t(78) = 5.90 < .001* .56 NART: full-scale IQ 118.23 (5.10) 118.38 (3.66) t(70) = .14 .89 .02 NART: verbal IQ 118.63 (5.35) 118.97 (4.02) t(70) = .31 .76 .04 NART: performance IQ 113.31 (3.7) 113.51 (2.71) t(70) = .26 .79 .03 WAIS-IV: Similarities 1 10.73 (2.21) 11.33 (1.79) t(78) = 1.34 .19 .15 WAIS-IV: Matrix Reasoning 1 10.75 (2.73) 11.58 (2.76) t(78) = 1.34 .90 .15 Note. * = p ≤ .05 criteria; 1 = scaled scores. LSAS = Liebowitz Social Anxiety Scale; CESD = Center for Epidemiologic Studies – Depression; NART = National Adult Reading Test; WAIS-IV = Weschler Adult Intelligence Scale, Fourth Edition. Table options Table 3. Axis I Diagnoses Present in the Study Groups Diagnosis SAD n (%) Non-SAD n (%) Major Depressive Disorder 27 (67.5%) 10 (25%) Current 0 1 (2.5%) Lifetime 26 (67.5%)⁎ 9 (22.5%)⁎⁎ Generalized Anxiety Disorder 13 (32.5%) 1 (2.5%) Panic Disorder (with and without Agoraphobia) 12 (30%) 1 (2.5%) Substance Abuse/Dependence 6 (15%) 1 (2.5%) Bipolar Disorder (I or II) 3 (7.5%) 2 (5%) Post Traumatic Stress Disorder 2 (5%) 0 Agoraphobia (without Panic Disorder) 1 (2.5%) 3 (7.5%) Obsessive Compulsive Disorder 1 (2.5%) 0 Bulimia 0 1 (2.5%) ⁎ Six of the 26 individuals in the SAD group met criteria for a single past major depressive episode; the remaining 20 individuals met criteria for recurrent MDD. ⁎⁎ Five of the nine individuals in the non-socially anxious group met criteria for a single past major depressive episode; the remaining four individuals met criteria for recurrent MDD. Table options Theory of Mind To test the hypothesis that individuals with social anxiety have ToM impairments that appear either independent of or only when under cognitive load, we conducted a 2 (group: SAD vs. non-SAD) × 2 (cognitive load: high vs. low) Analysis of Variance (ANOVA) for each ToM task. The dependent variable for each ANOVA was the number of correctly answered questions on the MIE and the MASC, respectively. Because participants received the MIE twice (the first as a baseline measure, the second either under cognitive load or not), we conducted a repeated measures ANOVA for this task. The socially anxious group performed worse than the comparison group did on the MIE, F(1, 76) = 6.73, p = .01, r = .29, and participants in the load condition performed worse than those in the no-load condition, F(1, 76) = 5.10, p = .03, r = .22, whereas the Group × Cognitive Load interaction fell short of significance, F(1, 76) = 2.67, p = .11, r = .18 (see Figure 1). Likewise, there were no significant main effects or interactions for the repeated measure (MIE at time one versus MIE at time two). Performance on Reading the Mind in the Eyes Revised.Note. Dotted line represents ... Figure 1. Performance on Reading the Mind in the Eyes Revised. Note. Dotted line represents mean score of subjects with Asperger Syndrome or High Functioning Autism ( Baron-Cohen et al., 2001). Figure options To evaluate the type of errors participants made on the Mind in the Eyes task, we used our classification of the MIE items to determine if valence was related to group performance. A repeated measure ANOVA showed an interaction effect between group and the valence of the eyes in question. Using follow-up t-tests, we found that participants with SAD made significantly more errors than did comparison participants on questions about negative, t(78) = 3.40, p = .001, r = .36, valenced sets of eyes, and this difference remained significant after we applied a Bonferroni correction for multiple comparisons (p < .02). The groups did not differ, however, in the number of errors they made in response to positive, t(78) = .24, p = .81, r = .03, or neutral, t(78) = .72, p = .47, r = .08, valenced expressions. We categorized each incorrect answer on the MIE as more negatively valenced, more positively valenced, or the same valence as the correct answer. For example, if the correct answer for a particular item was positive in valence, but the participant chose an incorrect answer that was either neutral or negative in valence, this would be coded as a “more negative error.” Alternatively, if a participant chose an incorrect answer that was more positive in valence than the correct answer, this was coded as a “more positive error.” Finally, if someone chose an incorrect answer that was the same valence as the correct answer (e.g., the correct answer is negative in valence and the person chose an incorrect answer that was also negative in valence), this was coded as a “same valence error.” After a repeated-measures ANOVA indicated a significant interaction between group and error type, we performed follow-up t-tests, which revealed no group differences in the number of more positive, t(78) = .39, p = .70, r = .04, or more negative answers, t(78) = .61, p = .54, r = .07. However, socially anxious participants chose significantly more incorrect answers that were the same valence as the correct answer, t(78) = 3.5, p = .001, r = .37. This difference remained significant after we corrected for multiple comparisons (Bonferroni corrected p < .02). Analysis of MASC scores (with a second 2 × 2 ANOVA) showed a similar pattern of findings to the MIE, though performance on the two tasks was uncorrelated when we controlled for the presence of SAD (partial r = .17, p = .13). Participants with SAD were less accurate on the task than were non-socially-anxious participants, F(1, 76) = 9.37, p = .003, r = .33. Moreover, participants in the cognitive load condition performed worse overall on the MASC than did participants under no load, F(1, 76) = 4.02, p = .05, r = .22. There was no significant interaction between group and load, F(1, 76) = 1.12, p = .29, r = .12; see Figure 2). To determine the extent to which participants were taking the perspective of the individuals in this task, we examined the types of errors they made on the MASC with independent sample t-tests. On the MASC, participants with SAD made significantly more “excessive ToM” responses than did non-socially-anxious participants, t(78) = 2.92, p = .005, r = .19, but not more “no ToM,” t(78) = 1.11, p = .27, r = .12, or “less ToM,” t(78) = 1.01, p = .48, r = .11, errors. Finally, we compared groups’ performance on the MASC control questions to determine if the ToM impairments in the SAD group were specific to social information or generalized to nonsocial information. We found that socially anxious and comparison participants did not differ in their accuracy for control questions, F(1, 75) = 0.37, p = .54, r = .07, indicating that deficits in the SAD group occurred only for socially relevant information. Finally, there was a significant main effect of condition such that participants under cognitive load were less accurate in answering control questions than were participants in the no-load condition, F(1, 75) = 21.80, p < .001, r = .47. The interaction between group and condition was nonsignificant. Performance on the Movie for the Assessment of Social Cognition.Note. Dotted ... Figure 2. Performance on the Movie for the Assessment of Social Cognition. Note. Dotted line represents mean score of subjects with Asperger Syndrome ( Dziobek et al., 2006). Figure options To rule out the possibility that group differences on ToM tasks were driven by the presence of comorbid anxiety disorders, we repeated the main analyses after excluding participants with other anxiety disorders; we excluded 19 in the SAD group and 4 in the non-SAD group. We conducted another power analysis to confirm that we had sufficient power to conduct these analyses with the remaining 57 participants (power = .85 to detect a large effect). Our findings remained consistent. Using the full sample of participants, we also examined if there was an association between depressive symptoms, as measured by the CESD, and performance on the ToM. Depression was not significantly correlated with performance on either the MIE (r = -.03, p = .78) or the MASC (r = -.03, p = .82). We compared participants’ difficulty ratings of the memory task (range: 1–10) to determine if the task successfully taxed working memory for both the SAD and the non-SAD groups. Collectively, the participants gave the first memory task a mean difficulty rating of 7.82 and the second task a mean rating of 7.70. The groups did not differ on ratings for task one, t(36) = .58, p = .57, r = .13, or task two, t(36) = .56, p = .58, r = .12, nor did the groups differ in how accurately they recalled the symbols of memory task one, t(39) = .58, p = .57, r = .12, or two, t(35) = .42, p = .26, r = .11. Finally, to test the association between different factors that may be associated with ToM performance, we examined the relation between ToM accuracy and cognitive ability and sex. Scores on the NART and Similarities subtest predicted performance on the MIE (r = .23, p = .05 and r = .22, p = .06, respectively), but not on the MASC. Finally, men performed worse on the MASC than did women, t(78) = 3.14, p = .002, r = .33, though SAD still significantly predicted performance when we controlled for sex in a simultaneous multiple regression (β = -2.34, t[76] = 2.53, p = .014, r = .28). There was no difference between men and women’s performance on the MIE, t(78) = .27, p = .79, r = .09. Sex did not significantly interact with group (SAD vs. non-SAD) on either the MIE, F(1, 76) = .40, p = .53, or the MASC, F(1, 76) = .06, p = .82.