آیا جانبداری دقیق پیش بینی کننده واکنش پذیری اتونوم در واکنش به عوامل استرس زا هست؟

Abstract Biased processing of threatening information may play a casual role in the development of anxiety disorders. Even though empirical evidence points to the fact that preattentive bias can predict subjectively experienced distress in response to a stressor, it is still unknown whether it could be useful in predicting the physiological reactivity in response to a stressor. In the present study, the emotional Stroop task was used to measure preattentive bias. Whereas Stroop interference for masked threat words (i.e., preattentive bias) was found to be positively associated with emotional distress (self-reported) in response to a laboratory stressor, this association was reversed when the autonomic reactivity (electrodermal activity) was used as a measure of emotional response to the very same stressor. Also, neither of these effects were a function of pre-existing anxiety levels. The negative association between preattentive bias and autonomic reactivity corresponds to the autonomic inflexibility seen in clinical anxiety (or very high scores of trait anxiety) when exposed to stressful events. Results were discussed in terms of an inability to automatically inhibit the processing of threatening cues that seems to be a vulnerability marker for anxiety

Results No individual included in the Stroop task had a color naming error rate above 5%. There were no significant relationships found between trait anxiety (M = 41.2, S.D. = 9.64) and any of the Stroop interference indices (p > .17). When the relationships between the dependent variables were examined, there was a significant correlation between SCV and SCL (r = .43, p = .004). None of these autonomic measures were, however, significantly associated with ED (p > .13). When the associations between the predictor/covariates and the dependent variables were examined, the analyses (se Table 2) revealed, as expected, a significant correlation between pre-task state anxiety and emotional distress, and age was also significantly associated with both measures of autonomic reactivity. More importantly, the two threat indices for masked words (i.e., negative self-evaluative and physical) were strongly associated with emotional distress, and also significantly associated with one of the autonomic reactivity measures (i.e., SCV). In addition, whereas the threat index for masked physical words was significantly (p = .013) associated with the second measure of autonomic reactivity (i.e., SCL), the threat index for masked negative self-evaluative words was, however, only weakly associated (p = .14). Also, none of the dependent variables were significantly correlated with the index for masked positive self-evaluative words, nor significantly correlated with any of the indices for unmasked words (p > .07). Table 2. Bivariate correlations between predictors/covariates and the dependent variables; emotional distress (ED), skin conductance variance (SCV), skin conductance level (SCL). Measure ED SCV SCL Age −.26 −.33* −.32* State anxiety .36* −.12 −.13 Trait anxiety .30 −.14 −.17 Masked Stroop interference indices Negative self-evaluative words .46 −.43 −.15 Physical threat words .48** −.31* −.31* Positive self-evaluative words .15 −.09 −.17 Unmasked Stroop interference indices Negative self-evaluative words .28 −.06 −.08 Physical threat words .22 −.12 −.05 Positive self-evaluative words .08 −.16 −.17 * P < .05. ** P < .01. Table options 2.1. Predicting emotional responses to the stressor Based on the correlational analyses, the effects of threat indices for masked negative self-evaluative and physical threat words, controlling for current anxiety status and age, were primarily tested. Thus, to examine the main effects of trait anxiety and Stroop interference, and the interaction between trait anxiety and Stroop interference, hierarchical multiple regression analyses (separately for each threat word category) were conducted on ED. To control for pre-task levels of state anxiety, state anxiety (M = 38.2, S.D. = 7.14) was entered along with trait anxiety in step one. In step two, masked Stroop interference index was entered. Finally, in step three, the interaction terms (cross-product) of trait anxiety and Stroop interference index (centered scores) for masked words were entered. When the analyses were conducted on skin conductance reactivity data (SCV and SCL), the order of entries was identical, apart from age being entered in step one along with trait anxiety. This procedure was performed separately for each Stroop interference index (negative self-evaluative and physical threat) on each dependent variable. 2.1.1. Emotional distress in response to the stressor The model including Stroop interference for negative self-evaluative words and the model including Stroop interference for physical threat words were significant. As seen in Table 3, there was a significant increment in explained variance from step 1 to step 2 due to the effect of Stroop interference for masked threat words; higher interference scores being associated with higher emotional distress in response to the stressor, after trait and pre-task state anxiety being accounted for. There was, however, no significant increment in explained variance due to the inclusion of the interaction term of trait anxiety and Stroop interference in step 3 (p > .64). Table 3. Hierarchical multiple regression analyses predicting ED in response to the stressor. Predictor variables Criterion variable, ED Step 1 β Step 2 β Model 1 State anxiety .29 Trait anxiety .18 ΔR2 .16* State anxiety .28 Trait anxiety .12 Threat index (masked Negative self-evaluative words) .42** ΔR2 .17** R2 .33** Model 2 State anxiety .26 Trait anxiety .15 Threat index (masked Physical words) .44** ΔR2 .19** R2 .35** * P < .05. ** P < .005. Table options 2.1.2. Autonomic reactivity in response to the stressor When the SCV was used as criterion, again the model including Stroop interference for negative self-evaluative words as well as the model including Stroop interference for physical threat words was significant. As seen in Table 4, there was a significant increment in explained variance from step 1 to step 2 due to the effect of Stroop interference for masked threat words; higher interference scores being associated with lower autonomic reactivity in response to the stressor, after trait anxiety and age being accounted for. Table 4. Hierachical multiple regression analyses predicting changes in autonomic reactivity (as measured with skin conductance variance [SCV], skin conductance level [SCL]) in response to the stressor. Predictor variables Criterion variables SCV SCL Step 1 β Step 2 β Step 1 β Step 2 β Model 1 Trait anxiety −.08 −.10 Age −.31* −.27 ΔR2 .11 .09 Trait anxiety .02 −.05 Age −.41** −.32* Threat index (masked negative self-evaluative words) −.50** −.28 ΔR2 .23** .08 R2 .35** .17 Model 2 Trait anxiety −.03 −.05 Age −.40** −.37* Threat index (masked Physical words) −.39** −.46** ΔR2 .14** .20** R2 .26** .30** * P < .05. ** P < .005. Table options As seen in Table 4, when the SCL was used as criterion, there was a significant increment in explained variance from step 1 to step 2, due to the effect of Stroop interference for masked physical threat words, higher interference scores being associated with lower scores on SCL, after trait anxiety and age being accounted for. The model including threat index for masked negative self-evaluative words contributed only to a marginal significant increment in explained variance (p = .08) Finally, there was no significant increment in explained variance due to the inclusion of the interaction term of trait anxiety and Stroop interference in step 3 (p > .09). 2.1.3. Additional analyses The correlational analyses showed significant effects only on the Stroop interference indices for masked threat words, and as seen in Table 2, there were only weak correlations between measures of emotional responding and the threat indices for unmasked words and the index for masked positive self-evaluative words. However, despite the non-associations between these indices and emotional responding, a significant interaction effect between anxiety and a threat index for unmasked words is still a possible outcome. However, further explorative analyses revealed no significant increment in explained variance due to the interaction term of trait anxiety and Stroop interference (p > .22)