The integration of dual-systems processing and cognitive bias in cognitive vulnerability to anxiety: Investigations of spider fear

A recent model of cognitive vulnerability to anxiety proposes that discrepant findings related to the time-course of attentional responding to threat in anxious samples may be explained by individual differences in associative and rule-based processing. Specifically, stronger fear-relevant associations and maladaptive rule-based processing are hypothesized to contribute jointly to impaired disengagement from threat. The current program of research was designed as a test of this model within the context of spider fear/phobia. In Study 1, unselected participants completed measures of spider-fear associations (Go/No-go Association Task; GNAT), attention bias (Visual Search Task; VST), and self-reported spider fear and avoidance (Fear of Spiders Questionnaire; FSQ; & Spider Phobia Questionnaire; SPQ). Results demonstrated that participants were slower on disengagement trials of the VST compared to engagement trials. Surprisingly, higher levels of spider fear/avoidance were related to faster engagement with and disengagement from threat; stronger spider-fear associations were related to reduced disengagement accuracy. Moreover, an indirect relationship between spider-fear associations and reported spider fear/avoidance via disengagement accuracy was observed, such that stronger spider-fear associations predicted reduced disengagement accuracy, which subsequently predicted higher scores on the FSQ and SPQ. In Study 2, participants were randomly assigned into 1 of 2 conditions, wherein they either repeatedly negated the threat value of spiders or repeatedly reappraised spider stimuli as safe. As in Study 1, they completed the VST, FSQ, and SPQ. Again, all participants were slower on disengagement trials of the VST than on engagement trials. Attentional responding was unrelated to condition or to reported spider fear/avoidance. Moreover, there was no effect of condition on FSQ/SPQ scores. Exploratory analyses conducted within a restricted subsample of participants who obtained reasonable accuracy rates on the VST demonstrated that assignment to the negate condition predicted increased reported spider fear/avoidance. This relationship, however, was not mediated by attentional bias, contrary to hypotheses. For both studies, methodological limitations were noted, particularly with respect to the use of the VST. The utility of assessing fear with both indirect and direct measures is discussed, as well as theoretical and clinical implications for cognitive models of anxiety disorders and their associated treatments