(PS10-D82) Eye Tracking Evidence of Threat-related Attentional Bias in Anxiety- and Fear-related Disorders: A Systematic Review and Meta-analysis

Context: Cognitive theories of anxiety- and fear-related pathology suggest that individuals with these forms of pathology (versus those without) exhibit greater threat-related attentional bias (AB). However, there are a multitude of mixed and null findings in this area of research. Unlike other commonly used measures of AB, eye-tracking indices exhibit acceptable reliability, and thus, may help clarify the relationship between AB and anxiety- and fear-related symptoms.

Objective: The purpose of this study was to conduct an updated meta-analysis, to determine whether there is evidence of a relationship between anxiety and fear-related symptoms and expressions of threat-related AB (i.e., reflexive orienting and maintenance) measured via free-viewing eye-tracking tasks.

Method: MedLine, PsychArticles, and PsychInfo databases were used to identify articles that reported a relationship between anxiety- and fear-related symptoms and threat-related AB assessed via eye-tracking technology. The following search string was used: (anxi* or phobia or PTS* or panic or GAD) AND (eye tracking or eye-tracking or gaze or fixation or dwell time or saccade). Studies were excluded if a) the sample consisted of children and/or adolescents, b) anxiety groups were created via a manipulation, c) the study did not measure trait anxiety or anxiety symptoms consistent with the diagnostic and statistical manual of mental disorders (DSM; American Psychiatric Association, 2013), or d) a manipulation preceded the collection of eye tracking data.

Data Synthesis: A total of 40 articles were retained for this meta-analysis. Significant relations were observed between anxiety and fear-related symptoms and both reflexive orienting toward threat (r = .13, 95% CI: .03, .22) and maintenance of attention on threat (r = .15, 95% CI: .05, .25). Although both effects exhibited considerable heterogeneity (reflexive orienting: Q = 86.78, p < .001, I² = 64.28%; maintenance: Q = 105.93, p < .001, I² = 72.62%), power analyses revealed low power (.06 - .64) to detect moderator effects and as such moderator analyses were not conducted. Funnel plots indicated possible publication bias; however, fail-safe N indicated that a substantial number of null findings would be needed to render associations nonsignificant (reflexive orienting: N = 218, maintenance: N = 388).

Conclusions: Study findings provide support for both vigilance and maintenance hypotheses of AB. Additionally, small effect sizes may be the result of potential moderators such as individual differences in attentional control. Study results suggest that it may be important to develop attention bias modification interventions that target AB at both early (bottom-up) and later (top-down) stages of information processing to reduce anxiety- and fear-related pathology.