Threat belief determines the degree of costly safety behavior: Assessing rule-based generalization of safety behavior with a dimensional measure of avoidance. (September 2022)

• Record Type: Journal Article
• Authors: Wong, Alex H.K.; Pittig, Andre
• Is Part Of: Behaviour research and therapy. Volume 156(2022)
• Publication Date: 2022-09
• Subjects: Fear conditioning -- Safety behavior -- Fear generalization -- Rule-based generalization -- Threat beliefs -- Extinction; Cognitive therapy -- Periodicals; Psychotherapy -- Periodicals; 616.891
• DOI: 10.1016/j.brat.2022.104158
• Languages: English
• ISSNs: 0005-7967

Abstract

Abstract: Excessive generalization of safety behavior to innocuous stimuli that resemble a feared stimulus is oftentimes pathological especially with inflicted impairments. Safety behavior is conventionally assessed dichotomously, requiring multiple presentations of each test stimulus for assessing the proportion of safety behavior executed. Thus, the generalization gradient confounds with ongoing extinction learning during non-reinforced test trials. The present study employed a recently developed dimensional measure of avoidance to examine the extent of safety behavior generalization. We found that a dimensional measure of avoidance was able to assess the generalization gradients of safety behavior even when each test stimulus was presented once, thus minimizing the effect of ongoing extinction learning. Of equal importance is whether higher-order cognitive processes shape generalization of safety behavior. We found a range of distinct generalization gradients in safety behavior, which were highly consistent with participants' verbally reported relational rules. This rule-based generalization parallels to how clinically anxious individuals develop different threat beliefs after trauma exposure, and models how these distinct threat beliefs determine the extent of safety behavior engagement. Highlights: A dimensional measure of avoidance assessed generalization of safety behavior (SB). This measure minimizes confounding extinction learning. Distinct generalization gradients of SB corresponded with inferred relational rules. This finding aligns with how distinct threat beliefs shape the extents of SB.