Decomposing conditioned avoidance performance with computational models. (October 2020)

• Record Type: Journal Article
• Authors: Krypotos, Angelos-Miltiadis; Crombez, Geert; Meulders, Ann; Claes, Nathalie; Vlaeyen, Johan W.S.
• Is Part Of: Behaviour research and therapy. Volume 133(2020)
• Publication Date: 2020-10
• Subjects: Pain -- Anxiety-related disorders -- Fear -- Escape -- Computational modeling; Cognitive therapy -- Periodicals; Psychotherapy -- Periodicals; 616.891
• DOI: 10.1016/j.brat.2020.103712
• Languages: English
• ISSNs: 0005-7967

Abstract

Abstract: Avoidance towards innocuous stimuli is a key characteristic across anxiety-related disorders and chronic pain. Insights into the relevant learning processes of avoidance are often gained via laboratory procedures, where individuals learn to avoid stimuli or movements that have been previously associated with an aversive stimulus. Typically, statistical analyses of data gathered with conditioned avoidance procedures include frequency data, for example, the number of times a participant has avoided an aversive stimulus. Here, we argue that further insights into the underlying processes of avoidance behavior could be unraveled using computational models of behavior. We then demonstrate how computational models could be used by reanalysing a previously published avoidance data set and interpreting the key findings. We conclude our article by listing some challenges in the direct application of computational modeling to avoidance data sets. Highlights: We argue that computational models can provide insights in avoidance's mechanisms. We support our arguments by reanalysing an avoidance learning data set. We describe some of the challenges of computational modeling. A paradigm shift is needed for identifying the key avoidance learning mechanisms.