Poster Presentation

Poster C in Poster Session C: Wednesday, August 5, 9:30 – 11:15 am, Kimmel Center, Shorin & Rosenthal Rooms

Computational Mechanisms of Decision Uncertainty During Decision-Making in Worry

Mahalia Prater Fahey¹, Toby Wise¹; ¹King's College London

Presenter: Mahalia Prater Fahey

Worry is a symptom that is pervasive across mental health diagnoses, and when making decisions under uncertainty, individuals who worry display increased indecision. Previous work by Tallis & Mathews (1991), observed that increased levels of worrying were associated with longer decision times when determining that targets are absent and suggested this is driven by higher evidence requirements. Crucially, this decision making process can be captured by the drift diffusion model (DDM), whereby evidence is accumulated and a response is made after passing a response threshold. Here, we present a pilot study that aims to replicate Tallis & Mathews (1991)'s finding that individuals who worry more take longer to decide something is not present, and investigate whether this is driven by increased evidence requirements. Methods: To test this, 41 participants completed a visual search task online. In line with Tallis & Mathews (1991), participants were classified as either high or low worriers based on their total score on the worry domains questionnaire (High: 22; Low: 19). Participants completed 112 trials searching for a target letter among 25 letters (50% target present). Results: Replicating prior work, participants were slower overall to respond correctly on target absent trials compared to target present, and this effect was amplified in high worriers. Preliminary hierarchical drift diffusion modeling suggests that this behavior is best captured by a model where thresholds interact with worry group such that thresholds are selectively elevated for target absent trials, although the effect is not conclusive. Conclusion: In this pilot study we replicate Tallis & Mathews (1991)'s finding that high worriers take longer to respond when a target is absent and preliminary modeling points toward a threshold-based account. Follow-up work will test this account in a larger sample.

Topic Area: Development, Individual Differences & Clinical Populations