Keynotes | K&Ts | GACs | Talks | Posters | Search

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

A baby-step toward belief updating constructs: generalizability across two task variants

Magdalena Del Rio¹, Prashanti Ganesh¹, Noham Wolpe², Matthew Nassar¹; ¹Brown University, ²Tel Aviv University, University of Cambridge

Presenter: Magdalena Del Rio

Belief updating, defined as the process of revising one's expectations in the face of new evidence, is crucial for adaptive behavior and its disruption is implicated in many psychiatric conditions. Hence, different studies have used a wide range of tasks in order to operationalize the same underlying construct and link it to underlying neural mechanisms (Gibbs-Dean et al., 2023). However, it is unclear to what extent individual differences measured in the lab will generalize to individual differences in real-world belief updating. Here, we take a first step towards addressing this question by assessing whether individual differences in belief updating generalize across two close variants, namely the helicopter (McGuire et al., 2014) and the predator task (Satti et al., 2025). We asked participants recruited through Prolific (N=276) to complete both these task in one experimental session where the task order was counterbalanced. We hypothesized therefore that participant-level model-agnostic metrics and model-based parameters would correlate across task variants. Our preliminary results show that performance measures correlate across tasks along with mean single-trial learning rates and absolute estimation errors while mean absolute updates do not correlate. These results show that descriptive belief updating metrics, with the exception of absolute updates, do generalize across two task variants, laying the groundwork for a model-based analysis as well as more demanding tests of convergent validity using more dissimilar tasks, naturalistic and clinically relevant measures.

Topic Area: Development, Individual Differences & Clinical Populations