Talk title: Characterizing heterogeneity in strategy use during social learning and its relevance for psychiatric symptom dimensions.

We all live in a complex connected social world where the information we learn and the decisions we make do not occur in isolation but are instead strongly influenced by observing the behavior and actions of others around us. People also exhibit remarkable variability in their behavior and the strategies they deploy to solve social learning tasks; yet, robust characterizations of such variability are lacking. In this talk I will present the findings of two studies, as well as preliminary analyses of a third dataset, focusing on several methods that leverage large-scale online studies, computational modelling, and data-driven clustering, to better identify groups of individuals who deploy different computational mechanisms when learning and making decisions in social contexts. I will also show how such heterogeneity can help identify relevant variability in psychiatric symptom dimensions, as well as possibly dissociate symptom dimensions that tend to coexist in the population, such as autistic traits and anxiety. Finally, promises and pitfalls associated with these exploratory findings and methods will be discussed, together with broader implications for the field of computational psychiatry.

Read more here: https://osf.io/preprints/psyarxiv/pcjg7, https://osf.io/preprints/psyarxiv/mfc7z