Quantifying prior beliefs for affective decision making and their relation to symptoms of depression


Marc Guitart-Masip

Project description

In Bayesian decision theory, prior beliefs refer to the probabilities that a decision-maker attributes to qualities of the world before any new evidence is considered. Such prior beliefs about the world impact choices about affectively significant (i.e., rewarding and aversive) outcomes and may be unusually pessimistic in conditions such as depression. However, the probability distributions comprising these prior beliefs have not yet been inferred from people’s choices and there is no direct evidence that these distributions are explicitly represented in the brain.

In this project, two experiments will be carried out. In the first, decision-making tasks and Bayesian methods will be used to infer task invariant prior beliefs about the world in healthy human volunteers. The researchers will use fMRI and multivariate pattern analysis to identify the brain regions where these prior beliefs are represented.

In the second experiment, decision-making tasks and Bayesian methods will be used to infer prior beliefs about the world in depressed patients. The researchers will disentangle the effects of potential aberrant and pessimistic prior beliefs from other potential computational mechanisms that may also give rise to depressive symptoms. They expect to characterize the operation of alternative computational mechanisms that may heterogeneously contribute to depression in different patients and account for some of the variability in clinical outcomes typically observed in patients.

This project will advance the understanding of the computational and neuronal underpinnings of affective decision making in healthy individuals and how these processes are disturbed in depression.

The project is funded by the Marianne and Marcus Wallenberg Foundation.

Researchers participating in the project

  • Johan Lundberg, Karolinska Institutet, Department of Clinical Neuroscience
  • Mate Lengyel, Cambridge University, UK
  • Peter Dayan, Tübingen University, Germany

Amount and duration

SEK 6,000,000 for four years (2021-2025)