Accounting for Perceptual Uncertainty: Characterizing Heterogeneity

We constantly face uncertainty as we perceive our environment. How we take this uncertainty into account can be understood with the Bayesian framework: our brains infer the causes of our perceptions through a two-stage process, combining sensory likelihood with prior knowledge and updating priors; with uncertainty being accounted for at each stage. Significant heterogeneity exists in this process of perceptual inference, particularly between studies or between individuals. Inter-study differences may arise from the format of priors – explicit or implicit – which are associated with distinct behaviors and computations, although clear evidence for a dissociation is lacking. Psychological traits, such as psychotic, autistic, and anxious traits, may contribute, for their part, to inter-individual differences, but findings are inconsistent and sometimes contradictory. Finally, heterogeneity could lie in the implementation of inference.  The aim of this PhD was to characterize the heterogeneity in accounting for uncertainty. In this PhD, we established a dissociation between the use of explicit and implicit priors, we highlighted the effect of anxious traits on inference and we explored the neural mechanisms of the influence of implicit priors on inference.