28/10/2016 - LPC - Learning reward uncertainty

     Vendredi 28 octobre 2016 - LPC
    11 heures, salle des Voûtes
    Rafal Bogacz
    Nuffield Department of Clinical Neurosciences
    University of Oxford

Learning reward uncertainty

To maximize their chances for survival, animals and humans need to base their decisions not only on the average consequences of chosen actions, but also on the variability of the rewards resulting from these actions. For example, when an animal’s food reserves are depleted, it should prefer to forage in an area where food is guaranteed over an area where the amount of food is higher on average but variable, thus avoiding the risk of starvation. To implement such policies, the animals need to be able to learn about variability of rewards resulting from taking different actions. This talk will present a simple mathematical model describing how such learning may be implemented in the basal ganglia. These models suggest how the information about reward uncertainty can be used during decision making, so that animals can make choices that not only maximize expected rewards but also minimize risks. The models account for a wide range of experimental data : from properties of individual dopaminergic receptors, to the effects of dopaminergic medications on choices involving risks, and they make multiple experimental predictions.