The Value of Information and Choice to Improve Control.

ended

Predictions and control about impending action and rewards or punishments require accumulating information about our environment and ultimately choosing the action that will maximize reward. Therefore, contexts containing a large amount of information about the environment and with several available options to gather reward or avoid punishment are generally preferred by agents. Because information intake and choice seeking are so important for survival and well-being, they can be considered as cognitive rewards, i.e. a higher order form of reward not directly related to immediate satisfaction like food or water. This proposal seeks to understand the neural mechanisms involved in the encoding of these cognitive rewards. How cognitive rewards like information and choice availability are encoded and manipulated in the brain remains a major unanswered question. Evidence suggests that cognitive rewards may be encoded by the dopaminergic system in a way similar to basic rewards like food or water. In contexts where the amount of basic reward will be equal, I will train monkeys to perform a task where they will get either more choice or more available information related to the type of upcoming basic reward. I hypothesize that monkeys will prefer conditions where more choice or more information are available compared to condition where these variables are lacking. Using simultaneous high-density electrophysiological recordings, we predict that midbrain dopaminergic neurons will reflect these preferences. By using fiber photometry targeting specifically dopaminergic neurons, we will also investigate how the neural signal related to cognitive reward are transmitted from midbrain dopaminergic neurons to the dorsolateral prefrontal cortex and the striatum, two structures receiving massive dopaminergic inputs and previously linked with different aspects of motivational and learning processes.