Choice and Information as Cognitive Rewards improving Control Level – CIRCLE

Predictions and control of impending reward require accumulating information about one’s environment and choosing actions that maximize the outcomes of these actions. Since information intake and choice seeking promote survival, agents should prefer contexts where these variables are available. Choice and information can be viewed as “cognitive rewards”, i.e. an abstract form of reward that is not directly related to immediate satisfaction such as food or water (basic rewards). In this research project, we seek to understand the neural mechanisms involved in cognitive reward encoding and their role in choice behavior. Humans and monkeys will express their preferences for cognitive rewards in experimental trials manipulating either choice or amount of information about upcoming basic reward. Combining this task with both computational modelling and functional neuroimaging (fMRI) in healthy adult subjects, we will first ask whether regions coding for basic rewards overlap with those encoding cognitive rewards, and whether these regions communicate with each other using functional and effective connectivity analyses. Our ambition here is to better understand how large-scale brain networks interact to shape behavioral preferences for either information or choice, or for both. Neuroimaging data will benefit from the modelling work carried out on behavioral data, which aims to provide a model-based characterization of such preferences through designing, testing and comparing different models of learning and decision-making. We will then use state-of-the-art neurophysiological techniques to investigate further how the dopaminergic system, known to encode basic reward, represents cognitive rewards in cortical and sub-cortical areas of non-human primates (NHPs). To do so we will perform high-density electrophysiological recordings in midbrain dopaminergic neurons. We will also use fiber photometry to measure specifically dopaminergic activity at their axon terminals in downstream structures including the striatum and prefrontal cortex. Finally, we will investigate how cognitive reward representations may be impaired when the dopaminergic system is altered in Parkinson’s disease patients, who will perform the task while they are ON or OFF treatment (levodopa and deep brain stimulation). Taken together, these data should significantly advance our understanding of the neurocognitive determinants of the so-called “intrinsic motivation”, and of why people value information and choice, as opportunities to exercise control over events that affect their lives.