Representation of decision variables by catecholamines during learning
Our ability to learn is directly proportional to our ability to detect the difference between a prediction and its outcome. Reinforcement learning theories, which are widely used in neuroscience and machine learning, are based on this idea that learning only happens when a context falsely predicts an outcome, otherwise learning is null if the outcome matches the context prediction. Here, our goal is to establish the biological substrates for this learning signal. To do so, we study the synergy between catecholamine signals – dopamine and norepinephrine – in the medial prefrontal cortex, a brain center for complex behaviors. This project will leverage newly engineered fluorescent sensors for the rapid and specific detection of norepinephrine and dopamine binding, to assess the magnitude and time-course of catecholamine release in the cortex. By supporting this method with optogenetics, behavior, and genetic tools available in mice, we will determine how parallel dynamics of norepinephrine and dopamine support reinforcement learning and decision. Ultimately with these experiments, we will formulate a unified theory of catecholamine function in reward processing and decision-making. Additionally, this work will impact our understanding of the neuronal processes involved in mental disorders linked with NE and DA such as ADHD, schizophrenia, and depression.
Vincent Breton-Provencher , Université Laval
Partners and Donors