Modelling brain reinforcement learning circuits in health and disease
M1 or M2 Internship · Paris, France
Address: 29 rue d’Ulm, 75005 Paris, France
Team: Mathematics of Neural Circuits
Adviser: Boris Gutkin
Length of internship: 1-6 months
Animals and humans learn to make nearly optimal decisions based on received and expected rewards they get from the environment. Recent data indicate that the prevalent learning signal for such reinforcement learning is the error between the expected (previously learned) and the received reward. In our team we recently developed a minimal biolgically based computational model for how these errors are computed in the brain. The model includes the dynamics of both the dopaminergic nuclei (the ventral segmental area) and the dynamics of the formation of the persistent-activity representations of the relevant stimuli in the prefrontal cortex. In this project, we will extend this model to understand the role of sensory learning during conditioning tasks. We will also extend our study to understand the role of addictive drug action in the dopaminergic brain system and in the prefrontal cortical circuits. Addictive drugs considered will be nicotine and alcohol and their interactions will be studied.