Striatal Mechanisms of Punishment Learning (21495)
The role of dopamine and the dorsal striatum in instrumental reward learning is well studied but its role in instrumental punishment learning remains poorly understood. Reinforcement learning accounts identify punishment as the opposite of reward. It predicts that punishers should reduce dopamine release in the striatum and increase the activity of dopamine 2 receptor-expressing neurons to support avoidance learning. To assess this, we used dLight fibre photometry to record dopamine binding in the dorsal striatum (dorsolateral [DLS], dorsomedial [DMS], tail [TS]) as well as GCaMP fibre photometry to record activity of dopamine 1 receptor (Drd1) and dopamine 2 receptor (Adora-2a) expressing neurons during punishment learning. Rats were trained to press two levers for food under a VI-30 schedule. Then one response was punished, via footshock, on an FR-10 schedule whereas the other response was not. There were action- related dopamine transients for both the punished and unpunished lever press. Punishment itself caused robust dopamine transients across the dorsal striatum. Within the DMS, punishment was associated with increased activity of Drd1 neurons but only a transient activation of Adora-2a neurons during early punishment sessions. Interestingly, punishment augmented both dopamine release and Drd1 activity in DMS during reward retrieval. These findings suggest that the profile of dopamine and neuronal activity during punishment learning is similar to instrumental reward learning. We suggest dopamine binding and Drd1 activity in the DMS report a teaching signal for instrumental learning that is unsigned with respect to the value of the outcome being learned about.