Title: More than a feeling: A comparison of amygdala and ventral striatum contributions to reinforcement learning

Abstract: If you ask a neuroscientist about emotion, they would probably give you the skinny about every amygdala study ever written. If you ask them about reward, they would probably give you a syllabus filled with references to the ventral striatum. But I’ll bet they can’t tell you what specific computations are computed in the amygdala relative to the ventral striatum. This is because traditional views of reinforcement learning have held that the striatum integrates dopaminergic reward prediction errors to facilitate learning, but have not ascribed a specific role for the amygdala. After decades of research suggesting the amygdala mainly contributes to emotion and simpler forms of associative learning, researchers are just beginning to explore the computational boundaries of amygdala function. The emerging perspective, which I will discuss, is that standard accounts of reinforcement learning that emphasize stratal and cortical contributions, should be revised to include a key role for the amygdala.