Dopemin Reward Circuits in the Mushroom Body?

i'm kind of wondering, thinking about the distributional reinforcement learning that mat talked about a there. You know, there might just be this trend toward more and more detail. So maybe shokyou you've done work in the mushroom body, drosophola, the oll factory, a since re processing area. And i really like the distributional story because it'sind of introducing one axis of variation here, but if it flies, it seems like there are at least three or four other access of variations. Can me to think about systems that are each learning a slightly different thing, and ten somehow combining each of these different predictions into a coherent action?