Reinforcement Learning And Feedback Inhibition