What can feelings tell us about human behavior that we cannot know from directly observable states alone? From a reinforcement learning perspective, adaptive behavior is calibrated based on deviations...

Does including emotions improve reinforcement-learning models? A recent EEG study by Heffner and colleagues presents separate neural signatures for reward and emotion prediction errors. This advance invites questions about, and even holds clues to, which ingredients of emotion and prediction errors most improve reinforcement-learning models.