Recent developments in reinforcement learning, cognitive control, and systems neuroscience highlight the complimentary roles in learning of valenced reward prediction errors (RPEs) and non-valenced salience prediction errors (PEs) driven by the...