This chapter illustrates the power of the diverse emerging field using examples drawn from computational investigations of schizophrenia. It focuses on two main approaches within computational psychiatry: reinforcement learning and predictive processing. The chapter considers whether computational approaches can offer a plausible philosophical account of delusions. A computational model is then constructed to describe the problem, and the problem-solving process, in a mathematically rigorous way, allowing specific predictions as to the nature of the computational aberration underlying the clinical phenomenon. The chapter explores surveyed model-based and model-free approaches to reinforcement learning deficits in schizophrenia and examines the central role of dopamine in these processes, and how abnormalities in the dopamine system in schizophrenia are linked to these deficits. It considers the asymmetry of learning to maximize reward and learning to avoid punishment, and the link to negative symptoms associated with schizophrenia, particularly apathy and amotivation.