Computational Psychiatry and the Bayesian Brain

This chapter considers recent advances in computational neuroscience that are especially relevant for psychiatry. We offer a review of computational psychiatry in terms of its ambitions, emerging domains of application, and promises for the future. Our focus is on theoretical formulations of brain function that accommodate subjective beliefs and behavior within formal (computational) frameworks—frameworks that can be grounded in neurophysiology down to the level of synaptic mechanisms. Understanding the nature and principles that underlie functional brain architectures is, we assume, essential for understanding and phenotyping psychopathology and its pathophysiological underpinnings. To illustrate computational approaches to psychiatric disorders, we focus on active (Bayesian) inference and predictive coding. Specifically, we try to explain how the basic principles of neuronal computation are being used to understand psychiatric phenomena, ranging from affiliative behavior and theory of mind in autism to abnormalities of smooth pursuit eye movements in schizophrenia.