Computational Psychiatry: New Perspectives on Mental Illness

Abstract

Psychiatry faces a number of challenges due largely to the complexity of the relation-ship between mind and brain. Starting from the now well-justifi ed assumption that the mind is instantiated in the physical substrate of the brain, understanding this relation-ship is going to be critical to any understanding of function and dysfunction. Key to that translation from physical substrate to mental function and dysfunction is the com-putational perspective: it provides a way of translating knowledge and understanding between levels of analysis (Churchland and Sejnowski 1994). Importantly, the compu-tational perspective enables translation to both identify emergent properties (e.g., how a molecular change in a receptor affects behavior) and consequential properties (e.g., how an external sociological trauma can lead to circuit changes in neural processing). Given that psychiatry is about treating harmful dysfunction interacting across many levels (from subcellular to sociological), this chapter argues that the computational per-spective is fundamental to understanding the relationship between mind and brain, and thus offers a new perspective on psychiatry.