How to learn about the brain: A simple framework for a complex system

Neuroscience is a broad discipline which, although vital for deep understanding of mental function in health and disease, can be an intimidating and/or ostensibly clinically irrelevant aspect of psychiatric practice. This may be a consequence of what has thus far been a predominantly inductive approach to neurobiological research. The idea that sufficient empirical data collection would somehow result in emergence of a deep understanding of the brain and behaviour has thus far not been realized, resulting in (A) discounting of neurobiologically-informed clinical reasoning, but (B) substantial growth in the field of theoretical neuroscience over the past 3 decades. The cooperation between theoretical and experimental neuroscientists has resulted in significant discoveries, including the reward prediction error signal encoded by midbrain dopaminergic neurons (which are implicated in the neurobiology of many psychiatric disorders). Adopting the approach of theoretical neuroscience, this presentation seeks to outline a general framework for the study of psychiatric neurobiology, beginning with a description of the levels of complexity in the brain, followed by review of Marr's levels of analysis, which can inform the process by which observable behaviours can be mapped to their neural implementations. We then review a simplified model of human-environment interactions, with a focus on the specific elements of these interactions, and finally discuss the broad functional architecture of the brain within which these computations are implemented. The framework outlined in this presentation can be used for further discussion of topics psychiatric neurobiology, such that data can be interpreted within a translational hierarchy from brain to mind.