UCL Institute of Neurology

Date: Friday 5th July 2013, 16.15-17.15. Speaker: Klaas Enno Stephan

Venue: 4th floor seminar room, 12 Queen Square, London, WC1N 3BG (Wellcome Trust Centre for Neuroimaging).

For further details, please contact: peter.smittenaar.10@ucl.ac.uk.

Synopsis: Psychiatry deals with complex and heterogeneous diseases whose diagnostics largely relies on descriptive schemes, such as the Diagnostic and Statistical Manual of Mental Disorders (DSM), which are agnostic about disease mechanisms but define disorders by combinations of symptoms. While facilitating reproducible diagnoses, these schemes have limited predictive validity with regard to clinical outcome and treatment response. In this presentation, I will argue that three things are needed to address this problem: (i) a theory which proposes key dimensions of psychiatric disease in both physiological and computational terms; (ii) model-based assays for identification of subject-specific disease processes along these dimensions; and (iii) longitudinal validation studies in patients. This approach highlights the importance of concrete clinical problems for providing an independent benchmark against which models of brain disease should be evaluated. I will illustrate the potential of this framework by initial examples from clinical and pharmacological studies and outline the opportunities and challenges that lie ahead.