Computational psychiatry

We apply computational and multivariate methods to psychiatric and mood disorders, where brain and cognitive changes are often subtle and heterogeneous. In bipolar disorder, we have used multivariate analysis to link patterns of cognitive impairment to structural brain differences, moving beyond traditional univariate, single-region comparisons. 

We have also set out future approaches to characterising cognition in bipolar disorder that can define cognitive phenotypes and relate them to brain abnormality. These quantitative descriptions of how an individual differs from the healthy population complement our wider normative-modelling work.