Computational psychiatry research typically takes a cross-sectional approach. This is because longitudinal, and treatment-oriented studies are much more challenging to conduct, especially at the scale that is needed to estimate what are likely to be small effects. In this talk, I will describe emergent methods to more easily study variation in cognition and symptoms over time and through treatment. I’ll discuss recent efforts to gamify and optimise mainstay tasks in computational psychiatry that capture individual differences in model-based planning and metacognition. Reducing participant burden further, I will describe a ubiquitous ‘passive’ measure of cognition based on digital questionnaire response times and show the benefits and pitfalls of this approach. I’ll close by bringing multiple sources of remotely gathered data together in a treatment prediction model for depression.

Associate Professor at Trinity College Dublin

Claire Gillan – Shifting research from lab to life