Full Program »
100. Searching for an Anchor in an Unpredictable World: A Bayesian Model of Obsessive Compulsive Disorder
Drawing from the well-developed framework of the Bayesian brain (predictive coding and active inference), we propose that the core computational dysfunction in OCD is an impairment in predicting transitions between states. Clinically, this corresponds to patients' difficulty in trusting their own actions, while being preoccupied with unlikely chains of events. We illustrate the face validity of this idea using numerical simulations, and use these to form specific empirical predictions. These predictions are evaluated in relation to existing evidence. We show how seemingly unrelated findings in OCD can be explained by excessive uncertainty over state transitions. For example, whereas many studies suggest perseveration and overreliance on habits, other studies suggest overly controlled (i.e. goal directed) and explorative behavior in OCD. Using active inference simulations, we show how both types of behavior flow naturally from increased uncertainty over state transitions. This implies that overreliance on habitual behavior is a secondary (compensatory) result of goal-directed behavior being highly unpredictable. Using simulations we also outline conditions in which OCD patients over-rely on habits, and conditions in which they exhibit overly volatile or explorative goal-directed behavior. Finally, we outline potential clinical and nosological implications and suggest lines for future research.