The neurocognitive correlates of compulsivity

Abstract

The shortcomings of classic psychiatric nosology, such as the vast similarity across and heterogeneity within disorder categories, have limited progress in psychiatric research. In this thesis, we outline three neurocognitive models proposed to be key to obsessive-compulsive disorder (OCD) (i.e. dysfunctions in goal-directed control, metacognition and error monitoring) but unfortunately these effects are inconsistent in OCD or generalisable to other disorder categories. Here we employed a transdiagnostic approach to resolve this issue by investigating if dysfunctions in these proposed neurocognitive correlates were captured by a transdiagnostic compulsive dimension in order to delineate a robust, replicable and specific neurocognitive characterisation of compulsivity. First, we found that the transdiagnostic framework was able to reveal metacognitive deficits to compulsivity which were previously not found in an OCD case-control reinforcement learning study. Specifically, high compulsive individuals exhibited inflated confidence and an insensitivity in adjusting their confidence reports with evidence. This gave the first hints that the mental model, i.e. the cognitive representation of environmental contingencies that goal-directed behaviour relies upon, was dysfunctional in compulsivity. We subsequently found that a neural representation of state transition probabilities critical for goal-directed performance, which also manifested as reduced reaction time sensitivity to state transitions, was diminished in high compulsive individuals. Altogether, these data suggest that compulsivity is characterised by failures in curating the mental model which may explain decision deficits common to OCD and high compulsive individuals. Finally, we did not observe any significant associations of dysfunctional error monitoring by the error-related negativity (ERN) with the compulsive dimension. Overall, this thesis supports goal-directed and metacognitive dysfunctions, but not ERN abnormalities, as neurocognitive correlates of compulsivity, and underscores the advantages of transdiagnostic methods in psychiatric research.