Altered reward-related effective connectivity in obsessive-compulsive disorder: an fMRI study

Abstract

Background: Obsessive–compulsive disorder (OCD) is characterized by anxiety-provoking, obsessive thoughts. Patients usually react to these thoughts with repetitive behaviours that reduce anxiety and are perceived as rewarding. Hence, reward plays a major role in the psychopathology of OCD. Previous studies showed altered activation in frontostriatal networks, among others, in association with the processing of reward in patients with OCD. Potential alterations in connectivity within these networks have, however, barely been explored.

Methods: We investigated a sample of patients with OCD and healthy controls using functional MRI and a reward learning task presented in an event-related design. Dynamic causal modelling (DCM) was used to estimate effective connectivity.

Results: Our sample included 37 patients with OCD and 39 healthy controls. Analyses of task-related changes in connectivity showed a significantly altered effective connectivity between the ventromedial prefrontal cortex (vmPFC) and the orbitofrontal cortex (OFC), among others, both in terms of endogenous connectivity as well as modulatory effects under positive feedback. Clinical measures of compulsion correlated with the effect of feedback input on visual sensory areas.

Limitations: The reported alterations should be interpreted within the context of the task and the a priori–defined network considered in the analysis.

Conclusion: This disrupted connectivity in parts of the default mode network and the frontostriatal network may indicate increased rumination and self-related processing impairing the responsiveness toward external rewards. This, in turn, may underlie the general urge for reinforcement accompanying compulsive behaviours.

Fig. 1

Experimental design.

Fig. 2

Localization of volumes of interest (VOIs) used in the dynamic causal modelling analysis. The regions illustrated correspond to 4 mm spheres at average Montreal Neurological Institute coordinates of the group of control participants. DS/Put = dorsal striatum/putamen; Ins = insula; OFC = orbitofrontal cortex; V1 = visual cortex; vmPFC = ventromedial prefrontal cortex.

Fig. 3

One of the 21 models constituting the model space, in this case with the modulatory effect between visual cortex and orbitofrontal cortex (OFC) for both forward and backward direction. The connections considered are backed up by previous anatomic studies, as indicated by the numbers1,, 2–4, 5, 6, 7 and 8., DS/Put = dorsal striatum/putamen; Ins = insula; l = left; r = right; V1 = visual cortex; vmPFC = ventromedial prefrontal cortex.

Fig. 4

(A) Model exceedance probabilities. (B) Winning model (highest exceedance probability) for patients with obsessive–compulsive disorder (OCD; model 3) and control (model 1) groups. DS/Put = dorsal striatum/putamen; Ins = insula; l = left; OFC = orbitofrontal cortex; r = right; V1 = visual cortex; vmPFC = ventromedial prefrontal cortex.

Fig. 5

(A) Compulsion measure of Yale–Brown Obsessive–Compulsive Scale (Y-BOCS) as a function of input effect on the left visual cortex for negative feedback in all patients, and (B) in medicated patients only. (C) Obsessive–Compulsive Inventory (OCI) as a function of input effect on the left visual cortex (V1) for negative feedback and for medicated patients only.