Doubt in the Insula: Risk Processing in Obsessive-Compulsive Disorder

Abstract

Extensive cleaning or checking of patients with obsessive-compulsive disorder (OCD) are often interpreted as strategies to avoid harm and as an expression of the widespread belief that OCD patients are more risk-averse. However, despite its clinical significance, the neural basis of risk attitude in OCD is unknown. Here, we investigated neural activity during risk processing using functional magnetic resonance imaging and simultaneously assessed risk attitude using a separate behavioral paradigm in OCD patients with different symptoms versus healthy controls (HCs). We found opposite insula responses to high versus low risk in OCD patients compared to HCs: a positive correlation between insula activity and risk-aversion in patients versus a negative correlation in controls. Although OCD patients overall were not more risk-averse than controls, there were differences between subgroups of OCD patients: patients with doubt/checking symptoms were more risk-averse than other patients. Taken together, OCD patients show a reversed pattern of risk processing by the insula compared to HCs. Moreover, the data suggest that increased activation of the insula signals an abnormal urge to avoid risks in the subpopulation of OCD patients with doubt and checking symptoms. These results indicate a role for the insula in excessive risk-avoidance relevant to OCD.

Keywords: fMRI; insula; obsessive-compulsive disorder; risk avoidance; risk processing.

FIGURE 1

Schematic overview of scanner task. After viewing a fixation cross participants are presented with two gambles in the first period and choose the left or right side. In choice trials (A,B) the gambles differ in risk, in no-choice trials (C,D) they do not. After 4.5 s the choice of participants is represented by a red square around the chosen gamble in slide 2. Examples of trials: (A) choice trial, participant chooses low risk gamble (B) choice trial, participant chooses high risk gamble; (C) no-choice trial, participant ‘chooses’ high risk gamble (D) no-choice trial, participant ‘chooses’ low risk gamble.

FIGURE 2

Preferential correlations between risk-related activation and risk-aversion of patients compared to controls. (A) Stronger correlation between risk-aversion and brain activation in high risk > low risk contrast in OCD patients compared to healthy participants in the insula, dorsolateral prefrontal cortex, precentral and postcentral gyrus. (B) Patients show a positive correlation between insula activation in the high risk > low risk contrast and risk-aversion (green) while controls show a negative correlation (blue); the red cluster reveals the comparison between groups. (C) Patients show a positive correlation between dorsolateral prefrontal cortex activation in the high risk > low risk contrast and risk-aversion (green), the red cluster reveals the comparison between patients and controls.

FIGURE 3

Correlations between insula activation and risk-aversion. Positive correlation for (A) patients and negative correlation for (B) controls between insula activation and risk-aversion. Blue points in (A) represent patients within doubt/checking symptom dimension.