Cognitive models of obsessive–compulsive disorder (OCD) posit dysfunctional appraisal of disorder-relevant stimuli in patients, suggesting disturbances in the processes relying on amygdala–prefrontal connectivity. Recent neuroanatomical models add to the traditional view of dysfunction in corticostriatal circuits by proposing alterations in an affective circuit including amygdala–prefrontal connections. However, abnormalities in amygdala–prefrontal coupling during symptom provocation, and particularly during conditions that require stimulus appraisal, remain to be demonstrated directly.

Amygdala–prefrontal connectivity was examined in unmedicated OCD patients during appraisal (v. distraction) of symptom-provoking stimuli compared with an emotional control condition. Subsequent analyses tested whether hypothesized connectivity alterations could be also identified during passive viewing and the resting state in two independent samples.

During symptom provocation, reductions in positive coupling between amygdala and orbitofrontal cortex were observed in OCD patients relative to healthy control participants during appraisal and passive viewing of OCD-relevant stimuli, whereas abnormally high amygdala–ventromedial prefrontal cortex coupling was found when appraisal was distracted by a secondary task. In contrast, there were no group differences in amygdala connectivity at rest.

Our finding of abnormal amygdala–prefrontal connectivity during appraisal of symptom-related (relative to generally aversive) stimuli is consistent with the involvement of affective circuits in the functional neuroanatomy of OCD. Aberrant connectivity can be assumed to impact stimulus appraisal and emotion regulation, but might also relate to fear extinction deficits, which have recently been described in OCD. Taken together, we propose to integrate abnormalities in amygdala–prefrontal coupling in affective models of OCD.

The enhanced error monitoring in patients with obsessive–compulsive disorder (OCD), typically measured with the error-related negativity (ERN), has been found to be temporally stable and independent of symptom expression. Here, we examined whether the error monitoring in patients with OCD could be experimentally modulated by individually tailored symptom provocation.

Twenty patients with OCD and 20 healthy controls performed a flanker task in which OCD-relevant or neutral pictures were presented prior to a flanker stimulus. An individualized stimulus set consisting of the most provoking images in terms of OCD symptoms was selected for each patient with OCD. Response-locked event-related potentials were recorded and used to examine the error-related brain activity.

Patients with OCD showed larger ERN amplitudes than did control subjects in both the OCD-symptom provocation and neutral conditions. Additionally, while patients with OCD exhibited a significant increase in the ERN under the OCD-symptom provocation condition when compared with the neutral condition, control subjects showed no variation in the ERN between the conditions.

Our results strengthen earlier findings of hyperactive error monitoring in OCD, as indexed by higher ERN amplitudes in patients with OCD than in controls. Importantly, we showed that the patients’ overactive error-signals were experimentally enhanced by individually tailored OCD-symptom triggers, thus suggesting convincing evidence between OCD-symptoms and ERN. Such findings imply that therapeutic interventions should target affective regulation in order to alleviate the perceived threatening value of OCD triggers.