ORIGINAL ARTICLE Reduced Orbitofrontal-Striatal Activity on a Reversal Learning Task in Obsessive-Compulsive Disorder Peter L. Remijnse, MD; Marjan M. A. Nielen, PhD; Anton J. L. M. van Balkom, MD, PhD; Danie¨lle C. Cath, MD, PhD; Patricia van Oppen, PhD; Harry B. M. Uylings, PhD; Dick J. Veltman, MD, PhD Context: The orbitofrontal cortex (OFC)–striatal cir- sponses relative to control subjects but showed ad- cuit, which is important for motivational behavior, is as- equate behavior on receipt of punishment and with re- sumed to be involved in the pathophysiology of obsessive- gard to affective switching. On reward outcome, patients compulsive disorder (OCD) according to current showed decreased responsiveness in right medial and lat- neurobiological models of this disorder. However, the en- eral OFC as well as in the right caudate nucleus (border gagement of this neural loop in OCD has not been tested zone ventral striatum) when compared with controls. Dur- directly in a cognitive activation imaging paradigm so far. ing affective switching, patients recruited the left poste- rior OFC, bilateral insular cortex, bilateral dorsolateral, Objective: To determine whether the OFC and the ven- and bilateral anterior prefrontal cortex to a lesser extent tral striatum show abnormal neural activity in OCD dur- than control subjects. No areas were found for which pa- ing cognitive challenge. tients exhibited increased activity relative to controls, and no differential activations were observed for punish- Design: A reversal learning task was employed in 20 pa- ment in a direct group comparison. tients with OCD who were not receiving medication and 27 healthy controls during an event-related functional Conclusions: These data show behavioral impairments magnetic resonance imaging experiment using a scan- accompanied by aberrant OFC-striatal and dorsal pre- ning sequence sensitive to OFC signal. This design al- frontal activity in OCD on a reversal learning task that lowed investigation of the neural correlates of reward and addresses this circuit’s function. These findings not only punishment receipt as well as of “affective switching,” confirm previous reports of dorsal prefrontal dysfunc- ie, altering behavior on reversing reinforcement contin- tion in OCD but also provide evidence for the involve- gencies. ment of the OFC-striatal loop in the pathophysiology of OCD. Results: Patients with OCD exhibited an impaired task end result reflected by a reduced number of correct re- Arch Gen Psychiatry. 2006;63:1225-1236 HE ORBITOFRONTAL COR- relevant stimulus on an object discrimi- tex (OFC) and the ventral nation reversal task.8,9 In a subsequent ex- striatum constitute the periment, a double dissociation in the pre- main components of 1 of a frontal cortex was observed: deficits on series of parallel, segre- affective switching but intact perfor- gated neural loops, which were first de- mance on attentional (extradimensional) Author Affiliations: T 1,2 scribed by Alexander et al. The func- switching were found in OFC-lesioned Departments of Psychiatry tional roles of these areas have been marmosets, whereas the opposite was (Drs Remijnse, Nielen, investigated extensively in both non- true for dorsolateral prefrontal cortex van Balkom, Cath, van Oppen, 10 and Veltman) and Anatomy human primates and humans. Electro- (DLPFC)–ablated animals. (Dr Uylings), VU University physiological studies in monkeys have In humans, research on the function of Medical Center, Amsterdam, the demonstrated that OFC neurons code the the OFC has focused primarily on rever- Netherlands; Graduate School context-dependent positive or negative sal learning and decision-making.11 Hu- of Neurosciences, Amsterdam reinforcement value of sensory stimuli3-6 man lesion studies have corroborated ani- (Drs Remijnse and Uylings); and register the rapid reversal of such mal experiments with respect to the Outpatient Academic Clinic for stimulus-reinforcement associations,3,7 disruption of reversal learning in OFC- Anxiety Disorders, GGZ which is important for motivational be- damaged patients12 and found a dissocia- Buitenamstel, Amsterdam 6 (Drs van Balkom, Cath, havior. Orbitofrontal involvement in re- tion in affective switching for patients with van Oppen, and Veltman); and versal learning (also termed affective OFC damage and those with DLPFC dam- 13,14 Netherlands Institute for Brain switching) had previously been shown in age. In addition, neuroimaging stud- Research, KNAW, Amsterdam OFC-ablated macaques, who exhibited ies in healthy subjects have repeatedly (Dr Uylings). perseverant responding to the previously shown the involvement of the OFC in the (REPRINTED) ARCH GEN PSYCHIATRY/ VOL 63, NOV 2006 WWW.ARCHGENPSYCHIATRY.COM 1225 ©2006 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/26/2021 processing of reward and punishment stimuli, either from was shown to recruit OFC and striatal regions in healthy a sensory quality15,16 or from an abstract (monetary) na- controls,53 data of which were also used in the present ture.17,18 Moreover, neuroimaging studies using reversal study. Since functional magnetic resonance imaging of learning paradigms have reported OFC activity during the OFC is notoriously difficult because of signal drop- affective switching.19,20 out,25,54 we applied a scanning sequence specifically sen- As stated earlier, the OFC is connected with the ven- sitive to OFC signal.55 Based on the previously reviewed tral sector of the caudate nucleus and these structures data on OFC-striatal function together with its pro- conjointly form a frontal-striatal circuit.1,2,6 Indeed, neu- posed role in the pathophysiology of OCD, we hypoth- roimaging studies have also demonstrated the ventral stria- esized that patients would show impaired performance tum to be engaged in reward processing21,22 and in affec- during the reversal learning task compared with control tive switching.23-25 Thus, the OFC and the ventral part subjects. Moreover, we expected that this would be ac- of the striatum are presumed to be crucial in an organ- companied by abnormal OFC-striatal activity during pro- ism’s processing of reward and punishment and in the cessing of reward, punishment, and affective switching. ability to alter behavior on changing stimulus- reinforcement contingencies, ie, in affective switching. METHODS Recent neurobiological models of obsessive- compulsive disorder (OCD) have stressed the role of dys- functional OFC-striatal circuitry in the pathogenesis of SUBJECTS this disorder26-29 based on several observations. First, from a phenomenological point of view, reward and punish- Twenty patients with OCD (14 women; mean age, 34 years; range, 19-54 years) and 27 healthy controls (19 women; mean ment perception appear to be abnormal in OCD; ie, pa- age, 32 years; range, 22-53 years) participated in this study. Pa- tients with OCD give the impressions of having an on- tients were recruited from the outpatient clinic for anxiety dis- going error sensation (“something is wrong”) when orders and by advertisements on the internet. Diagnoses were experiencing obsessions27 and of feeling insufficiently re- established by experienced clinicians with the Structured Clini- lieved by compulsive behavior that serves a rewarding cal Interview for DSM-IV Axis I disorders.56 Exclusion criteria goal.27,29 Moreover, the rigid behavior exhibited by pa- were the presence of alcohol or substance abuse and major in- tients with OCD that appears insensitive to reinforcing ternal or neurological disorders. The following comorbid dis- signals can be thought of as reflecting an inability to per- orders were diagnosed with the Structured Clinical Interview form affective switching. Second, neuropsychological tasks for DSM-IV Axis I disorders: major depressive disorder (n=7), that specifically address OFC function have shown im- dysthymia (n=4), social phobia (n=3), generalized anxiety disorder (n=3), panic disorder (n=2), agoraphobia (n=1), and paired performance in patients with OCD compared with 30,31 32,33 posttraumatic stress disorder (n=1). Moreover, comorbid healthy controls (but see other resources ). Third, Tourette disorder was clinically diagnosed in 2 patients, whereas structural and functional neuroimaging studies have re- 5 patients were diagnosed with “pure” OCD. At the time of the peatedly shown abnormalities associated with these brain study, all patients and control subjects were free of psycho- areas in OCD, although these findings have not been uni- tropic medication for at least 2 weeks and, in case of fluox- form: ie, increased34 or decreased35 OFC volumes and en- etine or antipsychotic medication, for at least 1 month. More- larged,35 normal,36 or diminished37 striatal volumes in mor- over, no patients were currently involved in a cognitive phometric studies in addition to either increased38,39 or behavioral therapy program. All participants gave written in- decreased40 activity in the OFC and hypoactivity41 or hy- formed consent and the study was approved by the ethical re- peractivity38,42 in the caudate nucleus during resting- view board of the VU University Medical Center (Amsterdam, the Netherlands). state imaging. Similarly, symptom provocation studies 43 To assess symptom characteristics and severity scores, the in OCD have demonstrated increased OFC activity next Yale-Brown Obsessive Compulsive Scale57 was administered (pa- 43,44 45 to both increased and decreased caudate activity. Fi- tients only), whereas the