Psychonomic Bulletin & Review 1998,5(2),161-172

Issues surrounding the cognitive neuroscience of obsessive-compulsive disorder

KEVIN D. WILSON University ofPennsylvania, Philadelphia, Pennsylvania

Obsessive-compulsive disorder (OCD) has been studied extensively in recent years, with increased emphasis on understanding OCD's biological substrates. There has been significant progress in docu­ menting abnormal brain function in OCD patients, particularly in the orbitofrontal cortex, basal gan­ glia, and thalamus. Similar progress has broadened our understanding of the cognitive and behavioral manifestations ofthe disorder, including deficits in set shifting, hyperattention, and visuospatial con­ struction abilities. Unfortunately, these results have not been replicated consistently. This report com­ prises a review of previous attempts to characterize the neurobiology and neuropsychology of OCD, and a discussion ofseveral factors in OCD research that can help to explain previous inconsistencies.

Obsessive-compulsive disorder (OCD) is a relatively logical features ofOCD. Only then will an adequate the­ common form ofpsychopathology with a lifetime preva­ ory ofthe cognitive neuroscience ofOCD be possible. lence of2%-3% (Robins et al., 1984). Although the ana­ tomical and behavioral regularities of the disorder have PHENOMENOLOGY OF OeD been studied extensively over the past 20 years, there has been relatively little success in the attempt to establish OCD is grouped under the general heading ofanxiety how the former give rise to the latter. Popular models of disorders. The DSM-IV(American Psychological Asso­ OCD have drawn upon a variety ofdisciplines, including ciation, 1994) defines OCD as "recurrent obsessions or learning theory (Mowrer, 1960), cybernetics (Pitman, compulsions that are severe enough to be time consum­ 1987) and psychoanalysis (Fenichel, 1945). Often, how­ ing (i.e., they take more than 1hour a day) or cause marked ever, these models have had relatively little to say about distress or significant impairment." It defines obsessions the dysfunctional aspects ofthe OCD brain. This is due as "persistent ideas, thoughts, impulses, or images that in part to the fact that, until recently, the neuroanatomical are experienced as intrusive and inappropriate and that substrates ofOCD were poorly understood. Although OCD cause marked anxiety or distress." It defines compulsions researchers had long suspected the involvement of the as "repetitive behaviors ... or mental acts ... the goal of frontal cortex and basal ganglia in OCD neuropathology, which is to prevent or reduce anxiety or distress, not to there was little empirical evidence to support such claims. provide pleasure or gratification." Unfortunately, these be­ Only through the use ofmodern neuroimaging techniques haviors are performed with excessive frequency, or with have many ofthese suspicions been confirmed. little direct connection to the original purpose ofthe action. Neurobiological studies ofOCD have produced a wealth OCD is differentiable from obsessive-compulsive per­ ofdata, but the inability to consistently replicate results sonality disorder (OCPD) by feelings of inappropriate­ has left many questions unanswered. Neuropsychological ness with respect to the obsessive and compulsive be­ studies have similarly generated much information, but haviors. Whereas OCD patients find these thoughts and with equal uncertainty. An adequate theory ofOCD must behaviors to be intrusive and inconsistent with their val­ integrate the findings from each ofthese two disciplines, ues and belief systems (ego-dystonic), OCPD patients but any such integration would be premature at the mo­ maintain that these thoughts and actions are consistent ment, given the variability present within each. The goal of with their self-image (ego-syntonic) (Hembree, Foa, & this review is therefore not to provide an all-encompassing Kozak, 1994). Although as many as 50% ofOCD patients theory ofOCD expression, but rather to delineate several meet the criteria for OCPD (Rasmussen & Tsuang, 1986), factors that may explain the discrepancies among previ­ the two remain quite distinct in etiology, maintenance, ous studies of OCD. Consideration ofthese factors will and treatment. help clarify the relevant neurobiological and neuropsycho- One ofthe most striking features ofthe disorder is the degree of regularity in obsessional thoughts and com­ pulsive behaviors. According to the DSM-IV, the most This research was funded by NINDS Grant ROI NS34030 (issued to common obsessions are fear ofcontamination, thoughts Martha 1. Farah) and an NSF STC grant to the Institute for Research in ofaggression, pathological doubt, excessive focus on bod­ Cognitive Science at the University of Pennsylvania. Correspondence concerning this manuscript should be addressed to K. D. Wilson, De­ ily functions, need for order, and sexual imagery. The man­ partment ofPsychology, University ofPennsylvania, 3815 Walnut Street, ual reports a similar regularity in compulsive behaviors, Philadelphia, PA 19104 (e-mail: [email protected]). the most common being washing and cleaning, counting,

161 Copyright 1998 Psychonomic Society, Inc. 162 WILSON checking, requesting or demanding assurances, repeti­ Severe cases ofOCD can be treated surgically. Several tion of previously performed actions, and ordering. Al­ procedures, including thermocapsulotomy, cingulotomy, though some patients experience a single thought obses­ subcaudate tractatomy, and limbic leukotomy (Jenike sionally, a majority experience more than one obsession et aI., 1991) have all been documented as effective in reliev­ or compulsion either simultaneously or over the course ing OCD symptoms, but each presents a host of poten­ ofillness (Rasmussen & Tsuang, 1986). tial complications. Although quite effective in most cases, Initial presentation of OCD typically occurs in ado­ the potential negative side effects justify their use only in lescence or in the early twenties (Black, 1974), with the the most intractable and debilitating cases. average onset occurring between ages 6 and 15 years for males and between ages 20 and 29 years for females. NEUROIMAGING OF OeD There is evidence for a genetic component, with concor­ dance rates considerably higher for monozygotic twins Recent advances in neuroimaging techniques have led than for dyzygotic twins (Stanley & Prather, "1993). Sim­ to a number ofsignificant findings in the OCD literature ilarly, first-degree relatives ofOCD patients are at higher (for a more complete review, see Trivedi, 1996). In several risk for psychopathology, with 25% being themselves di­ studies, researchers have looked for structural abnormal­ agnosed with OCD (Maxmen & Ward, 1995). ities that differentiate OCD patients and controls by using It is estimated that as many as 80% of OCD patients methods such as MRI and CT. In other studies, research­ suffer from depression (Rasmussen & Tsaung, 1986), ers have looked for differences in regional brain activity making it the most common co-occurring psychopathol­ during symptom provocation or in association with par­ ogy in OCD (Barlow, 1988). Tic disorders are also com­ ticular treatments, using methods such as PET, SPECT, mon: Tourette's syndrome is evident in 10%-15% of and fMRI. Although a number ofconflicting results will OCD patients (Pauls, Leckman, Towbin, Zahner, & Co­ be reviewed here, the majority ofstudies suggest neuro­ hen, 1986). Up to 13% offemale OCD patients are di­ pathology in three regions: the orbitofrontal cortex, the agnosed with eating disorders (including anorexia and basal ganglia (specifically, the head of the caudate nu­ bulimia nervousa) (Maxmen & Ward, 1995). Further­ cleus), and the thalamus. more, anxiety disorders (e.g., panic disorder), delusional disorders, hypochondriasis, and personality disorders Structural Brain Imaging Studies (e.g., histrionic) (Hembree et aI., 1994), as well as Hunt­ Insel, Donnelly, Lalakea, Alterman, and Murphy (1983), ington's disease, Sydenham's chorea, and postencephalitic using CT, conducted one of the earliest neuroimaging Parkinsonism (Cummings, 1993) have all been docu­ studies of OCD. These authors compared 10 OCD pa­ mented in cases of OCD. The overlap between schizo­ tients and 10 healthy controls on a number ofbrain mea­ phrenia and OCD appears to be relatively small, with less sures, including ventricular size and hemispheric asym­ than 3% of OCD patients being diagnosed as schizo­ metry, and found no significant differences between phrenic (Black, 1974). groups. Behar et al. (1984) conducted a similar CT study, Several treatment options are currently available for using 16 adolescent OCD patients, but reported ventric­ OCD. Rational-emotive (Warren & Zgourides, 1991), ular enlargement in the patient group relative to controls. psychoanalytic (Oppenheim & Rosenberger, 1991), and This finding was not replicated, however, by Luxenberg cognitive therapies (Salkovskis & Warwick, 1985) have et al. (1988), who found no significant difference in ven­ all been relatively unsuccessful. For example, Turner and tricular size between OCD and control subjects (n = 10 Michelson (1984) estimate that 20%-40% ofOCD pa­ in each group), although these authors did note a non­ tients show improvement with psychotherapy. Behav­ significant trend in that direction. Nevertheless, Luxen­ ioral interventions, including exposure therapy and desen­ berg et al. did find smaller caudate nuclei in the OCD sitization, have been more successful in clinical settings patient group relative to controls. Kellner et al. (1991), in (Williams & Chambless, 1994), with reviews estimating contrast, reported no difference between OCD patient and that 60%-80% ofOCD patients improve with behavioral control subjects on a number ofvolumetric measures, in­ interventions (Turner & Michelson, 1984). cluding caudate nucleus size. Pharmacological treatments of OCD have become In a more recent study, using MRI, Scarone et al. (1992) widely available in recent years. Selective serotonin re­ found increased caudate nucleus volume in the right uptake inhibitors (SSRIs) such as clomipramine (Ana­ hemisphere ofOCD patients (n = 20) relative to controls flaxine), fluoxetine (Prozac), and flouvoxamine (Floxy­ (n = 16), although the OCD patients in this study were fral) have been the most successful agents in alleviating all being treated with SSRIs at study. In contrast, Stein OCD symptoms (Perse, Greist, Jefferson, Rosenfeld, & et al. (1993), using CT, compared a group ofOCD (n = Dar, 1987). Response rates, however, are far from ideal. 16) patients and control (n = 8) subjects and found no cau­ For example, approximately 70% of OCD patients re­ date nucleus abnormalities. The authors did note, however, spond to clomipramine treatment, with typical patients that a subgroup ofOCD patients exhibiting high neuro­ seeing only a 40% reduction in symptoms (Turner & logical soft signs showed lateral ventricle enlargement. Michelson, 1984). Relapse occurs in approximately 85% Using MRI, Calabrese, Colombo, Bonfanti, Scotti, and ofpatients after termination oftreatment. Scarone (1993) found greater T 1 signal asymmetry in the OBSESSIVE-COMPULSIVE DISORDER 163

caudate nuclei ofOCD patients relative to controls. This patients. However, 10 ofthe 16 patients were undergoing contrasts with Garber, Ananth, Chiu, Griswold, and Old­ drug treatment at the time ofthe study. endorf (1989), who reported equivalent T I signal sym­ Using SPECT, Machlin et al. (1991) found increased metry in OCD patient and control subject groups. Robin­ metabolic activity in the medial frontal regions of 10 son et al. (1995) found reduced caudate volume in a group OCD patients, all ofwhom were medication free and had of OCD patients relative to controls (n = 26 in each no history of major depression. The authors noted that group), but did not exclude depressed or medicated OCD activity in the orbitofrontal regions was not significantly patients. Using MRI, Jenike et al. (1996) found a nonsig­ different between patient and control subjects. nificant trend toward reduced caudate volume in a group In a more recent study, Rubin, Villanueva-Meyer, ofOCD patients, as well as a significant reduction in total Ananth, Trajmar, and Mena (1992) employed SPECT to brain white matter, a significant increase in total cortex, compare groups ofOCD and control subjects (n = 10 in and a significant increase in opercular volume. Finally, each group) and found increased perfusion in the or­ Aylward et al. (1996) found no significant abnormalities bitofrontal cortex bilaterally, as well as decreased meta­ in the lateral ventricles and caudate nuclei of 24 OCD bolic activity bilaterally in the caudate nucleus. Finally, patients relative to controls (n = 21). Lucey et al. (1995), using SPECT, compared 30 OCD pa­ Although abnormalities have been reported in the lat­ tients with 30 healthy control subjects and found de­ eral ventricles and caudate nucleus, these results are not creased metabolic activity in the right caudate nucleus universal. A number offactors may be responsible for this and the right thalamus, as well as in several cortical areas, confusion, including (1) differences in neuroimaging including the right inferior frontal cortex. techniques, (2) differences in patient population charac­ Many of the criticisms leveled at structural neuro­ teristics, and (3) variable exclusion criteria, particularly imaging studies hold true for functional neuroimaging with respect to depression and medication. For example, studies ofOCD as well. The large number ofconflicting ofthe three studies in which caudate nucleus abnormal­ reports may be the result ofmarked differences in patient ities were reported, in one, the 20 OCD patients had all selection criteria, differences in patient characteristics received medication at the time of scanning (Scarone (e.g., disease onset, symptom subtypes, or gender), or et aI., 1992); in another, medicated and depressed OCD neuroimaging techniques. Nevertheless, abnormal base­ patients were included (Robinson et aI., 1995); and in line activity in the orbitofrontal cortex, caudate nucleus, one, the subjects were atypical childhood-onset OCD pa­ and thalamus are suggested in a number ofstudies. tients (Luxenberg et aI., 1988; see Alyward et aI., 1996, Symptom provocation. Provocation studies provide a for further comment). Attempts to increase methodolog­ more direct examination ofaberrant brain activity asso­ ical consistency across studies will certainly improve the ciated with obsessive-compulsive episodes. They also al­ replicability ofprevious results. low the OCD patient to serve as his or her own pre-episode control, although in many studies, groups ofhealthy con­ Functional Brain Imaging trol subjects have also been employed. Results consis­ Functional neuroimaging studies of OCD have con­ tently suggest the involvement ofthe orbitofrontal cortex centrated on three paradigms: resting-state brain metab­ and caudate nucleus. olism, symptom provocation, and behavioral or drug Zohar et al. (1989) scanned a group ofOCD subjects, treatment interventions. While structural neuroimaging using 133Xe inhalation under one of three conditions: studies have suggested caudal and ventricular abnormal­ rest, imaginal flooding, or contamination exposure. The ities, functional neuroimaging studies have focused on results showed that whereas imaginal flooding was asso­ dysfunctional brain circuits involving the caudate nucleus, ciated with increased cerebral blood flow (rCBF) in most the orbital frontal cortex, and the thalamus. cortical regions, actual contaminant exposure was asso­ Resting-state brain metabolism. In a number ofstud­ ciated with decreased CBF, particularly in the parietal and ies, researchers have looked at baseline differences in temporal regions. Using PET, Rauch et al. (1994) found metabolic activity between OeD patients and controls. that a symptom-provoked state produced significantly Baxter et al. (1988) used PET to compare 10 nonmed­ greater brain activation in the right caudate nucleus, in icated OCD patients (with no history ofdepression) with the left anterior cingulate gyrus, and bilaterally in the or­ 10 control subjects matched for age and gender. The au­ bitofrontal cortex. In addition, there was a nonsignifi­ thors reported increased overall metabolic activity in the cant trend toward greater activation in the thalamus as­ cerebral hemispheres, as well as increased activity in the sociated with the symptomatic state. orbital gyri and caudate nuclei. In two more recent stud­ McGuire et al. (1994) also used PET to examine dif­ ies, PET revealed increased activity in the orbitofrontal ferences associated with symptom-provoked obsessional regions, but no differential activity within the caudate states. The authors reported increased rCBF in the or­ nuclei (Nordahl et aI., 1989; Swedo, Schapiro, et aI., 1989). bitofrontal cortex and in several regions ofthe basal gan­ Using PET, Martinot et al. (1990) found decreased glia, including the caudate nucleus, the putamum, and metabolic activity within the prefrontal lateral cortex, as the globus pallidus. In addition, increased rCBF was re­ well as in the striatum and thalamus, in a group of16 OCD ported in the thalamus and the cingulate gyrus. Breiter 164 WILSON

et al. (1996) provide the only fMRI study to date ofrCBF examined the effects ofclomipramine on rCBF and found during symptom provocation. The authors report signif­ decreased metabolic activity in the orbitofrontal, postero­ icant activation in the orbitofronta1, anterior cingu1ate, frontal, and dorsal parietal regions following treatment. and insular cortex, as well as in the right caudate nucleus Treatment studies suggest that changes in the OC ex­ and left 1enticulate. perience that result from pharmacological or behavioral Symptom provocation studies provide the strongest intervention are accompanied by corresponding decreases and most consistent evidence of basal ganglia, orbito­ in regional CBF in certain brain regions, including the frontal, and thalamic involvement. Cumulatively, the re­ orbitofrontal cortex, caudate nucleus, and thalamus. These ports demonstrate abnormally elevated activity in the or­ results complement symptom-provocation studies, which bitofrontal cortex and caudate nucleus of OCD patients have consistently shown increased activity in these same in response to symptomatic episodes. In addition, abnor­ regions in response to obsessive or compulsive episodes. mal activity in the cingulate gyrus and thalamus has been Despite differences in neuroimaging techniques, patient reported in some studies. characteristics (e.g., gender, disease onset, and symptom Treatment studies. Neuroimaging studies have also subtypes), and exclusion criteria, neuroimaging studies been used to examine changes in brain activity associated ofOCD indicate a number ofbrain regions that seem to with behavioral or drug therapy in OCD. These reports, to­ be intimately involved in OCD expression, including the gether with symptom provocation studies, provide com­ orbitofrontal cortex, the basal ganglia and the thalamus. pelling evidence concerning the dynamic changes in brain activity associated with the ebb and flow ofOCD episodes. NEUROPSYCHOLOGY OF OCD The majority ofneuroimaging studies oftreatment inter­ vention involve drug therapy, with subjects being scanned In the last 15 years, efforts to characterize the cognitive prior and subsequent to the intervention. Except for the neuropsychology of OCD have been quite successful. studies done by Hoehn-Saric, Pearlson, Harris, Machlin, Although neuropsychological testing provides a less di­ and Camargo (1991) and Rubin, Ananth, Villanueva­ rect examination of the neuropathology of OCD than Meyer, Trajmar, and Mena (1995), who used SPECT, all does neuroimaging, the results provide complementary the studies presented below used PET methodology. information. Neuropsychological testing also allows re­ Baxter et al. (1987) treated 10 OCD patients with trazo­ searchers to document the range and severity ofcognitive done (an antidepressant) coupled with an MAO inhibitor dysfunction in OCD patients. Indeed, a broader under­ and found increased activity in the right caudate nucleus standing ofthe cognitive manifestations ofthe disorder relative to a pretreatment baseline. In a more recent report, (i.e., beyond the cognitive deficits most evident during Baxter et al. (1992) examined the effects ofdrug treatment symptom expression) will be critical for establishing a (fluoxetine) and behavioral therapy on rCBF and found cognitive neuroscience ofOCD. that activity in the right caudate nucleus decreased sig­ Neuropsychological testing is currently aimed at multi­ nificantly in subjects who responded to treatment, re­ ple levels ofcognitive function, ranging from more "basic" gardless ofform (i.e., pharmacological or behavioral). In sensorimotor programming to the highest level "execu­ addition, activity was significantly correlated in the or­ tive" functions. Whereas many ofthe preliminary inves­ bital frontal cortex, caudate nucleus, and thalamus be­ tigations were inspired by the paucity of results in this fore, but not after, successful treatment. Schwartz, Stoes­ area, more recent efforts have been motivated by specific sel, Baxter, Martin, and Phelps (1996) also examined the predictions based on neuroanatomical considerations. effects of behavioral modification on brain metabolism This has resulted in a complement of studies that not and found decreased metabolic rates in the caudate nucleus only catalogue the neuropsychological profile of OCD, bilaterally in OCD responders. The authors also found but that provide critical empirical tests for current mod­ that correlated activity in the caudate nucleus, orbital gyri, els ofOCD. and thalamus decreased significantly after treatment. Other groups have looked at the effects of SSRIs on Sensorimotor and Attentional Mechanisms brain activity in OCD patients. Benkelfat, Nordahl, Sem­ Cognitive deficits in OCD occur quite early in both the ple, and King (1990) found decreased activity in the left sensory and motor pathways. Deficits in sensory input caudate nucleus and right orbitofrontal cortex following gating (Swerdlow, Benbow, Zisook, Geyer, & Braff, a 16-week trial of clomipramine. Hoen-Saric, Pearlson, 1993) as well as in saccadic eye movement programming et al. (1991) found decreased activity in the medial-frontal (Sweeney, Palumbo, Halper, & Shear, 1992; Tien, Pearl­ cortex following a 12- to 16-week trial of fluoxetine. son, Machlin, Bylsma, & Hoehn-Saric, 1992) have both Swedo et al. (1992) found decreased activity in the or­ been established in OCD. Several reports have also bitofrontal cortices (both left and right), but not caudate demonstrated abnormal attentional mechanisms in OCD nucleus, after 1 year of clomipramine treatment. Perani patients (e.g., Towey et aI., 1994). Although it is true that et al. (1995) treated 11 patients with fluvoxamine, fluoxe­ saccadic programming, sensorimotor gating, and atten­ tine, or clomipramine and found decreased activity in the tion have all been linked to orbitofrontal cortex and basal cingulate cortex after 12 weeks. Rubin et al. (1995) also ganglia function (and thus provide confirmation ofneuro- OBSESSIVE-COMPULSIVE DISORDER 165

imaging results), interpretation ofthese results with re­ sured by a reduction in PPI, but they were careful to note spect to OCD symptomology is not at all clear. Oculo­ that the relationship between their results and the inabil­ motor programming deficits provide an excellent example. ity ofOCD patients to inhibit obsessional thoughts is un­ Proper oculomotor (e.g., saccadic) programming re­ clear. That is, whether a general deficit in sensorimotor lies on normal basal ganglia function. Diseases of the gating produces, or simply accompanies, the inability to basal ganglia, such as Huntington's disease, have been suppress the particular obsessions experienced by OCD shown to disrupt oculomotor programming (Cummings, patients remains uncertain. Their findings nevertheless 1993). Given the ability of Huntington's disease to in­ appear relevant to the cognitive and behavioral phenom­ duce OCD, and given the establishment (through neuro­ enology ofOCD. imaging) of basal ganglia dysfunction in OCD (e.g., Swerdlow et al.'s (1993) findings reflect, perhaps, an Baxter, Schwartz, Guze, Bergman, & Szuba, 1990), sim­ early attentional disturbance in OCD patients. Nelson, ilar oculomotor deficits might be expected in OCD pa­ Early, and Haller (1993) conducted a more direct exam­ tients. Tien et al. (1992) and Sweeney et al. (1992) each ination ofthe attentional processes ofOCD patients using describe a series ofexperiments designed to test this hy­ a Posnerian attention task (Posner, Nissen, & Ogden, pothesis. Tien et al. reported that OCD subjects showed 1978). In this task, a central fixation cross is flanked by greater error rates and greater numbers of inaccurate two target locations. Subjects are required to respond to saccades in a goal-directed antisaccade task. There was a target in either location as quickly as possible. On each no difference, however, in overall reaction time, saccadic trial, target presentation is preceded by a location cue velocity, or accuracy ofsaccades in a visually guided sac­ that is valid, invalid, or noninformative (neutral). Typi­ cade task. The authors conclude that this performance is cally, responses are faster in the valid cue condition than consistent with disruption ofbasal ganglia function, but in the neutral condition. Similarly, responses are faster in they offer no connection between oculomotor program­ the neutral condition than in the invalid condition. ming deficits and OCD symptomology. Although the performance of OCD patients was in Sweeney et al. (1992) reported similar results, with most respect equivalent to that ofcontrol subjects, there OCD patients having difficulty in smooth tracking of was a marked difference in their inhibition ofreturn. In­ targets, an increased rate of"reversal saccades" to com­ hibition ofreturn refers to the finding that valid precu­ pensate for target overshooting, and an increased rate of ing normally produces benefits in reaction time, long in­ "square-wave" (intrusive) saccades while attempting to tervals between valid cue and target presentation actually maintain fixation. The authors concluded that the results produce a cost. This inhibition of return is a common were consistent with dysfunction in the basal ganglia and finding in the Posnerian attentional task, and it reflects, frontal regions, but no connection was made between perhaps, an inherent preference for novelty. In Nelson these results and OCD symptom expression. et al.'s (1993) study, unlike control subjects who exhib­ The preceding examples are typical ofneuropsycho­ ited normal inhibition ofreturn, OCD patients showed a logical testing in OCD: Strides in documenting neu­ decreased inhibition ofreturn for targets presented in the roanatomical, cognitive, and behavioral deficits have not left visual field, and no inhibition of return for targets necessarily shed light on the mechanisms that produce presented in the right visual field. the disorder. To be sure, not all deficits witnessed in OCD Nelson et al. (1993) interpret their results as a disrup­ are fundamentally involved in the expression ofthe dis­ tion ofthe "novelty bias" in OCD patients, which could order. Indeed, some deficits co-occur because ofshared reflect a general impairment in the ability to suppress at­ neural substrates. This is perhaps the greatest challenge tended stimuli, such as obsessional thoughts. They con­ in constructing an adequate account ofthe mechanisms clude that this deficit may reflect a "breakdown of the ofOCD: knowing which deficits are intimately related to gating mechanism allowing normally ignored, exoge­ OCD expression and which are merely concomitant. nous stimuli to occupy endogenous resources" (p. 194). More relevant to OCD symptomology, Swerdlow et al. Clearly, this finding suggests a possible mechanism by (1993) hypothesized that the inability ofOCD patients to which obsessional thoughts fail to be suppressed in OCD gate inappropriate thoughts (obsessions) may reflect a patients, but again, any such mechanism is strictly spec­ more central sensorimotor gating deficit. Their depen­ ulative. Furthermore, this account does not address the dent measure was the prepulse inhibition (PPI) ofa star­ specificity ofobsessional thoughts-a qualification which tle reflex. Normally, a startle response can be suppressed will resurface repeatedly in this examination. That is, a when the startling stimulus is preceded by a weak "pre­ general impairment in the ability to suppress "normally pulse." Deficient sensorimotor gating, however, reduces ignored, exogenous stimuli" does not explain the marked the PPI. Striatal dysfunction could produce such a defi­ specificity ofobsessional thoughts, both within (i.e., pa­ cit, since this area is known to be important in the nor­ tients normally experience only one thought obsession­ mal suppression of the startle reflex, and interestingly, ally) and across (i.e., the range of obsessional thoughts the striatal regions have been implicated in OCD pathol­ is remarkably narrow) patients. ogy. Thus, the underlying neuroanatomy of OCD may The examination ofsensorimotor and basic attentional produce a general deficit in gating sensorimotor stimuli. mechanisms in OCD patients has been helpful in con­ Swerdlow et al. (1993) reported that OCD patients did firming the involvement ofcertain brain regions but has in fact show a deficit in sensorimotor gating, as mea- contributed relatively little to our knowledge of the na- 166 WILSON ture ofOCD symptoms. Analysis ofhigher cognitive func­ right-hemisphere asymmetry but also highlighted the tions may prove more helpful in establishing such a link. role ofthe frontal cortex. Nevertheless, any indication ofhigher level cognitive dys­ In an attempt to address the discrepancies between pre­ function must be interpreted with respect to hyperacti­ vious findings, Head et ai. (1989) administered several vation ofattentional mechanisms. tests aimed at assessing the visuospatial and set-shifting abilities of 15 OCD patients. The authors used the Mod­ Traditional Neuropsychological Testing ified Wisconsin Card Sorting Task (MWCST) to assess Two results appear consistently across studies ofhigher set-shifting ability, the Block Design subtest ofthe WAIS level cognitive functions in OCD patients-namely, im­ (WAIS-BD) to assess spatial abilities, and the MRMT to pairments in visuospatial and cognitive set-shifting abil­ assess both set-shifting and visuospatial abilities within ities. The relationship between these findings and possible the same task. Reliable differences on the MWCST, the deficits in more basic attentional mechanisms remains WAIS-BD, and the MRMT were found, leading the au­ unclear. It should be noted that for many ofthe measures thors to propose that OCD patients had difficulties in discussed below, both positive and negative results have nonverbal tasks that required set-shifting. Results ofmore been reported (e.g., OCD patient performance on the Wis­ general tests ofset-shifting and visuospatial skills were consin Card Sorting Task [WCST] has been reported as equivocal, leading the authors to withhold any conclusions impaired [Head, Bolton, & Hymas, 1989] and as within about more general impairments in either domain. normal ranges [Abbruzzese, Bellodi, Ferri, & Scarone, The results of early efforts to capture the neuropsy­ 1995], depending on the study). Similar disagreement chological profile ofOCD patients were inconsistent in about laterality of impairment also exists (e.g., primar­ many cases. Researchers in more recent studies have at­ ily left hemisphere [Flor-Henry, Yeudall, Koles, & Ho­ tempted to address these inconsistencies and in some warth, 1979] involvement in some reports, and primarily cases have provided plausible explanations for previous right hemisphere [Insel et aI., 1983] involvement in oth­ discrepancies. For example, Martinot et ai. (1990) noted ers). Nevertheless, several common themes have emerged. inconsistencies with respect to laterality of impairment Flor-Henry et ai. (1979) offered perhaps the first ex­ and frontal lobe functioning in OCD patients. They argued tensive examination of neuropsychological test perfor­ that these discrepancies could be due to poor experimen­ mance in OCD. The research was motivated largely by tal control ofseveral factors: (1) the use ofdepressed ver­ the lack of clear documentation of cognitive deficits in sus nondepressed OCD patients, (2) medicated versus OCD patients. The authors compared the performance medication-free patients, and (3) control subjects versus ofOCD patients with age- and education-matched con­ normative data. They therefore compared the perfor­ trols, on tests of spatial abilities, frontal lobe function­ mance ofnondepressed OCD patients (using medicated ing, verbal fluency, and a range ofsimilar tasks. Although and medication-free subgroups) with age-matched con­ they found no significant difference on tasks such as Ver­ trol subjects on a variety ofneuropsychological tests. bal Fluency and the Vocabulary and Block Design sub­ Five general areas ofcognitive function-namely, lan­ tests of the WAIS, they did note differences on others guage, mental control, global memory, attention, and tasks such as Raven's Coloured Progressive Matrices, praxis-were tested (Martinot et aI., 1990). Language abil­ tactual performance tasks, the Purdue Pegboard task, and ities, assessed by tests ofverbal fluency, appeared normal the Digit Span and Digit Symbol subtests ofthe WAIS. as did performance on several tests of attentional abili­ The authors argued that this pattern ofdifficulties most ties, including the Trail Making and Graphic Alternating likely reflected frontal dysfunction, primarily in the left Sequence tasks. The authors did, however, find impair­ hemisphere. ment in the Stroop task as well as several memory tasks, Several of Flor-Henry et aI.'s (1979) initial findings including the Rey Figure and Rey Verbal Learning tasks. were not consistently replicated. Insel et ai. (1983), for The authors concluded that with experimental control of instance, attempted to replicate the results with a group the considerations mentioned above, OCD patients showed of 18 OCD patients. Results from only 1 of 11 neuro­ a selective attentional deficit. psychological tests (a nonvisual spatial organization test) Boone, Ananth, Philpott, Kaur, and Djenderedjian were replicated, although it must be noted that Insel et ai. (1991) conducted a similar study addressing the follow­ used normative data, not control subjects, in their repli­ ing cognitive areas: visual spatial skills, memory and at­ cation study. The authors argued that their findings were tention, frontal lobe functioning, and intelligence. They more accurately represented as a deficit in spatial per­ found no significant differences between 20 nondepressed ception, and therefore more consistent with right, rather OCD patients and control subjects on tests ofintelligence, than left, hemisphere damage. attention, verbal memory, or frontal lobe tasks. Interest­ Behar et ai. (1984) conducted a similar study and found ingly, the frontal lobe tasks included the WCST. They did, that OCD patients were impaired on several visuospatial however, note significant deficits on several nonverbal tasks, including the Stylus Maze task, the Money Road visual memory tasks, such as the Hooper Visual Organi­ Map Test (MRMT), and the Rey-Osterrieth complex fig­ zation and Rey-Osterrieth Complex Figure tasks. The ure task. The authors suggested a visuospatial deficit and authors concluded that these impairments were consis- OBSESSIVE-COMPULSIVE DISORDER 167

tent with basal ganglia dysfunction, and possibly right Brown, Kosslyn, Breiter, Baer, and Jenike (1994) ex­ hemisphere laterality. amined the performance ofOCD patients on several tasks The apparent disagreement concerning frontal lobe under the assumptions of signal detection theory. They functioning led Hoehn-Saric, Harris, et al. (1991) to exam­ tested a group of OCD patients on a task that required ine the possibly causal role ofpharmacological treatment discrimination ofreal and imagined words, with the hy­ in the development of frontal syndrome symptoms in pothesis that OCD patients are less capable ofdiscrimi­ OCD. They, like Martinot et al. (1990), were concerned nating between real and imagined events. That is, they with poor experimental control in previous reports, spe­ hypothesized that OCD patients are more greatly im­ cifically in the use of medicated patients. The authors paired in discriminating between actually performing an hypothesized that fluoxetine, a commonly prescribed drug action as opposed to only imaging it, and that this in­ in the treatment ofOCD, produced symptoms that were ability underlies OCD symptom expression. Contrary to characteristic of frontal lobe syndrome. Thus, deficient their predictions, however, OCD patients showed greater performance by OCD patients in frontally mediated neuro­ ability to discriminate between real and imaged events psychological tests may have resulted from the treatment in a signal detection task, as measured both by subjects' ofthese patients, rather than from the disorder itself. sensitivity to stimuli (d') and by criteria for decision ({3). Hoen-Saric, Harris, et al. (1991) administered a series Abbruzzese et al. (1995) conducted a more recent in­ of neuropsychological tests to an OCD patient before vestigation of cognitive neuropsychological test perfor­ and after treatment with fluoxetine. The subject scored in mance in OCD patients. The authors hypothesized that if the normal range on all tests prior to medication, includ­ orbitofrontal cortex (OFC), as opposed to dorsolateral ing Verbal and Design Fluency, Complex Figure Copy prefrontal cortex (DLPFC) was implicated in OCD, pa­ and Recall, Verbal Learning, Stroop, and Trail Making tients should show deficits on OFC-dependent tasks tasks. Performance on tests ofsustained attention and in­ (such as the object alternation tasks, or OAT), but not on hibition of prepotent responses (the Stroop and Trail DLPFC-dependent tasks (such as the WCST). Consistent Making tasks) were unaffected during treatment. Perfor­ with their hypothesis, Abbruzzese et at. found that 25 mance on tasks that involved the generation ofnovel re­ OCD patients were impaired on the OAT but not the sponses, set-shifting, and memory were all impaired dur­ WCST. As a control, they also found that 25 schizophrenic ing the administration offluoxetine. The authors argued patients, who are presumed to have only DLPFC dam­ that symptoms offrontal lobe syndrome were not intrin­ age, were impaired on the WCST but not the OAT. sic to the disorder, but rather were the result ofpharma­ Taken together, neuropsychological testing has sug­ cological intervention. The authors argue in conclusion gested several problems in OCD: hyperactivation ofat­ that this result is capable ofexplaining several inconsis­ tentional systems, difficulties with cognitive set-shifting, tencies in previous research. and visuospatial deficits. In contrast, general intelligence, Another group (Christensen, Kim, Dysken, & Hoover, verbal skills, and executive functions (with the exception 1992) examined the neuropsychological functions of 18 ofset-shifting) all remain preserved in OCD. OCD patients specifically in the areas ofverbal abilities, visual spatial abilities, recent verbal memory, recent non­ MODELS OF OeD verbal memory, and executive function. The authors found deficits in tests ofrecent nonverbal memory, as well as Models ofOCD expression have focused on a number visuospatial reasoning. They did not, however, find def­ ofissues, including the production and execution ofgoal­ icits in tasks involving executive functions, verbal abili­ directed motor programs, malfunction of behavioral in­ ties, or recent verbal memory. Although they are careful hibition systems, and the release ofspecies-specific adap­ to note that speeded performance had a profound impact tive behaviors. Although many ofthese models account on a variety of tasks, the most obvious deficiency was for the general characteristics ofOCD, they often lack the within the domain of recent nonverbal memory. They requisite specificity to explain the very constrained na­ argue further that this pattern ofimpairment is consistent ture ofOCD symptomology. with damage to mesial temporal areas ofthe right hemi­ Pitman (1987) draws on cybernetic research to develop sphere, possibly in the limbic and paralimbic regions. the notion of a malfunctioning homeostatic system in In many of the studies presented above, the perfor­ OCD, using the analogy ofathermostat. Accordingly, per­ mance of OCD patients was assessed on a wide variety ceptual systems are compared with internal reference ofneuropsychologiGal tasks without a specific model of signals, and an error signal denotes deviation ofthe for­ etiology. In more recent examinations, driven by specific mer from the latter. The system, experiencing this error theories ofOCD etiology, researchers have attempted to signal, recruits the behavioral output system in an at­ ask more focused questions about performance on spe­ tempt to reduce the discrepancy between the actual and cific tasks. These attempts have also focused more di­ desired states. rectly on the relationship between symptom expression OCD is thought to result from persistent, high error and neuropsychological test performance, with the for­ signals in this model (Pitman, 1987). These error signals mer often generating predictions for the latter. Two ex­ arise from several possible malfunctions: Internal con­ amples are presented below. flict between two control systems, intrinsic comparator 168 WILSON

malfunction, and attentional disturbances could each in the treatment ofsuch behavioral problems in lower an­ produce unwarranted error signals. According to the the­ imals, suggesting some consistency across species. ory, these signals are manifested as obsessional thoughts While Rapoport (1991) more directly addresses the which drive the behavioral output system into effecting specificity and family resemblance ofobsessional thoughts compulsive behaviors. Despite the compulsive behav­ and compulsive behaviors, the theory lacks evidence for iors, however, the high error signal is never sufficiently the described temporal relationship in humans. Further­ reduced. more, it cannot account for OCD cases in which the com­ Pitman (1987) notes that the basal ganglia have been pulsive behaviors are not externally observable, evolution­ implicated in the production of behavioral (motor) out­ arily adaptive behaviors, such as prayer, counting, and put, and that the limbic system can be viewed as an "in­ checking in some cases (e.g., a patient who obsesses about ternal comparator" within the goal-directed behavioral hitting a pedestrian while driving, and who retraces his system, thereby providing a neuroanatomical basis to his route while driving to check for such occurrences). theory. He does not, however, provide an explicit role for Researchers are beginning to consider the importance the orbitofrontal cortex, nor does he address the speci­ of information processing and cognitive dysfunction in ficity with which obsessive thoughts appear. the origin ofOCD. Stein and Hollander (1992,1994), for Another model of OCD symptom expression is pre­ example, have argued for the importance of adopting a sented by Baxter et al. (1990), who argue for the role of cognitive science approach to OCD, combining results the striatum in the production of obsessive symptoms. from a variety ofdisciplines in explaining how the cog­ The claim is that "striatal dysfunction may lead to ad­ nitive as well as the behavioral symptoms ofOCD arise. ventitious thoughts and sensations and actions that are They argue that connectionist simulations ofOCD may normally suppressed with little conscious effort by site­ provide crucial insights into the role ofinformation pro­ specific regions of the striatum" (p. 24). Hyperactivity cessing in the emergence of OCD. They also note the within the orbitofrontal regions, it is argued, reflects the success ofconnectionist modeling in other areas ofpsy­ recruitment ofcortical mechanisms in an (unsuccessful) chopathology research (e.g., Cohen & Servan-Schreiber's attempt to inhibit these thoughts. The authors highlight [1992] connectionist model relating the biological and the similarity between OCD patients, who are unable to cognitive deficits in schizophrenia) and appeal for simi­ prevent intrusion ofobsessional thoughts, and Hunting­ lar models in OCD research. ton's patients, who are unable to prevent intrusion ofun­ wanted motor programs. This similarity may reflect an DISCUSSION overlap in basal ganglia dysfunction in both groups. Although concerned with the same neuroanatomical Researchers have made tremendous progress in un­ circuits as Baxter et al. (1990), Rapoport (1991) offers a derstanding OCD over the past decade, particularly with very different model ofOCD symptomology. On the basis respect to the neuroanatomy and neurophysiology ofthe ofanimal research addressing the effects ofcaudate nu­ disorder. Nevertheless, a major stumbling block is the in­ cleus lesions, Rapoport argues that compulsive behav­ ability to consistently replicate findings, in both the neu­ iors are experienced before obsessional thoughts in OCD, roimaging and the neuropsychological testing literature. and that the latter are undertaken as justifications for the The failure to consistently replicate findings in OCD inappropriate nature of the former. The theory assumes research may result from a number ofmethodological dif­ that damage to the caudate nucleus releases an excess of ferences between studies, including differences in the certain species-specific, fixed-action patterns that have criteria for excluding medicated or depressed OCD pa­ evolutionary significance. An example is grooming be­ tients. Differential results may also reflect differences in havior, which in many lower species (as well as our own) the underlying neuropathology ofOCD subgroups. Spec­ is crucial to survival. Such "species-typical self-protective ifically, it is possible that distinct OCD patient subgroups behaviors" (p. 7) are normally regulated by the circuits manifest different impairments within the circuit con­ connecting regions of the cortex, thalamus, and basal necting the orbitofrontal cortex, the basal ganglia, and thal­ ganglia. Dysfunction within this system (particularly in amus. Although aberrant brain activity during OCD epi­ the caudate nucleus) causes excessive release ofsuch be­ sodes likely occurs throughout this circuit, the sources haviors, which are characterized as compulsive behav­ ofthe dysfunction might be different for different patients. iors. According to the theory, obsessional thoughts arise These differences would show up most clearly in struc­ as a justification for the inappropriate behaviors in which tural neuroimaging or baseline functional neuroimaging the OCD patient engages. studies, when the circuit is not actively engaged. Rapoport (1991) offers several examples ofanalogous In recent theories, researchers have attempted to di­ behavioral phenomena in lower animals. For example, vide OCD patients into obsessive and compulsive sub­ acral lick dermatitis (in dogs) and excessive preening (in groups (Baer, 1994; Hantouche & Lancrenon, 1996). On birds) are examples of excessive and ritualistic behav­ the basis of symptomology and OCD inventory scores, iors performed in a manner similar to the rituals ofOCD these studies have tended to place patients into one of patients. Interestingly, Rapoport demonstrates the effec­ three groups: purely obsessional (i.e., patients with ex­ tiveness ofpharmacological agents, specifically SSRIs, cessive sexual, religious, or violent imagery), purely com- OBSESSIVE-COMPULSIVE DISORDER 169

pulsive (i.e., patients with counting or checking rituals as imaging studies report the specific symptoms experi­ well as patients with symmetry or order obsessions), and enced by subjects (with the notable exception of symp­ mixed (i.e., patients with contamination fears or exces­ tom provocation studies). sive washing or grooming). It is possible that these sub­ Potential differences in the underlying neuropathology groups exhibit different neuropathologies within the or­ among OCD subtypes is one ofa number of potential bitofrontal-basal-thalamic-cortical circuit, and that these confounds in current OeD research. The effects ofmed­ differences underlie some ofthe inconsistencies within ication, the inclusion ofdepressed OCD patients, gender the OCD literature. differences, and differences between childhood- and adult­ Lesion studies have demonstrated that damage to the onset OCD also must be considered in interpreting the orbitofrontal cortex produces disinhibition, inappropri­ results ofprevious investigations. Until these factors are ate behavior, or emotional displays; inappropriate re­ systematically explored in an empirical manner, any the­ sponses to environmental cues; and utilization behavior ory of OCD neuropathology and symptom expression (Cummings, Gosenfeld, Houlihan, & McCaffrey, 1983; will be premature. Kolb, 1977; Lhermitte, Pillon, & Serdaru, 1986; Logue, Our review ofthe neuroanatomy and neuropsychology Durward, Pratt, Piercy, & Nixon, 1968; Rolls, 1985). ofOCD suggests several directions for further research. OCD patients with purely obsessional or purely compul­ One priority is a more systematic exploration ofseveral sive symptoms display such characteristics. The inabil­ patient selection criteria, including the use ofmedication, ity to inhibit thought patterns, particularly deviant or in­ the exclusion ofdepressed OCD patients, the age ofdis­ appropriate ones, is characteristic of many of these ease onset, the patient's gender, and the symptoms expe­ patients. In addition, checking, counting, and hoarding rienced by each subject. Greater uniformity in selection rituals can all be seen as a type ofutilization behavior or criteria across study will no doubt improve the replica­ as an inappropriate response to environmental cues. These bility ofprevious results. Specifically, attention to these two OCD patient subgroups therefore share much in com­ details will confirm the precise role ofthe orbitofrontal mon with "orbitofrontal" patients. cortex, the basal ganglia, and the thalamus in OCD Neurological studies have demonstrated the link be­ expression. tween damage to the basal ganglia and compulsive behav­ Another important issue in OCD research is that of iors in a number ofexcessive motor disorders, including understanding the role of set-shifting impairments and Sydenham's chorea (Swedo, Rapoport, et aI., 1989), Hun­ hyperattention in the emergence ofOCD. That is, are the tington's disease (Cummings & Cunningham, 1992), and cognitive impairments causally related to OCD, or vice Tourette's syndrome (Insel, 1992). Rapoport has simi­ versa, or do both emerge simultaneously? It seems likely larly reported that the basal ganglia may be involved in that the two reflect a common underlying mechanism, disorders ofspecies-specific grooming or washing behav­ but this mechanism has not been studied experimentally. iors such as excessive hair pulling (trichotillomania) and Furthermore, serious attempts to relate visuospatial def­ excessive nail biting in humans, and acral dermatitis or icits to OCD phenomenology are needed. Visuospatial excessive preening in animals (Rapoport, 1991). OCD pa­ deficits have been considered only in light ofwhat they tients with fears ofcontamination or washing and groom­ may tell us about possible underlying neuroanatomy. A ing compulsions (i.e., the "mixed" variety) can similarly more complete analysis, however, with examination of be characterized as having dysfunctional grooming or the relationship between this deficit and OCD sympto­ washing behaviors and therefore share much in common mology seems imperative. with these "basal ganglia" patients. One ofthe most difficult features ofOCD, and one for Although OCD episodes most likely result from aber­ which little has been said, is the rather restricted nature rant activity within the entire frontal-basal-thalamic loop, of obsessional thoughts. Although it is possible to con­ the origin of abnormal activity may be different in dif­ ceive ofobsessions taking many forms, remarkably few ferent cases. OCD symptoms may be triggered by dysfunc­ types have been reported. An adequate theory must ac­ tional orbitofrontal activity in some cases and by dys­ count for this restricted class ofpotential obsessions and functional basal ganglia activity in others. Thus, brain characterize the family resemblance they share. Few the­ imaging or neuropsychological testing may find differ­ ories have attempted to describe these characteristics, but ent patterns, depending on the source ofdysfunction. If cognitive neuroscience may uncover valuable clues about the patient group draws more heavily from one group the nature and basis of obsessional thoughts in particu­ than from another, we might expect to find differences in lar, and OCD in general. the observed neuropathology. That is, resting-state meta­ bolic studies and structural neuroimaging studies may CONCLUSIONS find a greater incidence oforbital frontal deficits in OCD patients who experience purely obsessional or purely com­ The question still remains: By what brain mechanisms pulsive symptoms and a greater incidence ofbasal gan­ do the particular symptoms experienced by the OCD pa­ glia dysfunction in OCD patients with mixed sympto­ tient arise? Although research on OCD has flourished at mology (i.e., those with contamination fears or washing many levels, a successful integration ofthis information and grooming rituals). Unfortunately, relatively few neuro- is still lacking. Before any such integration can be fruit- 170 WILSON

ful, however, several aspects of OCD research method­ compulsive disorder. Archives o/General Psychiatry, 49, 681-689. ology need to be addressed systematically. In this review, BAXTER, L. R., SCHWARTZ, J. M., GUZE, B. H., BERGMAN, K. S., & it is suggested that medication, depression, gender, dis­ SZUBA, M. P. (1990). Neuroimaging in obsessive-compulsive disor­ der: Seeking the mediating neuroanatomy. In M. Jenike, L. Baer, & ease onset, and symptom subtype may all conspire to W. Minichiello (Eds.), Obsessive-compulsive disorders: Theory & produce differential results, particularly in the neuro­ management (pp. 167-188). St. Louis: Mosby. imaging literature. Systematic exploration ofthese fac­ BAXTER, L. R., SCHWARTZ, J. M., MAZZIOTTA, J. C., PHELPS, M. E., tors will reveal the true neurobiological factors relevant PAHL, J. J., GUZE, B. H., & FAIRBANKS, L. (1988). Cerebral glucose to OCD expression, and only then will an adequate the­ metabolic rates in nondepressed patients with obsessive-compulsive disorder. American Journal 0/Psychiatry, 145, 1560-1563. ory ofOCD emerge. BAXTER, L. R., THOMPSON, J. M., SCHWARTZ, J. M., GUZE, B. H., Greater concentration on the cognitive aspects ofOCD, PHELPS, M. E., MAZZIOTTA, J. C., SELIN, C. E., & Moss, L. (1987). and specifically on the information-processing charac­ Trazodone treatment response in obsessive-compulsive disorder­ teristics, is also necessary. Although many ofthe currently correlated with shifts in glucose metabolism in the caudate nuclei. Psychopathology, 20(Suppl. I), 114-122. popular models take aim at the production ofcompulsive BEHAR, D., RAPOPORT, J. L., BERG, C. J., DENCKLA, M. B., MANN, L., behaviors, a more focused examination ofthe nature and COX, C., FEDIO, P., ZAHN, T., & WOLFMAN, M. G. (1984). Computer­ inception ofobsessional thoughts is equally important. It ized tomography and neuropsychological test measures in adoles­ seems imperative, therefore, that greater emphasis be cents with obsessive-compulsive disorder. American Journal 0/Psy­ chiatry, 141, 363-369. placed on understanding the nature ofobsessional thoughts BENKELFAT, C., NORDAHL, T. E., SEMPLE, W. E., & KING, A. (1990). and on their emergence within the framework ofgeneral Local cerebral glucose metabolic rates in obsessive-compulsive dis­ cognition. order: Patients treated with clomipramine. Archives o/General Psy­ Substantial evidence has been gathered concerning chiatry, 47, 840-848. the biological and cognitive foundations ofOCD, but our BLACK, A. (1974). The natural history ofobsessional neurosis. In H. R. Beech (Ed.), Obsessional states (pp. 19-54). London: Methuen. task is far from complete. Functional neuroimaging has BOONE, K. B., ANANTH, J., PHILPOTT, L., KAUR, A., & DJENDERED­ implicated the orbitofrontal cortex, basal ganglia, and HAN, A. (1991). Neuropsychological characteristics ofnondepressed thalamus in the expression ofOCD. Neuropsychological adults with obsessive-compulsive disorder. Neuropsychiatry, Neuro­ studies have highlighted impairments in cognitive set­ psychology, & Behavioral Neurology, 4, 96-109. BREITER, H. C., RAUCH, S. L., KWONG, K. K., BAKER, J. R., WEISSKOFF, shifting, hyperactivation of attentional systems, and R. M., KENNEDY, D. N., KENDRICK, A. D., DAVIS, T. L., JIANG, A., visuospatial deficits. These findings offer guidance for COHEN, M. S., STERN, C. E., BELLIVEAU, J. W., BAER, L., O'SULLI­ constructing a cognitive neuroscience of OCD, but fur­ VAN, R. L., SAVAGE, C. R., JENlKE, M. A., & ROSEN, B. R. (1996). ther work will be necessary before any such theory is Functional magnetic resonance imaging ofsymptom provocation in possible. obsessive-compulsive disorder. Archives o/General Psychiatry, 53, 595-606. Our knowledge ofthe nature ofOCD is growing rap­ BROWN, H. D., KOSSLYN, S. M., BREITER, H. c., BAER, L., & JENIKE, idly. Neurobiological studies of OCD, although rela­ M. A. (1994). Can patients with obsessive-compulsive disorder dis­ tively recent, have made great strides. To be sure, only criminate between percepts and mental images? A signal detection through further such empirical investigation will an ad­ analysis. Journal 0/Abnormal Psychology, 103,445-454. equate account ofOCD emerge. Itcan be concluded, how­ CALABRESE, G., COLOMBO, c., BONFANTI, A., SCOTTI, G., & SCARONE, S. (1993). Caudate nucleus abnormalities in obsessive-compulsive dis­ ever, that a greater emphasis on methodological consis­ order: Measurements of MRI signal intensity. Psychiatry Research: tency and an integration of results from neuroimaging Neuroimaging, 50, 89-92. and neuropsychological testing will be critical before CHRISTENSEN, K. J., KIM, S. W., DYSKEN, M. W., & HOOVER, K. M. any such account can be possible. (1992). Neuropsychological performance in obsessive-compulsive disorder. Biological Psychiatry, 31, 4-18. COHEN, J. D., & SERVAN-SCHREIBER, D. (1992). Context, cortex, and REFERENCES dopamine: A connectionist approach to behavior and biology in schizophrenia. Psychological Review, 99, 45-77. ABBRUZZESE, M., BELLODI, L., FERRI, S., & SCARONE, S. (1995). Mem­ CUMMINGS, J. L. (1993). Frontal-subcortical circuits and human behav­ ory functioning in obsessive-compulsive disorder. Behavioral Neu­ ior. Archives o/Neurology, 50, 873-880. rology,6,119-122. CUMMINGS, J. L., & CUNNINGHAM, K. (1992). Obsessive-compulsive AMERICAN PsYCHOLOGICAL ASSOCIATION (1994). Diagnostic and sta­ disorder in Huntington's disease. Biological Psychiatry, 31, 263-270. tistical manualo/mentaldisorders (4th ed.). Washington, DC: Author. CUMMINGS, J. L., GOSENFELD, L. E, HOULIHAN, J. P., & MCCAFFREY, T. AYLWARD, E. H., HARRIS, G. 1., HOEHN-SARIC, R., BARTA, P. E., MACH­ (1983). Neuropsychiatric disturbances associated with idiopathic cal­ LIN, S. R., & PEARLSON, G. D. (1996). Normal caudate nucleus in ob­ cification ofthe basal ganglia. Biological Psychiatry, 18, 591-60I. sessive-compulsive disorder assessed by quantitative neuroimaging. FENICHEL, O. (1945). The psychoanalytic theory 0/neurosis. New York: Archives o/General Psychiatry, 53, 577-584. Norton. BAER, L. (1994). Factor analysis ofsymptom subtypes ofobsessive com­ FLOR-HENRY, P., YEUDALL, L. T., KOLES, Z. J., & HOWARTH, B. G. pulsive disorder and their relation to personality and tic disorders. (1979). Neuropsychological and power spectral EEG investigations Journal o/Clinical Psychiatry, 55,18-23. of the obsessive-compulsive syndrome. Biological Psychiatry, 14, BARLOW, D. H. (1988). Anxiety andits disorders: The nature and treat­ 119-130. ment on anxiety andpanic. New York: Guilford Press. GARBER, H. J., ANANTH, J. v., CHIU, L. c., GRISWOLD, V. J., & OLDEN­ BAXTER, L. R., SCHWARTZ, J. M., BERGMAN, K. S., SZUBA, M. P., GUZE, DORF, W. H. (1989). Nuclear magnetic resonance study ofobsessive­ B. H., MAZZIOTTA, J. C., ALAZRAKI, A., SELIN, C. E., FERNG, H.-K., compulsive disorder. American Journal o/Psychiatry, 146, 1001-1005. MUNFORD, P., & PHELPS, M. E. (1992). Caudate glucose metabolic HANTOUCHE, E. G., & LANCRENON, S. (1996). Typologie moderne des rate changes with both drug and behavior therapy for obsessive- symptomes et des syndromes obsessionnels-compulsifs: Resultats OBSESSIVE-COMPULSIVE DISORDER 171

[Modern typology ofobsessive-compulsive symptoms and syndromes: A c., & COHEN, R. M. (1989). Cerebral glucose metabolic rates in Results]. Encephale, 22, 9-21. obsessive-compulsive disorder. Neuropsychopharmacology, 2, 23-28. HEAD, D., BOLTON, D., & HYMAS, N. (1989). Deficit in cognitive shift­ OPPENHEIM, S., & ROSENBERGER, J. (1991). Treatment ofa case ofob­ ing ability in patients with obsessive-compulsive disorder. Biological sessional disorder: Family systems and object relations approaches. Psychiatry, 25, 929-937. American Journal ofFamily Therapy, 19, 327-333. HEMBREE, E. A., FOA, E. B., & KOZAK, M. J. (1994). Obsessive­ PAULS, D. L., LECKMAN, J. E, TOWBIN, K E., ZAHNER, G. E., & COHEN, compulsive disorder. In B. B. Wolman & G. Stricker (Eds.), Anxiety D. J. (1986). A possible genetic relationship exists between Tourette's and related disorders: A handbook (pp. 177-193). New York: Wiley. syndrome and obsessive-compulsive disorder. Psychopharmacology HOEHN-SARIC, R, HARRIS, G. J., PEARLSON, G. D., Cox, C. S., MACH­ Bulletin, 22, 730-733. LIN, S. R., & CAMARGO, E. E. (1991). A fluoxetine-induced frontal PERANI, D., COLOMBO, C., BRESSI, S., BONFANTI, A., GRASSI, E, lobe syndrome in an obsessive compulsive patient. Journal ofClini­ SCARONE, S., BELLODI, L., SMERALDI, E., & FAZIO, E (1995). cal Psychiatry, 52,131-133. [18F]FDG PET study in obsessive-compulsive disorder: A clini­ HOEHN-SARIC, R, PEARLSON, G. D., HARRIS, G. J., MACHLIN, S. R, & cal/metabolic correlation study after treatment. British Journal of CAMARGO, E. E. (1991). Effects of fluoxetine on regional cerebral Psychiatry, 166,244-250. blood flow in obsessive-compulsive patients. American Journal of PERSE, T. L., GREIST, 1. H., JEFFERSON, J. W., ROSENFELD, R., & DAR, R Psychiatry, 148, 1243-1245. (1987). Flouvoxamine treatment of obsessive-compulsive disorder. INSEL, T R (1992). Toward a neuroanatomy of obsessive-compulsive American Journal ofPsychiatry, 144, 1543-1548. disorder. Archives ofGeneral Psychiatry, 49, 739-744. PITMAN, R. K. (1987). A cybernetic model of obsessive-compulsive INSEL, T R., DoNNELLY, E. E, LALAKEA, M. L., ALTERMAN, 1. S., & psychopathology. Comprehensive Psychiatry, 28, 334-343. MURPHY, D. L. (1983). Neurological and neuropsychological studies POSNER, M. 1., NISSEN, M. 1., & OGDEN, W. C. (1978). Attended and un­ ofpatients with obsessive-compulsive disorder. Biological Psychia­ attended processing modes: The role of set for spatial location. In try, 18, 741-751. H. L. Pick, Jr., & 1. 1. Saltzman (Eds.), Modes ofperceiving andpro­ JENIKE, M. A, BAER, L., BALLANTINE, T., MARTUZA, R L., TYNES, S., cessing information (pp. 137-157). Hillsdale, NJ: Erlbaum. GIRIUNAS, 1., BUTTOLPH, M. L., & CASSEM, N. H. (1991). Cingulotomy RAPOPORT, J. L. (1991). Recent advances in obsessive-compulsive dis­ for refractory obsessive-compulsive disorder: A long-term follow-up order. Neuropsychopharmacology, 5, 1-10. of33 patients. Archives ofGeneral Psychiatry, 48, 548-555. RASMUSSEN, S. A, & TSUANG, M. T. (1986). Clinical characteristic and JENIKE, M. A., BREITER, H. C., BAER, L., KENNEDY, D. N., SAVAGE, family history in DSM-Ill obsessive-compulsive disorder. American C. R, OLIVARES, M. J., O'SULLIVAN, R. L., SHERA, D. M., RAUCH, Journal ofPsychiatry, 143, 317-322. S. L., KEUTHEN, N., ROSEN, B. R., CAVINESS, V. S., & FILIPEK, P. A. RAUCH, S. L., JENIKE, M. A, ALPERT, N. M., BAER, L., BREITER, (1996). Cerebral structural abnormalities in obsessive-compulsive H. C. R, SAVAGE, C. R, & FISCHMAN, A J. (1994). Regional cerebral disorder. Archives ofGeneral Psychiatry, 53, 625-632. blood flow measured during symptom provocation in obsessive­ KELLNER, C. H., JOLLEY, R. R, HOLGATE, R. C., AUSTIN, L., LYDIARD, compulsive disorder using oxygen 15-labeled carbon dioxide and posi­ R B., LARAIA, M., & BALLENGER, 1. C. (1991). Brain MRI in obsessive­ tron emission tomography. Archives ofGeneral Psychiatry, 51, 62-70. compulsive disorder. Psychiatry Research, 36, 45-49. ROBINS, L. N., HELZER, J. E., WEISSMAN, M. M., ORVASCHEL, H., GRUEN­ KOLB, B. (1977). Studies on the caudate-putamen and then dorsomedial BERG, E., BURKE, 1. D., & REGIER, D. A. (1984). Lifetime prevalence thalamic nucleus ofthe rat: Implications for mammalian frontal-lobe of specific psychiatric disorders in three sites. Archives ofGeneral functions. Physiology & Behavior, 18,237-244. Psychiatry, 41, 949-958. LHERMITTE, E, PILLON, B., & SERDARU, M. (1986). Human autonomy ROBINSON, D., Wu, H., MUNNE, R A., ASHTARI, M., ALVIR, J. M.-J., and the frontal lobes: Part 1. Imitation and utilization behavior: A LERNER, G., KOREEN, A., COLE, K, & BOGERTS, B. (1995). Reduced neuropsychological study of 75 patients. Annals ofNeurology, 19, caudate nucleus volume in obsessive-compulsive disorder. Archives 326-334. ofGeneral Psychiatry, 52, 393-398. LOGUE, Y, DURWARD, M., PRATT, R. T C., PIERCY, M., & NIXON, W. L. ROLLS, E. T (1985). Connections, functions and dysfunctions oflimbic (1968). The quality oJ survival after an anterior cerebral aneurysm. structures, the prefrontal cortex and hypothalamus. In M. Swash & British Journal ofPsychiatry, 114, 13 7-160. C. Kennard (Eds.), Scientific basis ofclinical neurology (pp. 201­ LUCEY, J. Y, COSTA, D. c., BLANES, T, BUSATTO, G. E, PILOWSKY, 213). New York: Churchill Livingstone. L. S., TAKEI, N., MARKS, 1. M., ELL, P. J., & KERWIN, R. W. (1995). RUBIN, R. T, ANANTH, 1., VILLANUEVA-MEYER, 1., TRAJMAR, P. G., & Regional cerebral blood flow in obsessive-compulsive disorder pa­ MENA, 1. (1995). Regional 133xenon cerebral blood flow and cerebral tients at rest. British Journal ofPsychiatry, 167, 629-634. 99mTc-HMPAO uptake in patients with obsessive compulsive disor­ LUXENBERG, J. S., SWEDO, S. E., FLAMENT, M. E, FRIEDLAND, R. P., der before and during treatment. Biological Psychiatry, 38, 429-437. RAPOPORT, J., & RAPOPORT, S. 1. (1988). Neuroanatomical abnor­ RUBIN, R T, VILLANUEVA-MEYER, J., ANANTH, J., TRAJMAR, P. G., & malities in obsessive-compulsive disorder detected with quantitative MENA, 1. (1992). Regional xenon 133 cerebral blood flow and cere­ X-ray computed tomography. American Journal ofPsychiatry, 145, bral technetium 99m HMPAO uptake in unmedicated patients with 1089-1093. obsessive compulsive disorder and matched normal control subjects: MACH LIN, S. R, HARRIS, G. J., PEARLSON, G. D., HOEN-SARIC, R., JEF­ Determination by high-resolution single-photon emission computed FERY, P., & CAMARGO, E. E. (1991). Elevated medial-frontal cerebral tomography. Archives ofGeneral Psychiatry, 49, 695-702. blood flow in obsessive-compulsive patients: A SPECT study. Amer­ SALKOVSKIS, P. M., & WARWICK, H. M. C. (1985). Cognitive therapy of ican Journal ofPsychiatry, 148, 1240-1242. obsessive-compulsive disorder: Treating treatment failures. Behavioral MARTINOT, J. L., ALLILAIRE, J. E, MAZOYER, B. M., HANTOUCHE, E. G., Psychotherapy, 13,243-255. HURET, J., LEGAUT-DEMARE, E, DESLAURIERS, A., HARDY, P., PAP­ SCARONE, S., COLOMBO, C., LIVIAN, S., ABBRUZZESE, M., RONCHI, P., PATA, S., BARON, J., & SYROTA, A (1990). Obsessive-compulsive dis­ LOCATELLI, M., SCOTTI, G., & SMERALDI, E. (1992). Increased right order: A clinical, neuropsychological, and positron emission tomog­ caudate nucleus size in obsessive-compulsive disorder: Detection with raphy study. Acta Psychiatrica Scandanavica, 82, 233-242. magnetic resonance imaging. Psychiatry Research: Neuroimaging, 45, MAXMEN, J. S., & WARD, N. G. (1995). Essential psychopathology and 115-121. its treatment. New York: Norton. SCHWARTZ, J. M., STOESSEL, P. W., BAXTER, L. R, MARTIN, K M., & MCGUIRE, P. K., BENCH, C. J., FRITH, C. D., MARKS, 1. M., FRACKOWIAK, PHELPS, M. E. (1996). Systematic changes in cerebral glucose meta­ R. S. J., & DOLAN, R. J. (1994). Functional anatomy of obsessive­ bolic rate after successful behavior modification treatment ofobsessive­ compulsive disorder. British Journal ofPsychiatry, 164,459-468. compulsive disorder. Archives ofGeneral Psychiatry, 53, 109-113. MOWRER, O. H. (1960). Learning theory and behavior. New York: Wiley. STANLEY, M. A., & PRATHER, R C. (1993). Obsessive-compulsive dis­ NELSON, E., EARLY, T S., & HALLER, 1. W. (1993). Visual attention in order. In A. S. Bellack & M. Hersen (Eds.), Psychopathology in adult­ obsessive-compulsive disorder. Psychiatry Research, 49, 183-196. hood (pp. 164-178). Boston: Allyn & Bacon. NORDAHL, T. E., BENKELFAT, c., SEMPLE, W. E., GROSS, M., KING, STEIN, D. J., & HOLLANDER, E. (1992). Cognitive science and obsessive- 172 WILSON

compulsive disorder. In D. 1. Stein & 1. E. Young (Eds.), Cognitive tients with obsessive-compulsive disorder. Biological Psychiatry, 33, science and clinical disorders (pp. 235-246). San Diego: Academic 298-301. Press. nEN, A. Y., PEARLSON, G. D., MACHLlN, S. R., BYLSMA, E W., & STEIN, D. J., & HOLLANDER, E. (1994). A neural network approach to HOEHN-SARIC, R. (1992). Oculomotor performance in obsessive­ obsessive-compulsive disorder. Journal 0/Mind & Behavior, IS, compulsive disorder. American Journal o/Psychiatry, 149,641-646. 223-237. TOWEY, J. P., TENKE, C. E., BRUDER, G. E., LEITE, P., FRIEDMAN, D., STEIN, D. J. HOLLANDER, E., CHAN, S., DECARIA, C. M., HILAL, S., LIEBOWITZ, M., & HOLLANDER, E. (1994). Brain event-related po­ LIEBOWITZ, M. R., & KLEIN, D. E (1993). Computed tomography tential correlates ofoverfocused attention in obsessive-compulsive dis­ and neurological soft signs in obsessive compulsive disorder. Psy­ order. Psychophysiology, 31, 535-543. chiatry Research, 50, 143-150. TRIVEDI, M. H. (1996). Functional neuroanatomy of obsessive­ SWEDO, S. E., PIETRINI, P., LEONARD, H. L., SCHAPIRO, M. B., RETTEW, compulsive disorder. Journal o/ClinicalPsychiatry, 57(Suppl.), 26-35. D. C., GOLDBERGER, E. L., RAPOPORT, S. I., RAPOPORT, 1. L., & TuRNER, S. M., & MICHELSON, L. (1984). Obsessive-compulsive disor­ GRADY, C. L. (1992). Cerebral glucose metabolism in childhood­ der. In S. M. Turner (Ed.), Behavioral theories and treatment 0/anx­ onset obsessive-compulsive disorder: Revisualization during phar­ iety (pp. 239-277). New York: Plenum. macotherapy. Archives o/General Psychiatry, 49, 690-694. WARREN, R., & ZGOURIDES, G. D. (1991). Anxiety disorders: A rational­ SWEDO, S. E., RAPOPORT, J. L., CHESLOW, D. L., LEONARD, H. L., emotive perspective. New York: Pergamon. AYOUB, E. M., HOSIER, D. M., & WALD, E. R. (1989). High prevalence WILLIAMS, K. E., & CHAMBLESS, D. L. (1994). Behavioral therapies. In of obsessive-compulsive symptoms in patients with Sydenham's 8. 8. Wolman & G. Strucker (Eds.), Anxiety and related disorders: chorea. American Journal 0/Psychiatry, 146,246-249. A handbook (pp. 358-375). New York: Wiley. SWEDO, S. E., SCHAPIRO, M. B., GRADY, C. L., CHESLOW, D. L., ZOHAR, J., INSEL, T. R., BERMAN, K. E, FOA, E. 8., HILL, J. L., & WEIN­ LEONARD, H. L., KUMAR, A., FRIEDLAND, R., RAPOPORT, S. I., & BERGER, D. R. (1989). Anxiety and cerebral blood flow during be­ RAPOPORT, J. (1989). Cerebral glucose metabolism in childhood­ havioral challenge: Dissociation ofcentral from peripheral and sub­ onset obsessive-compulsive disorder. Archives 0/General Psychiatry, jective measures. Archives o/General Psychiatry, 46, 505-510. 46, 518-523. SWEENEY, J. A., PALUMBO, D. R., HALPER, J. P., & SHEAR, M. K. (1992). Pursuit eye movement dysfunction in obsessive-compulsive disorder. Psychiatry Research, 42, I-II. SWERDLOW, N. R., BENBOW, C. H., ZISOOK, S., GEYER, M. A., & BRAFF, (Manuscript received August 16, 1996; D. L. (1993). A preliminary assessment ofsensorimotor gating in pa- revision accepted for publication September 26, 1997.)