ORIGINAL ARTICLE Disorder-Specific Neuroanatomical Correlates of Attentional Bias in Obsessive-compulsive Disorder, Panic Disorder, and Hypochondriasis Odile A. van den Heuvel, MD, PhD; Dick J. Veltman, MD, PhD; Henk J. Groenewegen, MD, PhD; Menno P. Witter, PhD; Jille Merkelbach, MD; Danielle C. Cath, MD, PhD; Anton J. L. M. van Balkom, MD, PhD; Patricia van Oppen, PhD; Richard van Dyck, MD, PhD Context: Attentional bias to disease-relevant emo- Results: During incongruent vs congruent color nam- tional cues is considered to be pathogenetically relevant ing, all patient groups recruited additional posterior brain in anxiety disorders. regions relative to controls, but performance was im- paired only in OCD. In OCD, color naming OCD- Objective: To investigate functional neural correlates related, but not PD-related, words correlated with in- and disease specificity of attentional bias across differ- creased activation of frontal-striatal and temporal regions, ent anxiety disorders. although performance was unimpaired. In contrast, in PD, increased frontal-striatal involvement was found during color naming both OCD-related and panic-related words. Design: A cognitive and emotional Stroop task, consist- In PD, color naming panic-related words was slowed and ing of congruent and incongruent color words, obsessive- correlated with increased activation of the right amyg- compulsive disorder (OCD)–related and panic-related dala and hippocampus. Patients with hypochondriasis negative words, and neutral words, was used in 3 pa- showed a similar activation pattern to patients with PD. tient groups and a control group during functional mag- netic resonance imaging. Conclusions: Our results support the hypothesis of in- creased distractibility for irrelevant information in pa- Setting: Academic outpatient department for anxiety dis- tients with OCD, PD, and hypochondriasis associated with orders. frontal-striatal and limbic involvement compared with controls. Although patients with OCD did not display an attentional bias in behavior relative to controls, there was Patients and Participants: Medication-free patients a clear, specific neural response during color naming with OCD (n=16), panic disorder (PD) (n=15), and hy- OCD-related words, involving mainly ventral brain re- pochondriasis (n=13) and 19 controls. gions. In contrast, generalized emotional interference ef- fects were found in PD and hypochondriasis, involving Main Outcome Measure: Voxel-wise analyses of ce- ventral and widespread dorsal brain regions, reflecting rebral blood flow changes for contrasts of interest (in- not only unconscious emotional stimulus processing but congruent vs congruent color words, OCD-related vs neu- also increased cognitive elaboration. tral words, and panic-related vs neutral words) within and between groups. Arch Gen Psychiatry. 2005;62:922-933 Author Affiliations: Departments of Psychiatry IFFICULTY INHIBITING IR- sumably not simply a by-product of emo- (Drs van den Heuvel, Veltman, relevant information, for tional state but may play a major role in 4-6 Cath, van Balkom, van Oppen, instance obsessive symptom causation and maintenance. and van Dyck) and Anatomy thoughts and impulses, is The critical process of gating (ie, inhibit- (Drs Groenewegen and Witter), a key feature of obsessive- ing irrelevant information) has been linked VU University Medical Center, Dcompulsive disorder (OCD).1,2 Because to frontal-striatal function.7,8 Impaired Amsterdam, the Netherlands; most of their attentional resources are al- frontal-striatal function is considered to be Department of Psychiatry, Leiden located to threat cues related to their con- of etiological importance in the affective, University Medical Center, cerns, patients with OCD are limited in behavioral, and cognitive characteristics Leiden, the Netherlands their ability to selectively attend to rel- of OCD.9 In contrast, the brainstem and (Dr Merkelbach); Outpatient Department for Anxiety evant information while simultaneously ig- limbic regions, such as the amygdala and Disorders, GGZ Buitenamstel, noring irrelevant competing informa- (para)hippocampal region, are mainly im- 1 10-13 Amsterdam (Drs Cath, tion. Similar cognitive dysfunctions have plicated in the symptoms of PD. van Balkom, van Oppen, and been described for panic disorder (PD).3 Cognitive interference occurs when van Dyck). Attentional bias in anxiety disorders is pre- processing one stimulus impedes simul- (REPRINTED) ARCH GEN PSYCHIATRY/ VOL 62, AUG 2005 WWW.ARCHGENPSYCHIATRY.COM 922 ©2005 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/26/2021 taneous processing of another stimulus, as in the Stroop nificant activation in the affective (rostral) subdivision of color-word task.14 This task requires the ability to ac- the ACC during emotional interference in the absence of tively inhibit a fast automatic response (word reading) a behavioral interference effect. In contrast, the tradi- in favor of a slower voluntary response (color naming), tional Stroop task was correlated with increased activa- resulting in increased response latencies and error rates tion of the dorsal ACC (cognitive subdivision of the cin- during the interference condition (eg, “red” printed in gulate cortex).22 The authors suggested that rostral ACC blue ink) compared with the facilitation (or baseline) con- activation reflected a regulatory response during success- dition (eg, “red” printed in red [or black] ink). ful suppression of task-irrelevant emotional information. Frontal involvement in interference processes was first Neuropsychological studies using the emotional Stroop demonstrated in lesion studies, showing impaired Stroop task in various anxiety disorders, such as generalized anxi- performance in patients with lesions in the left15 and right16 ety disorder,44-46 PD,47-53 OCD,53-56 posttraumatic stress dis- lateral prefrontal cortex. More extensive investigation of order (PTSD),57-59 social phobia,46,60-64 and spider pho- the neural correlates of interference in healthy subjects bia,65,66 all (but 153) showed robust interference effects in was undertaken with the advent of positron emission to- patients during color naming threat-related words. It is not mography,17-21 functional magnetic resonance imaging yet clear, however, whether this effect is specific to disease- (fMRI),22-31 and magnetoencephalography.32 In many,* related threat words or extends to general threat domains but not all,18,21,26,29,32 studies, interference was associ- (or even positive emotional material), although some anxi- ated with increased activation of the anterior cingulate ety disorders seem to be more specific in their attentional cortex (ACC). Prefrontal involvement in interference was bias than others. Whereas patients with OCD54-56,67 and so- reported in dorsolateral (DLPFC),19,22,23,26-29,31,32 anterior cial phobia46,63,64 showed a predominantly disease- (aPFC),18,20,29,32 and ventrolateral (VLPFC)21,25,26,28,29 pre- specific attentional bias, threat word interference in gen- frontal cortex subregions. Recent imaging studies have eralized anxiety disorder46 and PD49 was found to be more shown that ACC is primarily involved in response- generalized.51,52 A second issue of consideration in emo- related processes, such as performance monitoring, re- tional interference is the explicit vs implicit nature of the sponse conflict, and error detection.24,27,28,31,33,34 In con- attentional bias. Most studies using subliminal stimulus trast, various prefrontal regions seem to be primarily presentation support the assumption that interference is responsible for the implementation of top-down atten- not dependent on conscious strategies.5,52,68,69 tional control (ie, higher order regulation of hierarchi- To summarize, evidence from imaging data in healthy cally lower attentional processes) per se, although the subjects as well as from neuropsychological studies in relative contribution of specific prefrontal subregions subjects with an anxiety disorder appear to be in agree- is insufficiently clear.24,27,28,31 Other regions associated ment with the hypothesis of frontal involvement in emo- with interference were (pre)motor,17,22,25,32 tem- tional Stroop interference. However, direct evidence for poral,17,22,23,25,29,31 parietal,18-20,22,23,28,29,32 the underlying neuronal substrate in anxiety disorders extrastriate17,19-21,23,26,31 and insular19,20,25 cortices, thala- is lacking, with a single exception.70 Shin et al70 found mus,19,20,23 and striatum.17,19,23,25 activation of the rostral ACC in Vietnam veterans with- In patients with OCD, neuropsychological studies us- out PTSD during trauma-related emotional interfer- ing the Stroop color-word task have provided mixed evi- ence, similar to the results of Whalen et al43 in healthy dence for impaired selective attention compared with con- volunteers, but not in veterans with PTSD. Shin et al sug- trol subjects.1,35-40 Some studies showed that patients with gested that the absence of rostral ACC activation in pa- OCD performed slower than control subjects1,35 and made tients with PTSD reflected the inability of this region to significantly more errors and had slower reaction times inhibit amygdalar hyperresponsivity, although no amyg- during the interference condition of the Stroop task.1,37 dala activation was found during threat-related emo- This interference effect in patients with OCD was aug- tional processing in their patients with PTSD. They did mented when situational