Volumetric Investigation of the Frontal-Subcortical Circuitry in Patients With Obsessive-Compulsive Disorder

Although the pathophysiology of obsessive-compulsive disorder (OCD) remains controversial, there is substantial evidence suggesting that disturbances in the frontal-subcortical circuitry may be implicated.^1,2 Evidence of this hypothesis stems mainly from the elevated rates of OCD in individuals with basal ganglia disorders^3,4 and from functional neuroimaging studies that indicate hyperfunction in the frontal-subcortical circuitry in patients with OCD.^5–7 However, despite evidence implicating functional abnormalities in the frontal-subcortical circuitry, these regions have not been well investigated by magnetic resonance imaging (MRI). The majority of studies have examined these areas separately, and have yielded contradictory results.^8–11 Recently, considerable evidence suggests that pathological changes in patients with OCD may be expressed at the level of the spatially distributed network that subsumes the multiple, densely interconnected cortical and subcortical cortices.^12–14 This shift from key regions to key neural circuits in the research on OCD may be attributed to the diverse clinical symptoms of the disorder and neuropsychological dysfunctions observed.

Our group recently reported on the increased gray matter density of the frontal subcortical circuitry in patients with OCD by employing a novel voxel-based analysis of segmented MR images.¹⁵ However, our finding of the increased regional density of the left orbitofrontal cortex is to some extent conflicting with previous reports by Szeszko et al.¹¹ In their study, Szeszko et al. reported reduced total orbital frontal volume in OCD employing the region of interest (ROI) method. In another MRI investigation using the semiautomated method for neocortical parcellation, Grachev et al.¹⁶ reported no volume difference in the orbital frontal regions of 10 female patients with OCD and 10 female normal control subjects. Furthermore, we could not find any regional density differences in the basal ganglia structures, which are consistently suggested to be implicated in the symptomatic expression of OCD, between the patients with OCD and normal control subjects.^1,2,4,8 Although these discrepancies between studies may partly arise from methodological differences between voxel-based morphometry (VBM) and ROI methods and the heterogenous study sample, it is necessary to reconcile this disparity in order to more accurately understand the pathophysiology of OCD.

To replicate our prior VBM reports employing ROI approach, we measured volumes of the orbitofrontal cortex, anterior cingulate gyrus, thalamus, and caudate nucleus, which are the main components of the frontal subcortical circuitry in patients with OCD and in normal subjects. On the basis of our previous findings of increased regional density of the left orbitofrontal cortex and thalamus, we hypothesized that patients with OCD would also show increased gray matter volumes in those regions in which ROI method was employed.

METHODS

Subjects

Thirty-six patients with OCD (28 men and eight women) were recruited from the OCD clinic at Seoul National University Hospital. These patients fulfilled the DSM-IV criteria¹⁷ for OCD, as diagnosed using the Structured Clinical Interview for DSM-IV (SCID).¹⁸ The normal control group consisted of 36 age- and sex-matched healthy volunteers (28 men and eight women) who were recruited from the community through newspaper advertisement. Exclusion criteria for OCD subjects and normal controls were the presence of a significant medical condition, neurological disorder, or history of other major psychiatric disorders such as substance abuse, schizophrenia, and bipolar disorder, which was screened using the SCID. This study was conducted under the use of human subjects established by the institutional review board. After a complete description of the scope of the study, written informed consent was obtained from all subjects.

The mean ages of the OCD group and the control group were 26.33 years (SD=6.18) and 26.33 years (SD=7.58), respectively. The mean periods of education were 13.86 years (SD=2.04) and 15.17 years (SD=1.89) (t=2.814, df=70, p=0.006), respectively, and the mean socioeconomic status¹⁹ was ranked 3.00 (SD=0.71) and 2.86 (SD=0.73) (t=−0.79, df=70, P=0.43), respectively. All patients with OCD and normal control subjects were right-handed.²⁰ At the time of the study, the patients had a mean duration of illness of 8.90 years (SD=7.00), ranging from 1 year to 26 years. Four patients with OCD had a major depressive disorder as a comorbidity, but had no other DSM-IV axis I disorder, and five patients with OCD had histories of transient tic disorder, which spontaneously resolved. Eleven patients were drug naïve, and 25 received a prior treatment for their disorder, including selective serotonin reuptake inhibitors or clomipramine (six included a history of combined therapy with neuroleptics), but they all remained psychotropic-free for a period of at least 4 weeks. Clinical assessment included the Yale-Brown Obsessive Compulsive Scale²¹ (Y-BOCS) for determining OCD symptom severity (mean score for obsessive symptom= 13.31, SD=3.54; mean score for compulsive symptom= 11.22, SD=4.26; mean total score=24.17, SD=5.94).

Magnetic Resonance Image Acquisition and Image Processing

Three-dimensional T₁-weighted spoiled gradient echo MR images were acquired on a 1.5-T GE SIGNA Scanner (GE Medical System, Milwaukee). Imaging parameters were: 1.5-mm sagittal slices, echo time=5.5 msec, repetition time=14.4 msec, number of excitation=1, rotation angle=20°, field of view=21×21 cm, and a matrix of 256×256. Magnetic resonance images were processed using an image-processing software package, ANALYZE (version 4.1, Mayo Foundation, Rochester, Minnesota). Images were resampled to 1.0 mm³ voxels, reoriented to the conventional position, and spatially realigned so that the anterior-posterior axis of the brain was aligned parallel to the intercommissural line, and the other two axes were aligned along the interhemispheric fissure. The data sets were then filtered using anisotropic diffusion methods to improve the signal-to-noise ratio. The semiautomated region growing method removed images of tissues exterior to the brain. The extracted brain images were segmented into gray matter, white matter, and cerebrospinal fluid by employing the fuzzy C-mean algorithm.²² Intracranial volume was calculated by adding the volumes of these three components.

Boundary of Region of Interests

We focused on 12 cortical and subcortical subregions (both right and left), namely, the orbitofrontal cortex, rostral-anterior cingulate gyrus, caudal-anterior cingulate gyrus, caudate, putamen, and thalamus. Tracing was performed manually on all coronal slices in the ANALYZE ROI module.

Anterior Cingulate Gyrus

Boundary definition and the method of tracing anterior cingulate gyrus were performed as described by Crespo-Faccoro et al.²³ As previously reported by Crespo-Faccoro et al., we chose 1) the coronal plane passing through the point where the central sulcus intersects with the midsagittal plane (plane A) as the border between the anterior and posterior cingulate gyrus, and 2) the coronal plane (plane B) passing through the anterior-most tip of the inner surface of the genu of the corpus callosum as the border between the rostral and caudal anterior cingulate gyrus. The deepest point of the callosal sulcus and the most medial point of the dorsal bank of the cingulate sulcus constitute the inner and outer boundaries of the anterior cingulated gyrus in each coronal slice, respectively (Figure 1). Tracing begins at the plane A and continues rostrally to the most anterior coronal slice containing anterior cingulate gyrus tissue.

Orbitofrontal Cortex

Methods for measuring the orbitofrontal cortex were adapted from Crespo-Faccoro et al.,²³ with several modifications for use in our MRI²⁴ (Figure 1). Differing from previous methods by Crespo-Faccoro et al.,²³ the straight gyrus and the medial frontal cortex was included in the delineation of the orbitofrontal cortex in this study because it is difficult to distinguish these areas in the absence of triangulated line.²⁴ Before beginning the tracing, the lateral orbital sulcus was identified on the intermediate aspect of the frontal lobe, where it was clearly seen. On the anterior portion of the frontal lobe, the frontomarginal sulcus constitutes the lateral boundary of the orbitofrontal cortex until the lateral orbital sulcus appears. Moving caudally on the serial coronal slices, the deepest point of the lateral orbital sulcus constitutes the lateral boundary of the orbitofrontal cortex. When the lateral orbital sulcus disappears on the coronal slices, the inferior margin of the circular sulcus of the insula defines the lateral boundary of the orbitofrontal cortex caudally. The deepest inferior medial point of the cingulated sulcus constitutes the medial boundary of the orbitofrontal cortex on the posterior and intermediate portions of the orbitofrontal cortex. On the anterior portion, the deepest point of the superior rostral sulcus represents the medial boundary of the orbitofrontal cortex. Anterior boundary is defined by anterior extreme of the frontal pole, and posterior boundary is defined as the most posterior coronal slice that contains some aspect of the posterior medial orbitofrontal gyrus.

Caudate and Putamen

The caudate and the putamen were traced on coronal slices from the rostral to caudal aspect, first on the right side of the image and then on the left side (Figure 1). The sequence of tracing was right caudate, left caudate, right putamen, and left putamen. Tracing commenced when the structure could be seen with the naked eye and ended when the structure was no longer discernible. Anatomical landmarks and boundaries for the caudate and putamen were determined with the aid of the human brain atlas.²⁵

Thalamus

To define the anterior and the posterior boundaries of the thalamus on coronal slices, the sagittal plane was used as a reference line. The main tracing was performed on the coronal plane, in which the boundaries of the thalamus were based on the detailed guidelines of Portas et al.²⁶ the internal capsule, laterally; the third ventricle and the habenular nucleus, medially; the lateral ventricle and the crux fornix, dorsally; and the zona inserta and the red nucleus, ventrally (Figure 1). The lateral and medial geniculate bodies were included as a part of the thalamus.

The interrater reliability for each structure volume was assessed in a randomly selected subset of 10 MR scans. As shown in Table 1, the intraclass correlation coefficient (ICC) for the reliability of gray matter volume ranged from a low of r=0.76 (right putamen) to a high of r=0.97 (right thalamus).

Statistical Analyses

We used t tests to examine group differences in demographic and clinical variables. All measures of cortical regional gray matter volumes were subjected to analysis of covariance (ANCOVA), with diagnosis as a between-group factor and the total brain volume as a covariate. The relations between the volumes of each region and subjects characteristics (age, duration, and symptom severity) were examined using Pearson correlation coefficients. All statistical analyses were two-tailed, and P<0.05 was used as the level of significance. Brain structures were examined separately because of their functional and neuroanatomical heterogeneity and the possibility that OCD may involve structural pathology in a single frontal region or subcortical region.

RESULTS

Mean intracranial volume was 1,454.81 cm³ (SD= 135.07) for the OCD group, and 1,408.98 cm³ (SD=144.33) for the normal comparison group, which was not statistically different (t=−1.396, df=70, P=0.679). Table 1 shows gray matter volumes of each ROI subregion and the results of the ANCOVA, taking total brain volume as a covariate and diagnosis as a grouping variable. As shown in Table 1, the left orbitofrontal volume was significantly smaller in patients with OCD (F=5.352, df=1, P=0.024). No significant differences were found in the volumes of the other cortical and subcortical regions measured in this study, which included the right orbitofrontal cortex, anterior cingulate gyrus, thalamus, caudate, and putamen. The volume of the left orbitofrontal cortex showed significant negative correlation (r=−0.402, P=0.017) with the total Yale-Brown Obsessive Compulsive Scale (Y-BOCS) scores of the OCD patients (Figure 2), and the right orbitofrontal cortex volume negatively correlated (r=−0.344, P=0.040) with the compulsive subscale of Y-BOCS. However, no significant association was found between brain structure volumes and the duration of illness or age at clinical symptom onset in patients with OCD.

DISCUSSION

The main finding of this study is that the left orbitofrontal volume is reduced in OCD patients. Although there have been relatively few volumetric studies of the orbitofrontal cortex in OCD, our finding is partially consistent with a previous report by Szeszko et al.¹¹ In their study, Szeszko et al. reported reduced total orbital frontal volume, which included both gray and white matter in OCD, while only the left orbital frontal volume showed a significant group difference. For a parcellation of the orbitofrontal cortex, however, Szeszko et al.¹¹ employed topographic methods based on surface sulcal landmarks, which involve the substantial variability of the interindividual and interhemispheric gyral and sulcal patterns.^23,27 Yet, a more detailed topographic method was used in our study for a delineation of the orbitofrontal cortex, as suggested by Crespo-Faccoro et al.,²³ to reflect interindividual variability of the orbitofrontal cortex. In another MRI investigation using semiautomated method for neocortical parcellation, Grachev et al.¹⁶ reported no volume difference in the orbital frontal regions of 10 female patients with OCD and 10 female normal control subjects. However, the statistical capacity of their study may have been limited by their small sample size. In the current study, reduced volume of the left orbitofrontal cortex in OCD was shown, using a more detailed topographic method for a delineation of the orbitofrontal cortex.

However, our finding of reduced volume of the left orbitofrontal cortex conflicts with that of our prior study, which employed VBM in OCD.¹⁵ In the previous study, we reported increased gray matter density of the left orbitofrontal cortex. There are several hypotheses accounting for these discrepant results. First, the regional differences detected in the previous study should be viewed as representing foci of maximal change rather than regions that are exclusively affected because VBM does not reflect changes of the whole region of the orbitofrontal structure. Second, the sources of differences may arise from the changes in the shape or displacement of structures in the course of spatial normalization.²⁸ Finally, statistical problems of multiple comparisons may cause the different findings between studies, although we attempted to apply a strict criterion (clusters of more than 50 contiguous voxels with a threshold of P<0.001) to avoid type I errors in the VBM study. In particular, Kubichi et al.²⁸ emphasized the importance of comparisons with manual ROI studies as well as careful interpretation of the results, given that factors other than volume alone influence VBM results.

In the present study, a significant negative correlation was found between the left orbitofrontal volume and the symptom severity, as measured by Y-BOCS in patients with OCD, but no significant correlations were found between the gray matter volumes demarcated and the duration of illness or the age at onset. In the previous functional studies, it was reported that the orbitofrontal cortex had increased glucose metabolic rates when compared to normal control subjects in resting states^7,29 and decreased rates after successful treatment of symptoms .^5,30 This suggests that the orbitofrontal cortex region may be involved in mediating the expression of obsessive-compulsive symptoms. Furthermore, there has been much experimental and clinical evidence that the orbitofrontal cortex is involved in the mediation of emotional response to biologically significant stimuli and in the inhibition of behavioral response.³¹ Our results suggest that a structural abnormality of the left orbitofrontal cortex plays a critical role in symptomatic expression of OCD.

On the other hand, no significant group differences were found in the anterior cingulate gyrus, thalamus, and basal ganglia structures. Prior reports of a structural investigation of the caudate nucleus are controversial.^8,9,32,33 Yet, our results are consistent with those of the previous studies, which reported normal caudate volume.^9,32,33 In a comprehensive analysis of prior structural and functional imaging studies, Aylward et al.⁹ suggested that the heterogenous nature of OCD might explain the inconsistencies between studies. In particular, they insisted that structural aberrations of the caudate nucleus are more evident in patients with ventricular enlargement and the soft neurological signs that are common in childhood onset. Patients in the present study were relatively free from soft neurological signs such as tics (only five patients had tic history). The nonsignificant group differences identified in the thalamus and in the anterior cingulate are consistent with previous MRI investigations in adults with OCD, which uniformly yielded negative findings.^10,11,32

The most striking finding in the present study is that only the left orbitofrontal volume was significantly reduced in patients with OCD, and there were no additional structural abnormalities in other brain regions composing the frontal-subcortical tracts (i.e., abnormalities that have been implicated in previous functional neuroimaging studies of OCD, as mentioned above). One possible explanation for this morphological and functional discrepancy might be related to the methodological limitations of delineating ROIs in the current study. These may have been large enough to prevent the detection of subtle neuronal changes, and thus a significant change in some neuronal field could have been obscured, as surrounding areas within the same ROI showed no changes. In fact, other measures such as 1H-magnetic resonance spectroscopy of N-acetylaspartate have identified abnormalities of the thalamus,^34,35 anterior cingulate,³⁶ and striatum^36,37 in OCD patients.

Despite these thought-provoking results, our study has several limitations. First, it does not provide information about white matter volumes because of difficulties concerning the localization of clear anatomic landmarks that define white matter subregions, which would have helped in the interpretation of our results. Second, we parceled most of the main subregions composing the frontal-subcortical circuitry, which inevitably leads to an increase in the number of comparisons, and thus increases the risk of a type I error in the interpretation of results. To conclude, we found reduced left orbitofrontal volume in patients with OCD and significant negative correlations between the bilateral orbital frontal cortex and symptom severity by simultaneously investigating the main structures composing the frontal-subcortical circuitry. These findings suggest that a structural abnormality of the left orbitofrontal cortex plays a critical role in both the clinical and cognitive expression of OCD. Future research to determine the involvement of white matter tracts of this circuitry may shed further light on the pathophysiology of this disorder.

ACKNOWLEDGMENTS

This study was supported by the Korean Research Foundation (2001-041-F00182).

TABLE 1. Absolute Volumes (cm³ of Cortical and Subcortical Areas for Patients Versus Normal Controls^a

Enlarge table