Evidence for Reduced Cerebellar Volumes in Trichotillomania

Evidence for Reduced Cerebellar Volumes in Trichotillomania Nancy J. Keuthen, Nikos Makris, John E. Schlerf, Brian Martis, Cary R. Savage, Katherine McMullin, Larry J. Seidman, Jeremy D. Schmahmann, David N. Kennedy, Steven M. Hodge, and Scott L. Rauch Background: Limited knowledge exists regarding the neurobiology of trichotillomania (TTM). Cerebellum (CBM) volumes were explored, given its role in complex, coordinated motor sequences. Methods: Morphometric magnetic resonance imaging (MRI) scans were obtained for 14 female subjects with DSM-IV diagnoses of TTM and 12 age-, education-, and gender-matched normal control (NC) participants. Parcellation was performed utilizing a recently developed methodology to measure subterritory volumes of the CBM. Regions were defined based on knowledge of the structural and functional subunits of the CBM. Results: As predicted, significant group differences were reported for CBM raw cortical volumes (p ⫽ .008) that survived correction for total brain volume (TBV; p ⫽ .037) and head circumference (HC; p ⫽ .011). A priori and post hoc group raw volume comparisons for CBM subterritories and functional clusters revealed many significant differences. However, most differences failed to withstand correction for total CBM volumes (TCV). Smaller volumes were consistently reported for the TTM versus NC cohorts. Total Massachusetts General Hospital Hair Pulling Scale (MGHHPS) scores were significantly inversely correlated with left primary sensorimotor cluster volumes (p ⫽ .008), with smaller volumes associated with more severe TTM symptoms. Conclusions: These findings implicate the CBM in the neurobiology of TTM, with reduced subterritory volumes reported for the TTM versus NC groups. Key Words: Cerebellum, hair pulling, morphometry, MRI, OC spectrum, trichotillomania

T

richotillomania (TTM) remains a poorly understood and inadequately treated disorder despite increasingly greater recognition of its prevalence (Christenson et al 1991). It is a heterogeneous disorder that can encompass considerable variability in symptom frequency, severity, and pattern. The clinical presentation of TTM involves repetitive and coordinated motor behaviors of touching and stroking the hair, ultimately culminating in hair extraction. From a phenomenological perspective, sensory and affective variables play cardinal roles in both symptom triggering and maintenance. Significant physical and psychological sequelae (e.g., O’Sullivan et al 1996; Soriano et al 1996) and functional impairments (Diefenbach et al 2003; Keuthen et al 2004) are known to accompany the disorder. Its negative impact on psychosocial functioning can involve interpersonal withdrawal, relationship conflict, limitations in career choice, and avoidance of leisure activities. Trichotillomania has most commonly been conceptualized as an obsessive-compulsive spectrum disorder, sharing similarities with obsessive-compulsive disorder (OCD) and Tourette syndrome (TS). To date, however, limited knowledge has been available regarding its neurobiology. Much of the available

From the Department of Psychiatry (NJK, LJS, SLR), Center for Morphometric Analysis (NM, JES, DNK, SMH), Department of Neurology (NM, JES, JDS, DNK), A. Martinos Center (NM, JES, LJS, DNK, SLR), NMR Center (DNK, SLR), and Psychiatric Neuroimaging Research Program (SLR), Massachusetts General Hospital, Boston, Massachusetts; Department of Psychiatry (BM), University of Michigan, Ann Arbor, Michigan; Hoglund Brain Imaging Center and Department of Psychiatry and Behavioral Sciences (CRS), University of Kansas Medical Center, Kansas City, Kansas; Washington University (KM), St. Louis, Missouri; and Department of Psychiatry (LJS), Beth Israel Deaconess Medical Center, Boston, Massachusetts. Address reprint requests to Nancy J. Keuthen, Ph.D., Trichotillomania Clinic and Research Unit, Simches Research Building, Floor 2, 185 Cambridge Street, Boston, MA 02114; E-mail: [email protected]. Received May 4, 2005; revised March 27, 2006; accepted June 1, 2006.

0006-3223/07/$32.00 doi:10.1016/j.biopsych.2006.06.013

research has been driven by the noted similarities in clinical presentation, phenomenology, and treatment response between TTM and both OCD and TS, and the existing neurobiological studies of TTM also reflect this trend. Current neurobiological models for OCD (Baxter et al 1990; Rauch et al 1998; Saxena et al 1998; Graybiel and Rauch 2000) and TS (Rauch et al 1998; Stern et al 2000) implicate parallel, segregated cortico-striato-thalamocortical (CSTC) circuits with differences in clinical presentation between the two disorders attributed to different sites of striatal pathology. The cognitive intrusions or obsessions of OCD have most often been linked to caudate abnormalities (e.g., Jenike et al 1996; Robinson et al 1995), whereas the sensorimotor intrusions of TS have been tied to putamen differences (Peterson et al 1993; Singer et al 1993), although there are more recent exceptions to this heuristic scheme (e.g., Peterson et al 2003; Bloch et al 2005). Notably, while TTM does not present with obsessions or other cognitive phenomena, it does involve sensorimotor phenomena. This observation led investigators in our group (O’Sullivan et al 1997) to compare putamen, pallidum, and lenticulate volumes in TTM and matched normal control (NC) groups. Consistent with earlier findings for TS patients, our TTM sample had smaller left putamen volumes than their matched counterparts, thus providing neurobiological evidence for a link between TTM and TS. In our current study, we sought to explore evidence for other structural abnormalities in TTM. Several pieces of converging evidence, including the hallmark stereotyped motor routines characteristic of TTM, led us to investigate the cerebellum (CBM). In addition, the landmark positron-emission tomography (PET) study of TTM conducted by Swedo et al (1991) demonstrated hypermetabolism in right and left cerebellar areas, in addition to hypermetabolism in the right superior parietal area and increased global metabolic rates. In that study, TTM severity was also inversely correlated with right cerebellar and left caudate metabolic activity. Furthermore, our advancing knowledge of the CBM, indicating its involvement in higher order brain functions above and beyond its role in the control and integration of motor activity, also prompted us to study its role in TTM. Schmahmann (1991, BIOL PSYCHIATRY 2007;61:374 –381 © 2007 Society of Biological Psychiatry

N.J. Keuthen et al 1996) and Schmahmann and Pandya (1997) have mapped out cerebellar pathways with multiple subsystems responsible for information processing in the motor, sensory, affective, cognitive, and autonomic domains. More specifically, these corticocerebellar-thalamo-cortical (CCTC) circuits, similar to the CSTC pathways implicated in OCD and TS, are proposed to connect the CBM with cerebral sensorimotor and association areas via feedforward (Schmahmann and Pandya 1997) and feedback pathways (Schmahmann 1996; Kelly and Strick 2003). Thus, the significance of both sensory and affective variables in TTM, as outlined earlier, further suggests a role for cerebellar abnormalities in TTM. In this study, we postulated a priori hypotheses predicting differences in CBM cortical volumes between TTM and matched normal control groups. We also predicted group differences in CBM volumes for those functional clusters involved in autonomic/emotional and sensorimotor processing (Schmahmann et al 2000; Makris et al 2005). Lastly, significant correlations were anticipated between TTM severity and both total CBM cortical volumes and the above-mentioned subterritory volumes. We investigated these hypotheses using morphometric magnetic resonance imaging (MRI) and a recently developed method for CBM parcellation (Makris et al 2003, 2005).

Methods and Materials Subjects The study sample consisted of 14 TTM and 12 NC participants. Study subjects were recruited from the Massachusetts General Hospital (MGH) Trichotillomania Clinic and Research Unit, online study advertisements posted on the MGH research and Trichotillomania Learning Center (TLC) bulletin boards, and publications including local newspapers and the TLC newsletter In Touch. Written informed consent was obtained from all subjects prior to study involvement, in accordance with the MGH Institutional Review Board. All subjects were female, between 18 and 45 years of age, and right-handed (confirmed by the Edinburgh Inventory; Oldfield 1971). Trichotillomania and NC subjects were group-matched for age (TTM: 29.2 ⫾ 6.7; NC: 28.3 ⫾ 2.06; t ⫽ .02, p ⫽ .89) and education (TTM: 17.1 ⫾ 2.1; NC: 17.2 ⫾ 2.4; t ⫽ .12, p ⫽ .74). All TTM subjects satisfied DSM-IV criteria for TTM established by a semistructured clinical interview conducted by the first author (N.J.K.). Study participants were administered the Structured Clinical Interview for DSM-IV diagnosis (SCID; First et al 1995). Subjects did not meet criteria for any obsessive-compulsive spectrum disorder or other psychiatric condition, with the exception of one TTM subject with generalized anxiety disorder and two NC subjects with specific phobias (small animal and blood injury types). None of the subjects had a history of significant head injury, seizures, or other neurologic or medical conditions that would interfere with study procedures or confound results. No participants had lifetime diagnoses of any psychotic disorder or substance dependence. None of the subjects used any psychotropic medication within 4 weeks of scanning. Additionally, none of the subjects used fluoxetine within 6 weeks of scanning or any neuroleptic medications within 12 months of study participation. All subjects were administered urine tests immediately prior to scanning to rule out the possibility of pregnancy. To further confirm TTM diagnoses and to assess TTM history and symptoms, participants were administered a battery of clinician-rated TTM scales: the National Institute of Mental Health

BIOL PSYCHIATRY 2007;61:374 –381 375 (NIMH) TTM scales (Swedo et al 1989) and the Psychiatric Institute Trichotillomania Scale (Winchel et al 1992). The MGH Hair Pulling Scale (MGHHPS) (Keuthen et al 1995), a seven-item patient-rated scale, was the primary measure of TTM symptom severity. The mean total MGHHPS score for the TTM group was 18.14 (SD ⫽ 3.44, range: 12–25), indicating moderate to severe TTM symptom severity. All TTM subjects exhibited hair-pulling symptoms for ⱖ4 months. No significant group differences were reported for either the Beck Depression Inventory (Beck et al 1961) (TTM: 2.93 ⫾ 3.47; NC: 1.17 ⫾ 1.95; t ⫽ 1.56, p ⫽ .13) or the Beck Anxiety Inventory (Beck et al 1988) (TTM: 2.07 ⫾ 2.56; NC: 1.42 ⫾ 1.93; t ⫽ .73, p ⫽ .47). Image Acquisition We used a Sonata 1.5 Tesla whole body high-speed imaging device (Siemens Medical Systems, Iselin, New Jersey) with a three-axis gradient head coil. Head movement was restricted with expandable foam cushions. After an automated scout image was acquired and shimming procedures were performed to optimize field homogeneity (Reese et al 1995), two high-resolution, three-dimensional magnetization-prepared rapid acquisition gradient echo (3-D MPRAGE) sequences (repetition time [TR] 7.25 milliseconds, echo time [TE] 3 milliseconds, flip angle 7°, voxel size 1.3 x 1 x 1 mm) were collected. These replicate structural images were then individually motion corrected and averaged. Parcellation Method Our goal in this study was to perform a volumetric assessment of the CBM, including total and cortical versus white matter compartments, as well as gyral-based, biologically relevant substructures of the CBM. While methods such as voxel-based morphometry or other surface-based methods could be employed, these methods tend to result in a localization of intergroup differences, not a comprehensive volumetric assessment of structure. Furthermore, the sensitivity of these other methods is related to the underlying intersubject registration methods employed. Our method used direct observation of the relevant anatomy and did not rely on automated registration methods. Our parcellation method does not affect volumetric measurements, as surface manipulations are for identification of the necessary anatomic landmarks, and all volumetric measurements are made in the original geometric space of the MRI scan. The coordinate system (Talairach and Tournoux 1988; Filipek et al 1988, 1994) utilized to analyze each brain defined the y axis (anterior-posterior) as the bicommissural line (anterior commissure-posterior commissure). The superior-inferior z axis was orthogonal to the y axis, passing through the interhemispheric fissure. The medial-lateral x axis was perpendicular to both the y and z axes. Within this coordinate system, the coronal plane was defined by the x-z axes, the sagittal plane by the y-z axes, and the transaxial planes by the x-y axes. This positional normalization procedure allowed the reconstruction specified by the x-z axes of a new set of coronal images at the slice thickness of the original acquisition (1.5 mm). Parcellation was performed by one rater utilizing a previously published methodology (Makris et al 2003, 2005). Parcellation of the CBM divided the cortex into 32 individual parcellation units (PUs) per hemicerebellum (64 for the entire CBM), which were then clustered together. Subdivision into individual PUs was based on a grid of a constant set of 11 mediolateral “limiting” fissures, which intersect with a longitudinal paravermian plane and two more laterally situated parasagittal planes (Makris et al 2003, 2005). To www.sobp.org/journal

376 BIOL PSYCHIATRY 2007;61:374 –381 parcellate the CBM, two different environments were used interactively, namely CardViews (Center for Morphometric Analysis, Massachusetts General Hospital, Boston, Massachusetts) (Caviness et al 1996), which allows manipulations in the volume domain, and FreeSurfer (CorTechs Labs, Inc., La Jolla, California, and the Athinoula A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital, Boston, Massachusetts) (Dale et al 1999) for operations on the cerebellar surface. Once the CBM was segmented using CardViews, the surface was created in FreeSurfer and subsequently flattened. On the flattened surface, the fissures and paravermian sulcus were traced. The fissures were then used in CardViews to manually subdivide the cortex into lobules and to subdivide lobules VI through X into vermis and hemisphere. Finally, lobules III through V were automatically subdivided into vermis and hemisphere, and the hemisphere was automatically subdivided into medial and lateral zones. The MRI Atlas of the Human Cerebellum (Schmahmann et al 2000) was referenced to identify landmarks and fissures in the three cardinal planes, without referring to the specific coordinate space. Individual PUs were then clustered according to putative functional considerations. It has recently been empirically demonstrated that these functional clusters can be assigned with a high degree of reliability (Makris et al 2005). Ten functional clusters were formed in total and treated as anatomic units. The first was the balance and posture cluster consisting of vermal lobules I through V shown in light green in Figure 1. The second was the autonomic/emotional cluster consisting of vermal lobules VI through X shown in light blue in Figure 1. The third and fourth were the right and left primary sensorimotor clusters including right and left medial hemispheric lobules IIIm through Vm shown in dark green in Figure 1. The fifth and sixth were the right and left secondary sensorimotor clusters of medial hemispheric zone lobules VIIIAm and VIIIBm shown in yellow in Figure 1. Seventh and eighth were the right and left oculomotor clusters of hemispheric lobules IX and X shown in dark blue in

N.J. Keuthen et al Figure 1. Finally, the ninth and tenth were the right and left cognitive clusters that were formed from hemispheric lobules VI through VIIB and the lateral hemispheric zone in lobules V and VIII as shown in red in Figure 1. To control for brain volume, we measured both head circumference (HC) and automated total brain volume (TBV). Head circumference was found by identifying the exterior border of the subcutaneous fat in an axial level that contains the glabella (the most anterior point) and the inion (the most posterior point or occipital protuberance) of the skull. Procedurally, we used our positional normalization software to identify the above landmarks and to reslice the scan to obtain the axial slice. Subsequently, the CardViews segmentation program was used to extract exterior contour. This contour was quantified by its area and perimeter (HC). For automated brain volume, we utilized the Brain Extraction Tool (Center for Functional Magnetic Resonance Imaging of the Brain, Oxford University, Oxford, England) (Smith et al 2004), part of the Oxford Centre for Functional Magnetic Resonance Imaging of the Brain (FMRIB) Software Library from Oxford University, to generate a brain and nonbrain mask from each subject’s MRI scan. The brain mask includes cerebrum, cerebellum, and brainstem. These masks were manually reviewed for accuracy. The volume of the brain was calculated by summation of all voxels contained within the brain mask, multiplied by the volume per voxel as determined by the scan acquisition. Rater Training and Reliability Our two study raters were experienced bachelors-level MRI research assistants who had basic college-level background in neuroanatomy. They were trained and supervised on an ongoing basis by our neuroanatomist (N.M.) and were blind to all subject characteristics. To assess intrarater reliability, our primary rater (J.E.S.) blindly reanalyzed scans for 10 randomly selected study participants. Intraclass correlation coefficients (ICCs) were determined from

Figure 1. Functional considerations of parcellation unit (PU) clusters. Shown are functional considerations of the cerebellum used for grouping the PUs into clusters, which are then treated as anatomical units. PU, parcellation unit.

www.sobp.org/journal

BIOL PSYCHIATRY 2007;61:374 –381 377

N.J. Keuthen et al volumetric ratings for those regions comprising the 10 functional clusters defined earlier. To assess interrater reliability, another rater (M.D.A.; see acknowledgments) independently performed parcellation on the same brains. Mean ICCs were .94 and .93 for intrarater and interrater reliability, respectively (see also Makris et al 2005). Statistical Analyses We utilized the statistical package JMP version 5.0.1.2 (SAS Institute, Carey, North Carolina) to test our study hypotheses. Two-way t tests and analyses of covariance (ANCOVAs) (with both HC and TBV as covariates) were utilized to compare groups on total CBM, total CBM cortex, and total CBM white matter volumes. A significance level of p ⬍ .05 was established. Our small number of a priori hypotheses allowed us to investigate our predictions with limited risk of type I and II errors despite our modest number of study participants. Second-tier analyses were subsequently conducted to decompose findings of group differences in TCV. Given consistently smaller volumes in the TTM versus NC cohort, we used one-way t tests and ANCOVAs with TCV as the covariate to identify group differences in CBM subterritories and functional clusters. We also performed ANCOVAs with total CBM cortical volume as the covariate to identify group differences in the functional clusters. We did not correct for the number of multiple comparisons in these analyses, given our stated intention of merely generating hypotheses for future investigation. Lastly, Pearson product moment correlations were computed between CBM volumes and MGHHPS total scale scores. We a priori predicted correlations with MGHHPS scores only for CBM cortical volumes and the autonomic/emotional and sensorimotor functional clusters. Once again, in light of the hypothesisgenerating function of these analyses, we did not perform Bonferroni corrections on our results.

Results CBM Regional Volumes Table 1 reports the raw group volumes for the total, right and left CBM, CBM cortex, and CBM white matter in addition to t test and ANCOVA results. A Priori Hypotheses. As predicted, significant differences were found for between-group comparisons of total CBM raw cortical volumes (t ⫽ 2.88, p ⫽ .008, df ⫽ 24). Separate analyses

of right (t ⫽ 2.89, p ⫽ .008, df ⫽ 24) and left (t ⫽ 2.81, p ⫽ .010, df ⫽ 24) CBM cortical volumes were both similarly significant, thus providing no evidence to suggest a lateralized betweengroup effect. Trichotillomania participants consistently had smaller CBM cortical volumes than their NC counterparts. Analyses of covariance with HC and TBV as the covariates also demonstrated significant group differences in total, right, and left CBM cortical volumes. Post Hoc Analyses. Between-group comparisons of total CBM volumes (TCV) also revealed significant group differences (t ⫽ 2.67, p ⫽ .013, df ⫽ 24), as did separate analyses for right (t ⫽ 2.68, p ⫽ .013, df ⫽ 24) and left (t ⫽ 2.63, p ⫽ .015, df ⫽ 24) total CBM raw volumes. Analyses of covariance with HC as the covariate similarly revealed significant group differences in total, right, and left CBM volumes. Analyses of covariance with TBV as the covariate resulted in probability values with marginal significance levels (p’s ⫽ .054 to .058) for total, right, and left CBM volumes. No group differences were found for total, right, or left white matter volumes. CBM Cortex Subterritory Volumes Table 2 reports raw group volumes for total, right, and left CBM subterritories in addition to t test and ANCOVA results. Significant group differences were found for total vermis (t ⫽ 1.72, p ⫽ .049, df ⫽ 24), medial hemispheric zone (t ⫽ 2.64, p ⫽ .007, df ⫽ 24), lateral hemispheric zone 1 (t ⫽ 2.06, p ⫽ .025, df ⫽ 24), and lateral hemispheric zone 2 (t ⫽ 2.19, p ⫽ .019, df ⫽ 24). Group differences between the right (t ⫽ 1.75, p ⫽ .047, df ⫽ 24) and the left (t ⫽ 1.40, p ⫽ .087, df ⫽ 24) vermis suggest a lateralized effect with TTM cohort volumes significantly smaller than NC volumes only in the right vermal subterritory. For the medial and lateral hemispheric zones 1 and 2, there was no evidence of a lateralized group effect. In all cases, subterritory volumes were smaller for the TTM versus NC cohorts. Analysis of covariance results reveal no significant group differences for any of the CBM subterritories with TCV as the covariate. CBM Functional Cluster Volumes Table 3 reports the raw volumes for the 10 functional cluster areas for both the TTM and NC groups, as well as t test and ANCOVA results examining group differences by cluster. A Priori Hypotheses. As predicted, significant group differences (t ⫽ 2.41, p ⫽ .012, df ⫽ 24) were found for the

Table 1. Volumes of Total Cerebellum

Total CBM R CBM L CBM Total CBM Cortex R CBM Cortex L CBM Cortex Total White Matter R White Matter L White Matter

NC

TTM

n ⫽ 12

n ⫽ 14

Raw Volume T

Mean

SD

Mean

SD

139.8 70.1 69.8 108.3 54.3 54.0 31.5 15.7 15.8

10.90 5.7 5.2 9.5 5.0 4.6 2.7 1.5 1.3

128.6 64.5 64.1 98.7 49.4 49.3 29.9 15.1 14.9

10.5 4.9 5.6 7.6 3.7 4.0 4.0 1.8 2.3

p

ANCOVA HCa

ANCOVA TBVa

T

T

(df ⫽ 24) 2.67 2.68 2.63 2.88 2.89 2.81 1.15 .99 1.26

p (df ⫽ 23)

.013 .013 .015 .008 .008 .010 .263 .330 .220

2.57 2.58 2.52 2.77 2.79 2.71 1.04 .91 1.14

p (df ⫽ 23)

.017 .017 .019 .011 .010 .013 .310 .373 .264

1.99 2.03 1.92 2.21 2.24 2.13 .63 .57 .68

.058 .054 .067 .037 .035 .044 .534 .575 .505

Volumes in CC. NC, normal control subjects; TTM, trichtillomania; ANCOVA, analysis of covariance; HC, head circumference; TBV, total brain volume; CBM, cerebellum; R, right; L, left. a T and p values are for the group effect while independently covarying for head circumference (HC) and total brain volume (TBV).

www.sobp.org/journal

378 BIOL PSYCHIATRY 2007;61:374 –381

N.J. Keuthen et al

Table 2. Volumes of Subdivisions of the Cerebellum Cortex NC

TTM

n ⫽ 12

n ⫽ 14

ANCOVA TCVa

Raw Volume T

Region

Mean

SD

Mean

SD

Total Vermis R Vermis L Vermis Total Medial Hemispheric Zone R Medial Hemispheric Zone L Medial Hemispheric Zone Total Lateral Hemispheric Zone 1 R Lateral Hemispheric Zone 1 L Lateral Hemispheric Zone 1 Total Lateral Hemispheric Zone 2 R Lateral Hemispheric Zone 2 L Lateral Hemispheric Zone 2

11.9 5.9 6.0 33.9 17.1 16.8 45.1 22.4 22.7 17.2 8.9 8.4

1.0 .7 .4 3.3 2.0 1.7 4.2 2.0 2.4 2.3 1.2 1.3

11.0 5.4 5.6 30.4 15.1 15.2 41.7 20.7 20.9 15.4 8.0 7.4

1.5 .8 .8 3.5 1.6 1.9 4.2 2.0 2.3 2.1 1.0 1.2

p

T

(df ⫽ 24) 1.72 1.75 1.40 2.64 2.72 2.29 2.06 2.09 1.87 2.19 2.01 2.00

p (df ⫽ 23)

.049 .047 .087 .007 .006 .016 .025 .024 .037 .019 .028 .028

⫺.32 ⫺.06 ⫺.51 .68 .92 .21 .10 ⫺.07 .20 .63 .52 .58

— — — .253 .184 .416 .459 — .420 .268 .305 .284

Volumes in CC. NC, normal control subjects; TTM, trichotillomania; ANCOVA, analysis of covariance; TCV, total cerebral volume; R, right; L, left. a T values are for the group effect while covarying for total cerebellum volume (TCV); p-values are for the alternative hypothesis of smaller volumes in the TTM group; p-values are not reported for group comparisons in which NC volumes are greater than TTM volumes.

autonomic/emotional functional cluster located in the vermis. Once again, volumes for this region were smaller for the TTM versus NC group. When volumes for this cluster were corrected for TCV, however, group differences were no longer significant. Contrary to prediction, no significant group differences were found for the primary sensorimotor or the secondary sensorimotor regions. Group differences for these sensorimotor areas remained nonsignificant after correction for TCV. Similarly, ANCOVAs with total CBM cortical volume as the covariate failed to reveal group differences for either the autonomic/emotional or sensorimotor functional clusters. Post Hoc Analyses. Group raw volume comparisons for the other functional clusters unexpectedly revealed significant differences for the right (t ⫽ 2.80, p ⫽ .005, df ⫽ 24) and left (t ⫽ 2.71, p ⫽ .006, df ⫽ 24) cognitive hemispheric clusters. Significant group differences were also found for those regions involved in the left oculomotor functional cluster (t ⫽ 2.31, p ⫽ .015, df ⫽ 24). Similar to findings for the autonomic/emotional functional cluster, the right and left cognitive cluster group differences also failed to achieve statistical significance upon correction for TCV.

The left oculomotor cluster, however, remained significantly different between study groups even after correction for TCV (t ⫽ 1.93, p ⫽ .033, df ⫽ 23). All three functional clusters demonstrating group differences in the post hoc analyses failed to meet thresholds for statistical significance when covaried for total CBM cortical volume. Correlations Between CBM Volumes and TTM Severity Ratings Table 4 reports the results of linear correlational analyses between the CBM regions and CBM functional cluster areas with total MGHHPS scores, our index of TTM severity. A Priori Hypotheses. We predicted significant correlations between TTM severity ratings and volumes for the total CBM cortex and the autonomic/emotional, primary sensorimotor, and secondary sensorimotor functional areas. The only significant correlation reported for these regions was between TTM severity and the left primary sensorimotor cluster (r ⫽ -.67, p ⫽ .008). Volumes for this functional area were inversely correlated with TTM severity ratings; thus, smaller volumes were associated with more severe TTM.

Table 3. Volumes of Ten Functional Clusters of Parcellation Units of the Cerebellum Cortex NC

TTM

n ⫽ 12

n ⫽ 14

ANCOVA TCVa

Raw Volume T

Region

Mean

SD

Mean

SD

Balance and Posture Autonomic and Emotional R Primary Sensorimotor L Primary Sensorimotor R Secondary Sensorimotor L Secondary Sensorimotor R Oculomotor L Oculomotor R Cognitive L Cognitive

6.8 5.1 2.4 2.4 3.2 2.7 3.4 3.5 39.3 39.3

.9 .3 .8 .9 .6 .7 .6 .6 3.9 4.1

6.4 4.6 2.3 2.1 2.9 2.8 3.1 3.1 35.6 35.6

1.1 .6 .6 .6 1.1 1.1 .5 .5 2.7 2.8

p

T

(df ⫽ 24) 1.03 2.41 .63 1.05 .79 ⫺.25 1.54 2.31 2.80 2.71

p (df ⫽ 23)

.156 .012 .268 .152 .220 — .068 .015 .005 .006

⫺.78 .80 ⫺.82 ⫺.28 ⫺.18 ⫺.75 1.39 1.93 .81 .67

— .217 — — — — .089 .033 .212 .255

Volumes in CC. NC, normal control subjects; TTM, trichotillomania; ANCOVA, analysis of covariance; TCV, total cerebral volume; R, right; L, left. a T values are for the group effect while covarying for total cerebellum volume (TCV); p-values are for the alternative hypothesis of smaller volumes in the TTM group; p-values are not reported for group comparisons in which NC volumes are greater than TTM volumes.

www.sobp.org/journal

BIOL PSYCHIATRY 2007;61:374 –381 379

N.J. Keuthen et al Table 4. Pearson Product Moment Correlations for Cerebellar Volumes and Massachusetts General Hospital Hair Pulling Scale Scores Region Total CBM R CBM L CBM Total CBM Cortex R CBM Cortex L CBM Cortex Total White Matter R White Matter L White Matter Total Vermis R Vermis L Vermis Total Medial Hemispheric Zone R Medial Hemispheric Zone L Medial Hemispheric Zone Total Lateral Hemispheric Zone 1 R Lateral Hemispheric Zone 1 L Lateral Hemispheric Zone 1 Total Lateral Hemispheric Zone 2 R Lateral Hemispheric Zone 2 L Lateral Hemispheric Zone 2 Functional Clusters Balance and Posture Autonomic and Emotional R Primary Sensorimotor L Primary Sensorimotor R Secondary Sensorimotor L Secondary Sensorimotor R Oculomotor L Oculomotor R Cognitive L Cognitive

r

p

⫺.24 ⫺.25 ⫺.23 ⫺.34 ⫺.35 ⫺.32 .00 ⫹.03 ⫺.02 ⫺.17 ⫺.27 ⫺.05 ⫺.22 ⫺.16 ⫺.27 ⫺.37 ⫺.38 ⫺.34 .00 ⫺.08 ⫹.06

.399 .382 .422 .236 .213 .266 .989 .918 .954 .570 .345 .857 .447 .593 .353 .193 .178 .235 .989 .798 .847

⫺.12 ⫺.20 ⫺.30 ⫺.67 ⫺.36 ⫺.07 ⫹.28 ⫹.23 ⫺.25 ⫺.29

.684 .502 .299 .008 .202 .811 .326 .422 .383 .312

CBM, cerebellum; R, right; L, left.

Post Hoc Analyses. Additional correlational analyses failed to reveal significant results for the right primary sensorimotor cluster, the left medial zone, or the other CBM regions listed in Table 4.

Discussion These morphometric MRI findings represent, to our knowledge, the first structural investigation of CBM involvement in TTM. Intuitively, it is reasonable to hypothesize TTM-related differences in CBM volumes, given our historic conceptualization of this region as the neuroanatomic locus of control for complex, coordinated motor sequences. Additionally, the early functional neuroimaging data of Swedo et al (1991) and our evolving understanding of CBM function also point to the CBM as a particular region of interest in TTM. As predicted, our results provided evidence for reduced volumes for the total, right, and left CBM cortex in TTM versus NC individuals. Given the significant motor involvement in TTM symptomatology, these results are not surprising. In all cases, TTM sufferers demonstrated reduced CBM cortex volumes. In addition, also as predicted, the TTM cohort exhibited significantly reduced volumes for the autonomic/emotional functional cluster in comparison with the NC group. The affective accompaniments to TTM are similarly noteworthy; as acknowledged in the DSM-IV TTM criteria (American Psychiatric Associ-

ation 1994), these include tension prior to pulling or when attempting to resist and pleasure, relief, or gratification upon pulling. Diefenbach et al (2002) documented affective changes prepulling to postpulling, including decreases in boredom, anxiety, and tension and increases in relief, guilt, sadness, and anger. Recognition of the role of affective features in the phenomenology of TTM has led several investigators to behaviorally conceptualize TTM as a manifestation of dysfunctional mechanisms for the control of distressing affective phenomena (Diefenbach et al 2002; Keuthen et al 2005). Importantly, however, when group differences in raw volumes for the autonomic/emotional functional cluster were corrected for TCV and total CBM cortex volume, they were no longer significant. These results may be attributable to limits in the statistical power of our analyses or, conversely, may indicate more widespread group volumetric differences in the CBM or specifically CBM cortex. Given the statistical power limitations of the current study, these findings warrant replication in larger cohorts. Our unanticipated results of significant group volumetric differences for the right and left cognitive and the left oculomotor regions merit further investigation. The findings for the cognitive functional clusters were initially somewhat puzzling, given the noted absence of cognitive phenomena in TTM. Preliminary investigations of neuropsychological functioning in TTM, however, provide evidence of impaired performance in multiple cognitive domains including executive functioning (Bohne et al 2005; Keuthen et al 1996), nonverbal memory (Keuthen et al 1996; Rettew et al 1991), and divided attention (Stanley et al 1997). These findings are consistent with studies of cerebellar function in healthy individuals and patients with cerebellar lesions. These investigations demonstrate a cerebellar contribution to planning and executive functions (e.g., Kim et al 1994; Kish et al 1988, 1994; Schmahmann and Sherman 1998), visuospatial memory (e.g., Schmahmann and Sherman 1998), and attention shifting and modulation (e.g., Akshoomoff and Courchesne 1994; Le et al 1998). As these findings were the results of post hoc analyses and no corrections were made for multiple comparisons, it should be emphasized that they require replication in future studies. The similar failure of these findings to withstand statistical corrections for TCV and total CBM cortex volumes could again be attributable to power limitations in our study or more widespread group CBM volumetric differences. The significant group difference identified in the left oculomotor region with post hoc analysis is intuitively appealing. Visual stimuli can trigger hair-pulling behavior in many sufferers and the visual system plays a prominent role in the tracking of hairs for extraction. Although the left oculomotor region is the only functional cluster to withstand correction for TCV, it also did not survive correction for total CBM cortical volume. While it is premature to draw conclusions from these results alone, future studies should further explore group volumetric differences for this cluster and ascertain the extent to which hair pullers utilize visual and tactile cues to identify targets for hair extraction. The robust correlation between left primary sensorimotor cluster volumes and TTM severity is noteworthy. Individual TTM sufferers frequently exhibit idiosyncratic motor routines focused on the identification of hairs with specific textures or colors. Sensory paresthesias antecedent to or accompanying pulling often resemble the sensory phenomena in the TS prodrome (Stein and Hollander 1992; Minichiello et al 1994) and hairpulling behavior can function to alleviate the discomfort associated with paresthesias. The cerebellar lobules that comprise the primary sensorimotor functional cluster are known to participate www.sobp.org/journal

380 BIOL PSYCHIATRY 2007;61:374 –381 in upper extremity motion (Allen et al 1997; Grafton et al 1992, 1993; Nitschke et al 1996). Although our TTM sample was exclusively right-handed, the laterality of this finding may be explained by the anecdotal observation that sufferers often tend to pull with their nondominant hand to prevent interference with routine activities. Future studies should assess dominance for hair extraction, as well as general hand dominance. Although we failed to demonstrate volumetric differences between groups for this functional cluster, our correlational analyses do reveal smaller volumes in the left primary sensorimotor cluster for those sufferers with more severe TTM. This finding suggests a relationship between illness severity and CBM sensorimotor subterritory volume. Of note, the recent work of Laforce and Doyon (2001) has elucidated different roles for the striatum and CBM in motor behavior. These researchers have documented the involvement of the CBM in the integration of isolated learned movements into smooth sequences, whereas the striatum (which our group previously studied in TTM) has been identified as the locus of control for perceptual-motor learning mechanisms. Accordingly, our present results and our earlier published findings (O’Sullivan et al 1996) in an independent cohort of subjects suggest both cerebellar and striatal involvement in TTM. Future neuroimaging investigations should include functional imaging studies employing activation paradigms to more thoroughly investigate cerebellar dysfunction in TTM and its relationship with these observed volumetric differences. In addition, future TTM studies should investigate changes in brain structure and function associated with effective treatment, including both psychopharmacological and cognitive behavioral interventions.

Financial support was provided from a grant from an anonymous benefactor to the first author. We thank Matthew D. Albaugh who assisted in interrater reliability analyses. We also gratefully recognize Mary Foley and Larry White for their technical assistance in scanning and Mariko Jameson for her assistance in manuscript preparation. Akshoomoff NA, Courchesne E (1994): Intramodality shifting attention in children with damage to the cerebellum. J Cogn Neurosci 6:388 –399. Allen G, Buxton RB, Wong EC, Courchesne E (1997): Attentional activation of the cerebellum independent of motor involvement. Science 275:1940 – 1943. American Psychiatric Association (1994): Diagnostic and Statistical Manual for DSM-IV Disorders. Washington, DC: American Psychiatric Press. Baxter LR, Schwartz JM, Guze BH, Bergman K, Szuba MP (1990): Neuroimaging in obsessive-compulsive disorder: Seeking the mediating neuroanatomy. In: Jenike MA, Baer L, Minichiello WE, editors. Obsessive Compulsive Disorder: Theory and Management, 2nd ed. Chicago: Year Book Medical Publishers, 167–188. Beck AT, Epstein N, Brown G, Steer RA (1988): An inventory for measuring clinical anxiety. The Beck Anxiety Inventory. J Consult Clin Psychol 56: 893– 897. Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J (1961): An inventory for measuring depression. Arch Gen Psychiatry 4:561–571. Bloch MH, Leckman JF, Zhu H, Peterson BS (2005): Caudate volumes in childhood predict symptom severity in adults with Tourette syndrome. Neurology 65:1253–1258. Bohne A, Keuthen NJ, Tuschen-Caffier B, Wilhelm S (2005): Cognitive inhibition in trichotillomania and obsessive-compulsive disorder. Behav Res Ther 43(7):923–942. Caviness VS Jr, Makris N, Meyer JW, Kennedy DN (1996): MRI-based parcellation of human neocortex: An anatomically specified method with estimate of reliability. J Cogn Neurosci 8:566 –588. Christenson GA, Pyle RL, Mitchell JE (1991): Estimated lifetime prevalence of trichotillomania in college students. J Clin Psychiatry 52:415– 417.

www.sobp.org/journal

N.J. Keuthen et al Dale AM, Fischl B, Sereno MI (1999): Cortical surface-based analysis. I. Segmentation and surface reconstruction. Neuroimage 9:179 –194. Diefenbach G, Mouton-Odum S, Stanley M (2002): Affective correlates of trichotillomania. Behav Res Ther 40:1305–1315. Diefenbach GJ, Tolin DF, Hannan S, Crocetto J, Worhunsky P (2005): Trichotillomania: Impact on daily functioning and quality of life. Proceedings of the 37th Annual Meeting of the Association for the Advancement of Behavior Therapy, Boston, Massachusetts. Filipek PA, Kennedy DN, Caviness VS Jr (1988): A method of morphometric analysis of the human brain based upon magnetic resonance imaging. Ann Neurol 24:356. Filipek PA, Richelme C, Kennedy DN, Caviness VS Jr (1994): The young adult human brain: An MRI-based morphometric analysis. Cereb Cortex 4:344 – 360. First MB, Spitzer RL, Gibbon M, Williams JBW (1995): Structured Clinical Interview for DSM-IV Axis I Disorders-Patient Edition. New York: New York State Psychiatric Institute, Biometrics Research Department. Grafton ST, Mazziota JC, Presty S, Friston KJ, Frackowiak RS, Phelps ME (1992): Functional anatomy of human procedural learning determined with regional cerebral blood flow and PET. J Neurosci 12:2542–2548. Grafton ST, Woods RP, Mazziota JC (1993): Within-arm somatotopy in human motor areas determined by positron emission tomography imaging of cerebral blood flow. Exp Brain Res 95:172–176. Graybiel AM, Rauch SL (2000): Toward a neurobiology of obsessive-compulsive disorder. Neuron 28:343–347. Jenike MA, Breiter HC, Baer L, Kennedy DN, Savage CR, Olivares MJ, et al (1996): Cerebral structural abnormalities in obsessive-compulsive disorder: A quantitative morphometric magnetic resonance imaging study. Arch Gen Psychiatry 53:625– 632. Kelly RM, Strick PL (2003): Cerebellar loops with motor cortex and prefrontal cortex of a nonhuman primate. J Neurosci 23:8432– 8444. Keuthen NJ, Bohne A, Himle M, Woods D, Ling BE (2005): Obsessive-Compulsive Disorder Research. Hauppauge, NY: Nova Biomedical Books, 1–29. Keuthen NJ, Dougherty DD, Franklin ME, Bohne A, Loh R, Levy J, et al (2004): Quality of life and functional impairment in individuals with trichotillomania. J Appl Res 4:186 –197. Keuthen NJ, O’Sullivan RL, Ricciardi JN, Shera D, Savage CR, Borgmann AS, et al (1995): The Massachusetts General Hospital (MGH) hairpulling scale: I. Development and factor analyses. Psychother Psychosom 64:141–145. Keuthen NJ, Savage C, O’Sullivan RL, Brown HD, Shera DM, Cyr P, et al (1996): Neuropsychological functioning in trichotillomania. Biol Psychiatry 39: 747–749. Kim SG, Ugurbil K, Strick PL (1994): Activation of a cerebellar output nucleus during cognitive processing. Science 265:949 –951. Kish SJ, El-Awar M, Schut L, Leach L, Oscar-Berman M, Freedman M (1988): Cognitive deficits in olivopontocerebellar atrophy: Implications for the cholinergic hypothesis of Alzheimer’s dementia. Ann Neurol 24:200 –206. Kish SJ, El-Awar M, Stuss D, Nobrega J, Currier R, Aita JF, et al (1994): Neuropsychological test performance in patients with dominantly inherited spinocerebellar ataxia: Relationship to ataxia severity. Neurology 44: 1738 –1746. Laforce R Jr, Doyon J (2001): Distinct contribution of the striatum and cerebellum to motor learning. Brain Cogn 45:189 –211. Le TH, Pardo JV, Hu X (1998): 4 T-fMRI study of nonspatial shifting of selective attention: Cerebellar and parietal contributions. J Neurophysiol 79:1535– 1548. Makris N, Hodge SM, Haselgrove C, Kennedy DN, Dale A, Fischl B, et al (2003): Human cerebellum: Surface-assisted cortical parcellation and volumetry with magnetic resonance imaging. J Cogn Neurosci 15:584 –599. Makris N, Schlerf JE, Hodge SM, Haselgrove C, Albaugh MD, Seidman LJ, et al (2005): MRI-based surface-assisted parcellation of human cerebellar cortex: An anatomically specified method with estimate of reliability. Neuroimage 25:1146 –1160. Minichiello WE, O’Sullivan RL, Osgood-Hynes D, Baer L (1994): Trichotillomania: Clinical aspects and treatment strategies. Harv Rev Psychiatry 1:336 – 344. Nitschke MF, Kleinschmidt A, Wessel K, Frahm J (1996): Somatotopic motor representation in the human anterior cerebellum. A high-resolution functional MRI study. Brain 119:1023–1029. Oldfield RC (1971): The assessment and analysis of handedness: The Edinburgh Inventory. Neuropsychologia 9:97–113. O’Sullivan RL, Keuthen NJ, Gumley G (1996): Trichotillomania and carpal tunnel syndrome. J Clin Psychiatry 57:174.

N.J. Keuthen et al O’Sullivan RL, Rauch SL, Breiter HC, Grachev ID, Baer L, Kennedy DN, et al (1997): Reduced basal ganglia volumes in trichotillomania measured via morphometric magnetic resonance imaging. Biol Psychiatry 42:39 – 45. Peterson B, Riddle MA, Cohen DJ, Katz LD, Smith JC, Hardin MT, et al (1993): Reduced basal ganglia volumes in Tourette’s syndrome using threedimensional reconstruction techniques from magnetic resonance images. Neurology 43:941–949. Peterson BS, Thomas P, Kane MJ, Scahill L, Zhang H, Bronen R, et al (2003): Basal ganglia volumes in patients with Gilles de la Tourette syndrome. Arch Gen Psychiatry 60:415– 424. Rauch SL, Whalen PJ, Dougherty DD, Jenike MA (1998): Neurobiological models of obsessive-compulsive disorders. In: Jenike MA, Baer L, Minichiello WE, editors. Obsessive-Compulsive Disorders: Practical Management. Boston: Mosby, 222–253. Reese TG, David TL, Weisskoff RM (1995): Automated shimming at 1.5T using echo planar image frequency maps. J Magn Reson Imaging 5:739 –745. Rettew DC, Cheslow DL, Rapoport JL, Leonard HL, Lenane MC (1991): Neuropsychological test performance in trichotillomania: A further link with obsessive-compulsive disorder. J Anxiety Disord 5:225–235. Robinson D, Wu H, Munne RA, Ashtari M, Lavir JM, Lerner G, et al (1995): Reduced caudate nucleus volume in obsessive-compulsive disorder. Arch Gen Psychiatry 52:393–398. Saxena S, Brody AL, Schwartz JM, Baxter LR (1998): Neuroimaging and frontal-subcortical circuitry in obsessive-compulsive disorder. Br J Psychiatry 35:26 –37. Schmahmann JD (1991): An emerging concept: The cerebellar contribution to higher function. Arch Neurol 48:1178 –1187. Schmahmann JD (1996): From movement to thought: Anatomic substrates of the cerebellar contribution to cognitive processing. Hum Brain Mapp 4:174 –198. Schmahmann JD, Doyon J, Toga A, Petrides M, Evans A (2000): MRI Atlas of the Human Cerebellum. San Diego: Academic Press.

BIOL PSYCHIATRY 2007;61:374 –381 381 Schmahmann JD, Pandya DN (1997): The cerebrocerebellar system. Int Rev Neurobiol 41:31– 60. Schmahmann JD, Sherman JC (1998): The cerebellar cognitive affective syndrome. Brain 121:561–579. Singer HS, Reiss AL, Brown JE, Aylward EH, Shih B, Chee E, et al (1993): Volumetric MRI changes in basal ganglia of children with Tourette’s syndrome. Neurology 43:950 –956. Smith SM, Jenkinson M, Woolrich MW, Beckmann CF, Behrens TEJ, Johansen-Berg H, et al (2004): Advances in functional and structural MR image and implementation as FSL. Neuroimage 23(suppl 1):S208 –S219. Soriano JL, O’Sullivan RL, Baer L, Phillips KA, McNally RJ, Jenike MA (1996): Trichotillomania and self-esteem: A survey of 62 female hair pullers. J Clin Psychiatry 57:77– 82. Stanley MA, Hannay HJ, Breckenridge JK (1997): The neuropsychology of trichotillomania. J Anxiety Disord 11:473– 488. Stein DJ, Hollander E (1992): Low-dose pimozide augmentation of serotonin reuptake blockers in the treatment of trichotillomania. J Clin Psychiatry 53:123–126. Stern E, Silbersweig DA, Chee KY, Holmes A, Robertson MM, Trimble M, et al (2000): A functional neuroanatomy of tics in Tourette syndrome. Arch Gen Psychiatry 57:741–748. Swedo SE, Leonard HL, Rapoport JL, Lenane MC, Goldberger EL, Cheslow DL (1989): A double-blind comparison of clomipramine and desipramine in the treatment of trichotillomania (hair pulling). N Engl J Med 321:497–501. Swedo SE, Rapoport JL, Leonard HL, Schapiro MB, Rapoport SI, Grady CL (1991): Regional cerebral glucose metabolism of women with trichotillomania. Arch Gen Psychiatry 48:828 – 833. Talairach J, Tournoux P (1988): Co-Planar Stereotaxic Atlas of the Human Brain. New York: Thieme Medical Publishers, Inc. Winchel RM, Jones JS, Molcho A, Parsons B, Stanley B, Stanley M (1992): The Psychiatric Institute Trichotillomania Scale (PITS). Psychopharmacol Bull 28:463– 476.

www.sobp.org/journal