Voxel-Based Morphometry of Cerebellar Lobules in Essential Tremor

Background The extent of neurodegeneration underlying essential tremor (ET) remains a matter of debate. Despite various extents of cerebellar atrophy on structural magnetic resonance imaging (MRI), previous studies have shown substantial heterogeneity and included a limited number of patients. Novel automated pipelines allow detailed segmentation of cerebellar lobules based on structural MRI. Objective To compare the volumes of cerebellar lobules in ET patients with those in healthy controls (HCs) using an automated segmentation pipeline. Methods Structural MRI scans of ET patients eligible for deep brain stimulation (n = 55) and of age-matched and gender-matched HCs (n = 55, from the IXI database) were segmented using the automated CEREbellum Segmentation pipeline. Lobule-specific volume differences between the ET and HC groups were evaluated using a general linear model corrected for multiple tests. Results Total brain tissue volumes did not differ between the ET and HC groups. ET patients demonstrated reduced volumes of lobules I-II, left Crus II, left VIIB, and an increased volume of right X when compared with the HC group. Conclusion A large cohort of ET patients demonstrated subtle signs of decreased cerebellar lobule volumes. These findings oppose the hypothesis of localized atrophy in cerebellar motor areas in ET, but not the possibility of cerebellar pathophysiology in ET. Prospective investigations using alternative neuroimaging modalities may further elucidate the pathophysiology of ET and provide insights into diagnostic and therapeutic approaches.


INTRODUCTION
Essential tremor (ET) is the most prevalent movement disorder, affecting approximately 4% of the population over 40 years of age (Zesiewicz et al., 2010). Although it was initially believed to be a benign motor disorder, reports have demonstrated increased occurrences of mood disorders, cognitive disorders, and early mortality in ET patients (Benito-Leon et al., 2006;Louis et al., 2007a,b). A neurodegenerative origin of ET has been proposed based on neuropathological analyses of ET patients showing cerebellar pathology (Rajput et al., 2012;Benito-León, 2014). Mechanistically, a cerebello-thalamocortical network dysfunction has been suggested to cause tremor in ET (Hallett, 2014;Buijink et al., 2015) and lesions or deep brain stimulation (DBS) in this pathway may suppress tremor (Dupuis et al., 2010;Fytagoridis et al., 2012;Elias et al., 2016;Barbe et al., 2018).
Structural magnetic resonance imaging (MRI), which allows for voxel-based morphometric (VBM) analysis, has demonstrated diverse cortical and cerebellar volumetric changes in ET patients. Decreased cerebellar gray matter has been observed in groups of ET patients (Benito-Leon et al., 2009;Bhalsing et al., 2014;Gallea et al., 2015;Dyke et al., 2017). However, other studies have reported no differences (Daniels et al., 2006;Fang et al., 2013;Nicoletti et al., 2015) and abnormalities in white matter (Pietracupa et al., 2019). A recent meta-analysis of 16 structural MRI investigations proposed that previously reported gray matter abnormalities in ET might be related to methodological heterogeneity and small or diverse cohorts, thereby indicating the need for larger studies using standardized protocols (Luo et al., 2019).
In the present study, cerebellar volumes of 55 ET patients eligible for DBS (Fytagoridis et al., 2012) were compared with those of matched healthy controls (HCs) from the IXI repository. We aimed to test the hypothesis of regional atrophy in the cerebellar motor areas of ET patients using an automated segmentation pipeline (Romero et al., 2017).

Patient and Control Selection
Patients diagnosed with ET (n = 70) who underwent T1weighted MRI before DBS between 2006 and 2015 at Umeå University Hospital, Sweden were included. The exclusion criteria were the use of alternative MRI protocols (n = 13), previous DBS electrode implantation (n = 1), or major brain lesions (n = 1). T1-weighted MRI scans of HCs (n = 55) were retrieved from the IXI repository, which included volunteers from three UK hospitals (https://braindevelopment.org/ixi-dataset/). Subject characteristics (age and sex) and MRI data were collected. To reduce the technical Data are shown as means (SD). Groups were compared using t-tests.
bias introduced by different scanners and sites, volumes were normalized to total brain tissue, which did not differ significantly between the HC and ET groups ( Table 1 and Supplementary Figure 1).

Volumetric Analysis
T1-weighted MRI images in DICOM format were anonymized and converted to the NIFTI format using dcm2niix (Li et al., 2016). Cerebellar lobule volumes were determined using the CEREbellum Segmentation (CERES) pipeline, which is a patchbased multi-atlas segmentation pipeline that parcellates the cerebellum into 12 lobules and determines the respective volumes in the native space (Romero et al., 2017). CERES has been demonstrated to outperform several cerebellar pipelines in terms of segmentation accuracy (Romero et al., 2017). Total lobule volumes including both white and gray matter were extracted from the results. The anatomical segmentations in CERES are shown in Figure 1. Total brain volumes were determined by analyzing the T1-weighted sequences using the volBrain pipeline (Manjón and Coupé, 2016).

Statistical Analysis
Baseline characteristics (age and brain volume) were compared using Student's t-test. Cerebellar regions were compared between the ET and HC groups using a multivariate general linear model with age and total brain volume as covariates. The significance level was adjusted for multiple testing using the Benjamini-Hochberg method (Benjamini and Hochberg, 1995

RESULTS
The ET patients were matched with HCs from the IXI repository based on age and sex ( Table 1). The images were matched based on the MRI scanner type and magnetic field strength. Total brain volumes determined using volBrain (Manjón and Coupé, 2016) did not differ significantly between the HC and ET groups (p = 0.2403, Table 1 and Supplementary Figure 1). Cerebellar lobules were segmented and volumetrically quantified using CERES (Romero et al., 2017) and analyzed using a multivariate general linear model with gender, age, and total brain volume as  Table 2). The boxplots show the 25th, 50th, and 75th percentiles for all data. A multivariate general linear model was applied to the cerebellum and the 12 lobules (right and left sides) with gender, age, and total brain volume as covariates. The significance level (p < 0.05) was adjusted for multiple testing using the Benjamini-Hochberg method. Significance is denoted by *. R: right, L: left, Cr I: Crus I, Cr II: Crus II.
covariates. Total cerebellar volumes did not differ (right: p = 0.874 and left: p = 0.995, Figure 2A) between the groups, while volumes of lobules I-II (right: p < 0.001 and left: p < 0.001), left Crus II (p < 0.001), and left VIIB (p < 0.001) were lower in the ET group than in the HC group (Figures 2B,C and Table 2). The right lobule X was larger in the ET group (p < 0.001).

DISCUSSION
Structural MRI-derived cerebellar volumes of a large sample of ET patients eligible for DBS were compared with those of matched HCs. Among ET patients, subtle volume reductions were observed in bilateral lobules I-II, left Crus II, and left VIIB. The volume of the right lobule X was increased.
The observations from the present study are partly concordant with those of previous studies. A lobule-based analysis demonstrated reduced gray matter volumes of VI and VIIA in a pooled cohort of 39 patients with cerebellar and classic ET (Shin et al., 2016). Another study used the spatially unbiased infratentorial segmentation pipeline and reported reduced gray matter densities in the left lobules I-IV, V, VIIB, VIIIA, IX, and X; right lobules V and IX; and in vermal regions in a subgroup of ET A multivariate general linear model was applied to the cerebellum and the 12 lobules (right and left sides) with gender, age, and total brain volume as covariates.
The significance level p < 0.05 was adjusted for multiple testing using the Benjamini-Hochberg method. Significance denoted by * (bold). R: right and L: left.
patients with head or jaw tremor (Dyke et al., 2017). Laterality of volumetric changes has been reported in cerebellar gray matter volume (Quattrone et al., 2008). The reduced volumes of Crus II and VIIB in the left cerebellum may be related to technical or biological aspects, which varied between the ET and HC cohorts. Handedness is an example of the latter, although the implications on cerebellar gray matter volumes remain questioned (Ocklenburg et al., 2016). Interestingly, our results did not show pronounced volumetric reductions within the first (I-VI) and second (VIIIA-VIIIB) cerebellar motor representations (Buckner et al., 2011;Guell et al., 2018) among ET patients. Instead, the findings are consistent with a recent meta-analysis that challenged the hypothesis of distinct structural cerebellar atrophy patterns in ET and indicated that study size and heterogeneity might explain the diverse VBM findings reported previously (Luo et al., 2019). An alternative explanation for the reduced volumes in the present investigation might be the difference in MRI sites and scanner protocols, which influence the absolute volumetric measurements (Shinohara et al., 2017).
Histopathological examinations of ET patients demonstrated features suggesting neurodegeneration such as altered axon morphology and orientation of Purkinje cell bodies, reduced cell densities, and loss of dendritic spines (Louis, 2016). However, such changes may not necessarily cause volumetric differences observable on structural imaging. A recent resting-state functional MRI (fMRI) investigation indicated reduced cerebellar connectivity in ET, characterized by eigenvector centrality (Mueller et al., 2017). Additionally, evidence regarding the cerebellar involvement in ET has been obtained from fMRI investigations of ET patients wherein DBS affected cerebello-cerebral networks in task-dependent as well as task-independent manners (Awad et al., 2019). In addition, diffusion tensor imaging of patients with ET and Parkinson's disease demonstrated increased radial diffusivity in the superior, medial, and inferior cerebellar peduncles in ET, suggesting white matter involvement (Juttukonda et al., 2018).
The present study is limited by the use of MRI data from different sources, retrieved over a long period (2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015). However, similar comparisons using MRI scans from the IXI repository have previously been applied to case series (Corral-Juan et al., 2018). To reduce the technical interference introduced by different T1-weighted MRI parameters, volumes were normalized to total brain tissue, which did not differ significantly between the HC and ET groups. Although the CERES and volBrain pipelines are fully automated, differential segmentations based on MRI scan parameters and image quality may affect the results (Manjón and Coupé, 2016;Romero et al., 2017). The present study did not perform corrections for disease severity, ET subtype, hand dominance, and relevant comorbidities such as vascular disease and alcohol consumption. Patient selection was based on ET severity and only the patients eligible for DBS were included, although some previous studies have supported (Bagepally et al., 2012) and others have opposed (Quattrone et al., 2008) a relationship between disease severity and structural gray matter changes.

CONCLUSION
A cohort of 55 patients with ET demonstrated subtle signs of a decrease in cerebellar lobule volumes. These findings do not support the hypothesis of localized atrophy in the cerebellar motor areas in ET. Prospective investigations using structural or functional neuroimaging are needed to confirm these findings and to further elucidate the pathophysiology of ET to provide insights into diagnostic and therapeutic approaches.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation, to any qualified researcher.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by the Umeå Regional Ethical Review Board, Umeå, Sweden. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

AUTHOR CONTRIBUTIONS
RÅ, AA, and AF conceived and planned the study. PB provided the data. RÅ and AA performed the data analysis. PB and AF interpreted the data. RÅ prepared the manuscript with support from AA, PB, and AF. AF supervised the study. All authors approved the final manuscript.