Voxel-based morphometry and task functional magnetic resonance imaging in essential tremor: evidence for a disrupted brain network

Abstract

The pathophysiology of essential tremor (ET) is controversial and might be further elucidated by advanced neuroimaging. Focusing on homogenous ET patients diagnosed according to the 2018 consensus criteria, this study aimed to: (1) investigate whether task functional MRI (fMRI) can identify networks of activated and deactivated brain areas, (2) characterize morphometric and functional modulations, relative to healthy controls (HC). Ten ET patients and ten HC underwent fMRI while performing two motor tasks with their upper limb: (1) maintaining a posture (both groups); (2) simulating tremor (HC only). Activations/deactivations were obtained from General Linear Model and compared across groups/tasks. Voxel-based morphometry and linear regressions between clinical and fMRI data were also performed. Few cerebellar clusters of gray matter loss were found in ET. Conversely, widespread fMRI alterations were shown. Tremor in ET (task 1) was associated with extensive deactivations mainly involving the cerebellum, sensory-motor cortex, and basal ganglia compared to both tasks in HC, and was negatively correlated with clinical tremor scales. Homogeneous ET patients demonstrated deactivation patterns during tasks triggering tremor, encompassing a network of cortical and subcortical regions. Our results point towards a marked cerebellar involvement in ET pathophysiology and the presence of an impaired cerebello-thalamo-cortical tremor network.

Figure 1

Schematic of the cerebellum. Representation of the major anatomical subdivisions of the cerebellum according to FSL probabilistic cerebellar atlas (non-linearly registered to MNI space).

Figure 2

Activation and deactivation maps. Fixed-effects group analysis results for task 1 (patients with essential tremor [ET] and healthy controls [HC]) and task 2 (HC only). Statistical maps are thresholded by using clusters determined by Z > 6 (activations) and Z < − 6 (deactivations) with a (corrected) cluster significance threshold of p < 0.05.

Figure 3

Statistical comparisons for activations and deactivations. Brain areas statistically different across groups (patients with essential tremor [ET] and healthy controls [HC]) and tasks are illustrated in figure and color-coded by statistical significance. Activations and deactivations were compared separately in the statistical analyses, and significant clusters were determined by Z > 4 (activations) and Z < − 4 (deactivations) with a (corrected) cluster significance threshold of p < 0.05. Top: between-group comparison (task 1 ET vs HC). Bottom: between-group between-task comparison (task 1 ET vs task 2 HC).

Figure 4

Linear regression analysis. Regression results for two representative regions of interest are reported, one resulting as activated in the general linear model analysis for task 1 in patients with essential tremor (right pre/postcentral gyrus [PRG.r/POG.r]) and one as deactivated (left posterior cingulate gyrus [CGp.l]). Individual BOLD results, expressed as z-statistic values, were linearly regressed against the Fahn–Tolosa–Marin Tremor Rating Scale (TRS), part A + B. The correlation value (r-value) and the corresponding false-discovery rate (FDR)-corrected p-value (p_FDR) are also reported.