An Investigation of Heterogeneity in Cerebellar Cortical Pathology Across Essential Tremor Cases: A Cluster Analysis

Recent years have witnessed an evolution of our understanding of the underlying neuropathological features of essential tremor (ET). Studies have revealed a cascade of neurodegenerative changes in the cerebellar cortex, centered on the Purkinje cell (PC) population and its surrounding neuronal populations and connections. A fundamental question is whether there is heterogeneity across ET cases with respect to the expression of this cerebellar pathology. We examined postmortem tissue from 155 ET brains using a set of 15 quantitative metrics of cerebellar cortical pathology determined by histologic and immunohistochemical methods. These metrics reflect changes across the PC body (PC counts, empty baskets, heterotopias), PC dendrites (swellings), PC axon (torpedoes), basket cell axonal hypertrophy and climbing fiber-PC dendrite synaptic changes. Principal component analysis (PCA) was used to identify the major data dimensions. Outlier cases were then identified, followed by k-means clustering to identify subgroups of ET cases. The first 5 principal components explained 78.6% of the variance. Twenty-two (14.2%) cases were outliers. k-means clustering indicated that these 22 cases were distributed in 4 clusters. In each cluster, specific quantitative metrics of cerebellar pathology showed greater severity of change than observed in the main group of 133 ET cases. A small number of clinical features distinguished clusters from one another, although these analyses were limited by sample size. Cerebellar cortical pathology in ET was not homogeneously expressed across every case. It was possible to identify clusters of cases that differed with respect to the expression of postmortem features.