Optimal deep brain stimulation connectivity to treat Holmes tremor

Holmes tremor is a particular type of tremor that results from focal brain lesions. Dual-hit lesions affecting the cerbello-thalamo-cortical as well as the nigro-striatal pathways is the current hypothesis behind the mechanism with which Holmes tremor develop. With the paradigm shift in neurology from the traditional localizationist notion to a contemporary network perspective, Holmes tremor could be seen as a network disease with the causative lesions spatially located within a specific distributed network. Deep brain stimulation (DBS) represents an effective strategy to mitigate Holmes tremor with an acceptable rate of success. The relationship between the network targeted by DBS and the one linking the causative lesions is still elusive. To this end, we sought to investigate 14 Holmes tremor cases who have been treated with DBS in different centers. We first traced lesions in native-space and later warped them into a common average brain (MNI-space). Connectivity-profiles seeding from lesions were mapped using a normative, 1000-subjects, resting-state fMRI connectome. Similarly, we localized DBS electrodes and modelled their stimulation volumes to further estimate connectivity from the latter. By subtracting connectivity-profiles seeding from stimulation volumes and lesions, contrast-maps were derived for each patient. A voxel-wise Pearson-correlation was performed to correlate connectivity strength in the contrast-maps with the corresponding tremor improvement induced by DBS. We repeated voxel-wise correlation without contrasting, using connectivity-profiles of stimulation volumes. While the latter, frequently used model in DBS-connectomic studies, failed to explain variance in improvement (R=-0.27, p=0.35), the former, contrast-based model, predicted improvement successfully in a leave-one-out cross-validation (R=0.72, p<0.01). We conclude that beneficial therapeutic connectivity in Holmes tremor overcomes lesions-associated impaired networks by targeting the residual-network accessible to the modulatory effect of DBS. If our findings will be confirmed, they would open new avenues to target therapeutic network in secondary neurological phenomena by means of different neuromodulatory technologies.