Timing- and frequency-specific effects of dentate nucleus deep brain stimulation on somatosensory-evoked potentials in people with poststroke hemiparesis.

Abstract

Deep brain stimulation (DBS) of the dentate nucleus (DN) is being investigated as a therapy to enhance perilesional cortical excitability and promote motor recovery for individuals with chronic, poststroke motor deficits. Given that acute motor changes are not anticipated, DBS optimization would benefit from surrogate markers of stimulation's effect on cortical excitability. Here, we evaluate whether continuous and paired DN stimulation modulates somatosensory-evoked potentials (SSEPs), providing first in-human insight into their candidacy as a tool for device programming. SSEPs were collected from participants in a phase I DN DBS clinical trial to characterize the effects of continuous and paired stimulation on SSEP response characteristics. Continuous low-frequency DBS did not yield significant changes in short-latency peak-to-peak amplitude, though high-frequency stimulation yielded significantly lower peak-to-peak amplitude during double, but not single, pulse SSEP (64% of baseline, P < 0.05). As interstimulus interval (ISI) between SSEP and DBS was increased, short-latency power decreased (P < 0.005), with greatest power at an ISI of 0 ms (156% of baseline, P < 0.05). Our results support involvement of DN output in both early and late SSEP components. Modulation was modest and variable across subjects, limiting its potential role in therapeutic programming. Further work is required to elucidate the effects of lesion size and DBS lead placement.NEW & NOTEWORTHY We assessed SSEPs, a common clinical index of cortical excitability, as a candidate biomarker to optimize the programming of cerebellar neuromodulation devices for stroke recovery. Collected as part of a phase I clinical trial of deep brain stimulation for stroke, this work provides first in-human evidence that cerebellar stimulation acutely modulates both early and late stages of cortical sensory processing. We show that SSEPs, therefore, may be of future use for the programming of cerebellar neuromodulation devices.