Vigilance-dependent Modulation of Cortical Evoked Potentials Elicited by Dentate Nucleus Stimulation in Post-stroke Patients

Introduction: The deep cerebellar nuclei are increasingly being explored as potential targets for neuromodulation-based therapies for neurological and psychiatric indications due to their robust, widespread connections with cerebral cortical circuits (Dum & Strick, 2003). We use low-frequency DBS-cortical evoked potentials (CEPs) to characterize the effect of different contact locations and pulse parameters on cerebral cortical activity as part of an effort to improve therapeutic programming. Here, we report how fluctuations in participant state, or vigilance, can confound these measures, altering CEP component characteristics. Beyond that, our data provide insight into how participant state may alter cortical activation associated with our therapeutic approach.

Methods: As part of our phase I/II trials to enhance motor rehabilitation for individuals with chronic post-stroke motor deficits, participants undergo unilateral deep brain stimulation (DBS) lead implant targeting the dentate nucleus (DN) contralateral to their stroke-affected cerebral hemisphere (Baker et al, 2023). Scalp EEG data were recorded continuously as low-frequency ( < 6Hz) DBS was delivered using the implanted lead. Participants were seated comfortably in a reclined chair and asked to alternately open and close their eyes, with eyes-closed states extended for up to 20 minutes to encourage vigilance changes. EEG data were subsequently divided into 30-second epochs, and sleep staged (Awake, N1, N2, N3, REM) using USleep v1.0 (Perslev, M., Darkner, S., Kempfner, L. et al, 2021). Sleep stage-specific DBS CEPs were created by segregating pulse timing-locked EEG segments by stage and averaging. Comparisons between conditions were made using changes in peak-to-peak amplitude, latency, RMS power, and wavelets.

Results: CEP morphology was found to vary significantly across sleep stages, marked by amplitude attenuation and latency delays that increased with decreasing vigilance. Changes in the spectral content of the CEP were also observed, predominantly involving higher frequency activity. Specifically, we observed a shift from more prevalent gamma band activity in the high vigilance state towards low and high-beta band activity as vigilance reduced.

Conclusion: The observed drop in cerebellocortical modulatory effect likely reflects thalamic gating or increased cortical/hippocampal inhibitory feedback in the presence of sleep-related, slow-wave oscillations (Torres-Herraez et al, 2022). These results suggest that future closed-loop therapies may wish to consider how changes in vigilance impact real-time therapy delivery, and if DBS-CEPs are to be used for programming, it is imperative to ensure that the data are not misinterpreted due to unrecognized fluctuations in vigilance.