Many centers perform Deep Brain Stimulation (DBS) surgery under general anesthesia (GA), known as asleep DBS. Local field potential (LFP) of the Subthalamic Nucleus (STN) recorded in awake Parkinson's disease (PD) patients revealed important insights into disease mechanism and DBS optimization-strategies. In contrast, the spectral characteristics of oscillations recorded in the GA-induced unconscious state remain only partially understood.

To contrast the spectral and topographical characteristics of STN-LFPs recorded in both awake and asleep states and assess the clinical DBS response prediction based on neurophysiological hotspot positions.

STN-LFPs were recorded intraoperatively from 69 PD patients (128 hemispheres) awake and 26 patients (51 hemispheres) under propofol-anesthesia using multi-contact DBS electrodes. Spectral power (4 to 400 Hz), topographical hotspot distributions and their clinical predictive values were compared. The relationship between LFPs and frontal-EEG, anesthetic depth and dopamine withdrawal were also evaluated.

Asleep LFPs showed increased alpha (8-12 Hz), low-beta (13-20 Hz), and fast-gamma (110-140 Hz) activity, and decreased theta (4-7 Hz), high-beta (21–30 Hz), and low-gamma (35-45 Hz) power, while high-gamma (60-90 Hz), slow-HFO (205-295 Hz) and fast-HFO (305-495 Hz) activity remained unchanged compared to the awake state. Under asleep DBS the spectral topographical map shifted medially, posteriorly and inferiorly, hereby losing its clinical predictive value. STN-LFPs echo propofol-induced changes in frontal-EEG, while time of dopamine withdrawal did not impact asleep-LFP.

Unconsciousness reshapes the spectral and spatial topography of the STN in PD patients, hereby losing its predictive values for motor DBS-response. Dynamical changes of spectral features in space may inform future sleep-tailored DBS.