Investigating Basal Ganglia-Cortical Physiology During Sleep and Its Modulation by Adaptive Deep Brain Stimulation in Parkinson Disease

Parkinson disease (PD) is the second most common age-related neurodegenerative disease worldwide, and is the fastest growing neurologic disorder in prevalence, disability, and deaths. Sleep disturbances have a significant negative impact on quality of life for patients with PD and their caregivers, and may be correlated with faster disease progression. Poor sleep is a known risk factor for cognitive impairment in PD. PD patients spend less time in non-rapid eye movement (NREM) stage 3 (N3) and REM sleep, both of which are critical for memory consolidation and integration with existing knowledge. Reduction in N3 sleep and slow waves (SWs, a hallmark of N3) is associated with poorer cognitive performance and a higher likelihood of cognitive impairment. Targeted and highly efficacious therapies for sleep disorders in PD are lacking and therefore remain a crucial unmet need for the care of people with PD. Further studies are warranted to better understand the pathophysiologic mechanisms of sleep disorders in PD and aid in the development of improved treatments. Basal ganglia-thalamic-cortical circuits, as hubs for cortical-subcortical transmissions, may contribute to or be modulated by aberrant sleep circuits and thus provide critical insight into these complex disorders. Prior work by our group has demonstrated significant band power differences in the subthalamic nucleus (STN) across the sleep-wake cycle, which corroborate cortical patterns of sleep and thus provide support for the basal ganglia as important node for sleep-wake cycling. Patients treated with deep brain stimulation (DBS) represent a unique opportunity to sample relevant basal ganglia electrophysiology. Furthermore, treatment with DBS often leads to an improvement in both subjective and objective sleep measures, but the physiological mechanisms by which this may occur remain unknown. Recently FDA-approved adaptive DBS (aDBS) technology allows for real-time, closed-loop stimulation adjustment based on LFP activity. aDBS therapy provides at least comparable control of motor symptoms during wakefulness compared to cDBS, and is preferred by patients over cDBS for improvement in motor symptoms or reduction of adverse effects of stimulation. aDBS can operate in two modes: single-threshold (ST-aDBS) or dual-threshold (DT-aDBS). DT-aDBS is used more commonly by clinicians, and is preferred over ST-aDBS by patients, and so will be the focus of the present study. The dual threshold algorithm directs the neurostimulator to increase stimulation amplitude when the power of the LFP is above the upper threshold, and to decrease stimulation amplitude when the power of the LFP drops below the lower threshold, but to hold the stimulation amplitude when the power of the LFP remains between the thresholds. We hypothesize that wake-defined aDBS thresholds result in reduced adaptive engagement during sleep due to lower beta power, and that recalibration will restore dynamic threshold crossings.

Study Research Question How can adaptive deep brain stimulation (aDBS) therapy be optimized to address sleep dysfunction in people with advanced Parkinson disease already treated with DBS? Primary objective - Aim 1 To characterize the effect of aDBS on neurophysiological biomarkers in sleep and sleep physiology.

Primary outcome measure - Aim 1 Number of LFP threshold crossings during sleep (sleep stages N2 and N3) per hour.

Secondary outcome measure - Aim 1 Percentage of sleep time (N2, N3) that LFP power is above the upper threshold, below the lower threshold, and between thresholds. These metrics will be analyzed using linear mixed-effects models with a fixed effect for sleep stage and participant/hemisphere as a random intercept.

Exploratory outcome measures - Aim 1

• Average beta LFP power per 30-second epoch
• Beta LFP power in the 30 seconds preceding each scored arousal compared to matched control epochs of stable N2/N3 sleep
• Generalized linear mixed effects models with a logit link function and a random intercept will be used assess whether elevated beta power predicts the probability of transition from N2/N3 to wake
• Linear regression will be used to test whether baseline subjective sleep, quality of life, and disease severity measures predict aDBS and sleep outcomes (e.g., number of threshold crossings, N3 duration, arousal index).

Primary objective - Aim 2 To personalize and test sleep-tailored aDBS thresholds to restore adaptive engagement during sleep Primary outcome measure - Aim 2 Change in number of threshold crossings per hour of N2 and N3 sleep, comparing baseline (Aim 1) vs. sleep-tailored aDBS (Aim 2).

Secondary outcome measures - Aim 2 Change percentage of sleep time (N2, N3) that LFP power is between thresholds, below the lower threshold, and above the upper threshold, comparing baseline sleep study vs. sleep-tailored aDBS.

Exploratory outcome measures - Aim 2 Change in sleep architecture outcomes (N3 proportion, N2 proportion, REM proportion, sleep efficiency, total sleep time, arousal index) during lab-based PSG, comparing sleep-tailored aDBS (Aim 2) to the baseline daytime-optimized aDBS (Aim 1).

This is an interventional study of a cohort of patients with PD treated with DBS, and already utilizing adaptive DBS (dual threshold mode). In Aim 1, we will obtain baseline sleep and basal ganglia physiology data to inform aDBS parameter changes. Aim 2 involves 15 nights of in-home monitoring of subjective and objective sleep quality, during which participants will use 3 different sleep-tailored aDBS settings at night only. Participant preference among these 3 settings will determined the 'optimized sleep-tailored aDBS mode'. Participants will undergo repeat polysomnography, while in the optimized sleep-tailored aDBS mode, to compare aDBS behavior and sleep architecture to the baseline (Aim 1) sleep study.

Population: We will enroll 18 participants with PD treated with DBS from the Movement Disorders Center at the University of Colorado. Participants must have a Medtronic DBS system, including the Percept RC implanted pulse generator with SenSight model DBS leads (which are optimized for sensing and recording of LFP data), and be using aDBS therapy.