Low-Frequency Activity in Dorsal Subthalamic Nucleus Predicts Impulsivity Improvement Following Deep Brain Stimulation in Parkinsonian Patients

ObjectiveThe progression of impulsive-compulsive behaviors (ICB) in Parkinsons Disease (PD) following subthalamic deep brain stimulation (DBS) surgery displays a large inter-patient variability. However, the link between the subthalamic neural activity at the single-neuron level and the postoperative evolution of ICB remains unclear. In this study, we investigated neural features associated with postoperative ICB recovery and their spatial distribution within the subthalamic nucleus (STN). ApproachWe examined neural activity extracted from intraoperative microelectrode recordings within the STN of 22 PD patients undergoing STN-DBS. Ten patients were diagnosed with ICB, with half of them showing recovery (ICB-R) from impulsive symptoms following implantation, while the other half remained stable (ICB-S). Both groups presented similar motor symptoms and received similar drug treatments pre- and post-operatively. Following, we compared beta [12-30 Hz] and theta [4-8 Hz] oscillations, firing rate, regularity, and spiking patterns in non-ICB, recovered, and stable patients across STN regions. We adopted linear discriminant algorithms to classify the postoperative state at both single neuron and patient levels. Main resultsWe observed significantly weaker beta and theta oscillations and increased spiking regularity at the single neuron level (p<0.05, Mann-Whitney U test) in patients who displayed postoperative ICB recovery. Of note, this difference was significant only on the dorsal portion of the STN, close to the stimulation target region. The discrimination algorithms based on these features correctly classified the postoperative state of 9/10 ICB patients. SignificanceWe showed that low-frequency subthalamic neural activity next to the stimulation target could be an effective biomarker for the evolution of ICBs following STN-DBS surgery, independently from other clinical aspects. Our results also support broader implications of beta activity in PD pathology beyond the motor domain.