Project Summary/Abstract Dystonia is a disabling neurological condition characterized by sustained or repetitive muscle movements causing abnormal movements or postures. Studies that have investigated local field potentials (LFPs) recorded from deep brain stimulation (DBS) – direct electrical stimulation of subcortical brain structures using chronically implanted electrodes – leads in dystonia patients have been in-clinic or perioperative recordings while the patient was performing constrained movements. These studies have characterized pathologically increased low- frequency activity [i.e., theta (3-8 Hz)] and its relationship to dystonic symptoms. However, the functional relevance of this activity remains elusive and recent findings have implicated that additional oscillations (e.g., finely-tuned or narrowband gamma) may be related to pathophysiology. The objective of this application is to identify individualized LFP biomarkers of dystonia in a data-driven manner across the basal ganglia and cortex, while revealing the relationship between fluctuations in biomarkers to symptom suppression during DBS. We will accomplish this using a second-generation investigational, bidirectional device (Medtronic RC+S), which allows us to chronically sense LFPs and deliver DBS therapy while patients are in at-home settings. These novel recordings and devices provide insights into complex biomarkers and naturalistic behaviors, allowing a direct comparison of individualized biomarkers to symptom monitoring or suppression. This will advance our understanding of neural changes associated with dystonia and DBS, ultimately improving current neurostimulation therapy. Our central hypothesis is that naturalistic neural recordings, specifically those recorded both cortically and subcortically in conjunction with multimodal signal acquisition (i.e., video kinematics and acceleration), can detect personalized biomarkers of dystonic symptoms that can illuminate our network understanding of the disease and its symptom manifestation, while aiding in the optimization of DBS therapy. In Aim 1, we will identify and characterize individualized power bands associated with pathological activity in patients with dystonia using chronic subcortical and cortical recordings. We will decode pathological states using supervised and self-supervised machine learning techniques. In Aim 2, we will determine personalized power bands modulated during deep brain stimulation in patients with dystonia using chronic subcortical and cortical recordings and utilize these relationships to optimize stimulation programming. This research is significant and innovative because it will be the first chronic, multisite LFP recording paradigm to study effects of dystonic symptoms and DBS on pallidal LFPs, cortical LFPs, and subcortical-cortical interactions using the Medtronic RC+S, providing ample information about basal ganglia and cortical functions, network interactions in dystonia and the mechanisms of...