Electrocorticography is Superior to Subthalamic Local Field Potentials for Movement Decoding in Parkinson's Disease

Smart brain implants will revolutionize neurotechnology for improving the quality of life in patients with brain disorders. The treatment of Parkinson’s disease (PD) with neural implants for deep brain stimulation (DBS) presents an avenue for developing machine-learning based individualized treatments to refine human motor control. We developed an optimized movement decoding approach to predict grip-force based on sensorimotor electrocorticography (ECoG) and subthalamic local field potentials in PD patients undergoing DBS neurosurgery. ECoG combined with Bayesian optimized extreme gradient boosted decision trees outperformed multiple state of the art machine learning approaches. We further developed a whole brain connectomics approach to predict decoding performance in invasive neurophysiology, relevant for connectomic targeting of distributed brain networks for neural decoding. PD motor impairment deteriorated decoding performance, suggestive of a role for dopamine in human movement coding capacity. Our study provides an advanced neurophysiological and computational framework to aid development of intelligent adaptive DBS.