Summary
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 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 and subthalamic local field potentials in PD. We demonstrate that electrocorticography combined with Bayesian optimized extreme gradient boosted decision trees outperform other machine learning approaches. We elucidate a link between dopamine and movement coding capacity in PD, by showing negative correlations between decoding performance and motor symptom severity in the medication OFF state. Finally, we introduce an approach that leverages whole-brain connectomics to predict machine-learning based decoding performance in invasive neurophysiology. Our study provides a framework to aid development of intelligent adaptive deep brain stimulation.
Competing Interest Statement
The authors have declared no competing interest.
Footnotes
Timon Merk: timon.merk{at}charite.de, Victoria Peterson: vpeterson2{at}mgh.harvard.edu, Witold Lipski: lipskiw{at}upmc.edu, Benjamin Blankertz: benjamin.blankertz{at}tu-berlin.de, Robert Sterling Turner: rturner{at}pitt.edu, Ningfei Li: Ningfei.li{at}charite.de, Andreas Horn: andreas.horn{at}charite.de, Andrea Kühn: andrea.kuehn{at}charite.de, Robert Mark Richardson: Mark.Richardson{at}mgh.harvard.edu, Wolf-Julian Neumann: julian.neumann{at}charite.de