Abstract
Brain signal decoding promises significant advances in the development of clinical brain computer interfaces (BCI). In Parkinson’s disease (PD), first bidirectional BCI implants for adaptive deep brain stimulation (DBS) are now available. Brain signal decoding can extend the clinical utility of adaptive DBS but the impact of neural source, computational methods and PD pathophysiology on decoding performance are unknown. This represents an unmet need for the development of future neurotechnology. To address this, we developed an invasive brain-signal decoding approach based on intraoperative sensorimotor electrocorticography (ECoG) and subthalamic LFP to predict grip-force, a representative movement decoding application, in 11 PD patients undergoing DBS. We demonstrate that ECoG is superior to subthalamic LFP for accurate grip-force decoding. Gradient boosted decision trees (XGBOOST) outperformed other model architectures. ECoG based decoding performance negatively correlated with motor impairment, which could be attributed to subthalamic beta bursts in the motor preparation and movement period. This highlights the impact of PD pathophysiology on the neural capacity to encode movement kinematics. Finally, we developed a connectomic analysis that could predict grip-force decoding performance of individual ECoG channels across patients by using their connectomic fingerprints. Our study provides a neurophysiological and computational framework for invasive brain signal decoding to aid the development of an individualized precision-medicine approach to intelligent adaptive DBS.
Significance Statement Neurotechnology will revolutionize the treatment of neurological and psychiatric patients, promising novel treatment avenues for previously intractable brain disorders. However, optimal surgical and computational approaches and their interactions with neurological disorders are unknown. How can recent advances in machine learning and connectomics aid the precision and performance of invasive brain signal decoding strategies? Do the brain disorders treated with such approaches have impact on decoding performance? We propose a real time compatible advanced machine learning pipeline for invasively recorded brain signals in Parkinson’s disease (PD) patients. We report optimal movement decoding strategies with respect to signal source, model architecture and connectomic fingerprint and demonstrate that PD pathophysiology significantly and negatively impacts movement decoding. Our study has broad impacts for the development of smart brain implants for the treatment of PD and other brain disorders.
Competing Interest Statement
The authors have declared no competing interest.
Footnotes
Email: 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
Robert Mark Richardson: Mark.Richardson{at}mgh.harvard.edu
Wolf-Julian Neumann: julian.neumann{at}charite.de
Updated Analysis of Relation of beta bursting to performances and UPDRS symptom severity
https://github.com/neuromodulation/icn/tree/master/ECOG_vs_STN