Yohann Thenaisie†, Kyuhwa Lee†, Charlotte Moerman, Stefano Scafa, Andrea Gálvez, Elvira Pirondini, Morgane Burri, Jimmy Ravier, Alessandro Puiatti, Ettore Accolla, Benoit Wicki, André Zacharia, Mayte Castro Jiménez, Julien F. Bally, Grégoire Courtine†, Jocelyne Bloch†, Eduardo Martin Moraud
Abstract
Disruption of subthalamic nucleus dynamics in Parkinson’s disease leads to impairments during walking. Here,we aimed to uncover the principles through which the subthalamic nucleus encodes functional and dysfunctionalwalking in people with Parkinson’s disease. We conceived a neurorobotic platform embedding an isokineticdynamometric chair that allowed us to deconstruct key components of walking under well-controlled conditions.We exploited this platform in 18 patients with Parkinson’s disease to demonstrate that the subthalamic nucleusencodes the initiation, termination, and amplitude of leg muscle activation. We found that the same fundamentalprinciples determine the encoding of leg muscle synergies during standing and walking. We translated thisunderstanding into a machine learning framework that decoded muscle activation, walking states, locomotorvigor, and freezing of gait. These results expose key principles through which subthalamic nucleus dynamics encodewalking, opening the possibility to operate neuroprosthetic systems with these signals to improve walking inpeople with Parkinson’s disease.
Parkinson: scientists decode the neural signals that encode walking in the brain. NeuroRestore, EPFL, CHUV, UNIL
Link to the article: https://www.science.org/doi/10.1126/scitranslmed.abo1800