Subspace identification and decoding for movement restoration via brain-computer interfaces

Project summary

Stroke leads to debilitating paralysis, which if not systematically treated, can lead to permanent loss of motor function. In this project, state-of-the-art machine learning techniques are used to identify neural activity patterns altered by stroke-induced brain damage and use deep recurrent neural networks to decode the required motor signals to provide therapeutic and rehabilitative stimulation to restore hand function in affected individuals. Insights from the project will be utilized to optimize brain machine interfaces in stroke models, with aspirations for eventual realization in the clinic.