A novel approach to decode hand motor tasks from intraneural recordings in an amputee

In the field of neuroprosthesis, intraneural interfaces represent an interesting solution to restore lost functions in amputees. In particular, recent results have shown that transverse intrafascicular multichannel electrodes (TIMEs) can re-establish natural and sophisticated sensory feedback. However, the possibility to decode motor intentions in humans has not been explored yet. In this study, we developed a framework able to decode several hand movements intention from neural signals recorded from intrafascicular electrodes in an upper limb amputee. Four intraneural multichannel electrodes (TIMEs) were implanted in the median and ulnar nerve of the amputee, and the electroneurographic (ENG) activity was recorded during hand motor tasks imagined by the subject. We built a stable decoder, reliable in all the experimental sessions and able to decode four classes of movements plus rest with about 80% accuracy. This approach could improve the efficacy of sensitive hand prosthesis resulting in a more natural and dexterous device relying on a single device for both sides of the closed-loop.