The analysis of electroneurographic and electromyographic activity recorded in the medial nerve of a transhumeral amputee during phantom finger movements.

The control of bionic prostheses for upper-limb amputees is often conducted under electromyographic (EMG) control, where enacting individual finger movements presents a significant challenge. With the development of neuroprosthetical systems based on peripheral-nerve implants, electrical activity of the afferent and efferent nerve fibers can be employed to improve decoding and control. Notwithstanding these developments, it is still poorly understood how electroneurographic (ENG) recordings could be applied for neuroprosthetic motor control and what the relevant decoding algorithms would be. Here we show that mental efforts applied to perform movements of phantom fingers of a transhumeral amputee could be represented as ENG activity and assessed with time-frequency analysis. Participant who underwent implantation in peripheral nerves for the phantom limb pain treatment underwent an experiment performing 7 gestures including movements of each phantom finger independently. Our intrafascicular recording from the medial nerve showed that ENG power increased in the frequency range from 100 to 500 Hz for the movements of phantom index, middle and ring fingers. Using EMG recorded in the areas of electrode implantation in a residual limb as a control, we concluded that the ENG activity could not be always explained as an effect of overt movements but could be related to a more sophisticated mechanism instead.