ENG-Based Signal Processing and Neural Decoding Algorithms for Rehabilitative Prostheses Control

Techniques for extracting body motion intent from neural signals represent an important breakthrough in the field of rehabilitative prostheses. In concert with recent developments in microelectrode fabrication and multichannel neural signal recoding, cortical and peripheral neural interface implantation techniques may pose significant rehabilitation potential and have attracted increased attention. Neural signals from cerebral cortex or peripheral nerves may be used to control external devices such as a cursor or a mechanical arm. The procedure of processing neural activity recordings is summarized by three main steps: denoising, spike sorting and neural decoding. In this paper, we conduct a review of existing algorithms applied to the three steps.