Combined Natural and Artificial Neural Network Learning in Prosthesis Control

Authors: Diego Liberati

DOI: 10.23880/nnoaj-16000126

Volume 3, Issue 2

Abstract

One of the most difficult tasks in neurotherapy is to restore neural functionality after nerve surgery interfacing prostheses. A great amount of hope as well as comforting results are thus the effort towards nerveprosthesis interfacing where commands to the prosthesis are not given by vicariating muscle EMG as tradition, but through the very residual nerves interfacing the prostheses via a kind of artificial neural network. Being in fact quite difficult to model and reconstruct the real transfer function from nerve spiking to prosthesis actuation, a black box solution is given by learning such embedded transfer function via trials and errors, as we do when newborns, with the aid of an interposed artificial neural network

Keywords: artificial neural network; neural functionality

View PDF