Implementation of a real time decoder for real neuroprosthetic applications

There is substantial evidence that it is possible to predict movement intentions from single cells recording in monkeys. Such predictions, using decoding algorithms, have a large potential for clinical applications for paralyzed patients or amputees. In spite of these advances, it is however not clear how accurate and practical the movements obtained from real neuronal devices could be. In this work, an original single neurons decoding method to estimate the movement intention is proposed and studied in realistic simulations. The method provides an accurate estimation of the intention of movement location as a function of the number of neurons for its application in Brain Machine Interface.