Diseño e Implementación de un Sistema de Clasificación de Tareas Mentales a través de Redes Neuronales Artificiales

Abstract

A neuronal classifier for the recognition of mental tasks is a very important part in a Brain Computer Interface System (BCI) due to the fact that this part of the system have to be capable of interpreting the mental tasks that the subject is executing in a certain moment. For this reason arises the idea of looking for the most successful classification system for the interpretation of the mental tasks. In this article we show the implementation of two types of classifiers of cerebral activities across artificial neural network (Backpropagation and Radial base). The classifiers were developed under Matlab’s environment, using the electroencefalogram (EEG) of three different subjects. Also we show the rate of success classification for every net used in the different experiments.