BCI Data Classification for Hand Rehabilitation

This paper presents some results of a research work that aims to use Brain - Computer Interfaces to facilitate rehabilitation training of a patient who has lost motor functions. Brain Computer Interface would be connected to the computer controlled robotic glove based on signals acquired through this interface. The subject may imagine movements that will be transmitted to the hardware components of the robotic glove. The paper is focused on the study of classification algorithms that can be optimal candidates for signals classification obtained in order to identify different types of movements. Two types of neural networks-based classifiers are used and tested in order to compare the results: Multi-Layer Perceptron (MLP) and Radial Basis Function (RBF).