Resumen en español

A brain-computer interface (BCI) is a communication system that translates a brain signal (e.g. sensorimotor rhythms, evoked potentials) into a control signal and, therefore, constitutes an innovative communication alternative for people with severe motor disability (such as patients with amyotrophic lateral sclerosis). This project proposes the development of a BCI prototype based on the recording of P300 cognitive potentials using electroencephalography. The prototype uses a six-channel electroencephalograph to acquire the signals, and through a visual stimulation matrix containing the letters of the alphabet and icons associated to each one gives the user the possibility of writing words or elaborating messages with the icons. To process the signals (filtering and decimation, among other methods), the software BCI2000 and MATLAB 7.0 were used. The latter was used to program three different linear translation algorithms (linear discriminant analysis, least squares analysis and stepwise linear discriminant analysis), in order to translate the recorded evoked potentials into communication signals.