ABSTRACT

A prosthetic keyboard for a Brain Computer Interface (BCI) is a device that uses patterns of the electroencephalographic (EEG) signal evoked by specific visual stimuli to recognize symbols or commands. When the user looks to a stimulus, an associated pattern on the EEG signal will be evoked. The objective of this project is to develop a system for signal conditioning and acquisition, as well as for further identification of different patterns. The system uses a LED matrix as the source of visual stimuli and an appropriate data acquisition system, which were used with ten volunteers in different tasks. Acquired data was analyzed offline in both, time and frequency domains with the mean FFTs and spectrogram, using the MATLAB platform. In eight out of ten volunteers the system succeeded to identify relevant patterns. The analysis of the spectrogram showed that it is possible to identify two, four and up to eight different patterns. These results encourage further development towards a functional prosthetic keyboard.