EEG-based Brain Computer Interface

The Graz-BCI transforms EEG signals into control signals, whereby the signals from the brain must be modified by pure mental activity (thoughts). The Graz BCI prototype system runs on a notebook, operating in real-time under Windows. For rapid prototyping MATLAB and SIMULINK are used. Parameter estimation and classification of brain states use: autoregressive parameters, band power values, common spatial filters and hidden Markov models. At this time on-line differentiation between right and left motor imagery is possible with an accuracy of about 90 %.