Optimisation of a BCI system using the GA tehnique

This paper, that continues a previous research, has as primer goal the improvement of a brain computer interface (BCI) system that uses a new features extracting method named Adaptive Nonlinear Amplitude and Phase Process (ANAPP). The ANAPP method models the EEG signals as a combination of five a priori ¿spontaneous cortical oscillations¿ whose amplitudes and phases are established using an adaptive algorithm. While in a series of previous researches the amplitude features of the model were extensively used, in this research the opportunity of using supplementary the phase information within the BCI system is analyzed. In addition, in this paper, the number and the type of the input features that feed the classification system are optimized using a GA algorithm. The final goals are to obtain both a faster BCI system and better classification results.