Classification method for BCIs based on the correlation of EEG maps

This paper describes a new method of classification for a Brain–Computer Interface (BCI) based on a normalized cross-correlation of EEG maps which represent the mental activity of the brain. An optimization protocol has been designed to choose the main parameters of the classifier in order to increase the accuracy on the classification. This protocol has been tested with the registers provided by IDIAP Research Institute for BCI Competition 2003. Three different mental tasks based on motor imagery are performed in these sessions. The data have been processed and tested with the classifier to obtain the optimal success rate and reliability. To that end, the optimization protocol has been applied to select the suitable parameters for the classification. The results are very satisfactory and prove that this kind of classification can be successfully introduced in a real time BCI.