Reducing the false positives rate in a BCI system to detect error-related EEG potentials

Abstract

The goal of this paper is to detect error-related EEG potential (ErrP) to perform lower limb rehabilitation tasks. The detection of this potential can be used as a support mechanism to avoid error during this rehabilitation. For that purpose, a graphical interface has been used to simulate the error in the movement of a cursor that generate the appearance of ErrP on a user. The EEG signals were recorded using 16 electrodes and analyzed by two classifiers. Results from both classifiers were then combined in order to improve the performance of the system. Preliminary results suggest that it is possible to detect Errp with a low false positive rate.