P300 Detection Using a Multilayer Neural Network Classifier Based on Adaptive Feature Extraction

Abstract

This paper proposes two adaptive schemes for improving the accuracy and the transfer rate of EEG P300 evoked potentials: a feature extraction scheme combining the adaptive recursive filter and the adaptive autoregressive model and a classification scheme using multilayer neural network (MNN). Using the signals extracted adaptively, the MNN classifier could achieve 100% accuracy for all the subjects, and its transfer rate was also enhanced significantly. The proposed method may provide a real-time solution to brain computer interface applications