Neural Networks Training Based on Sequential Extended Kalman Filtering for Single Trial EEG Classification

The nonstationary nature of the brain signals provides a rather unstable input resulting in uncertainty and complexity in the control. Intelligent processing algorithms adapted to the task are a prerequisite for reliable BCI applications. This work presents a novel intelligent processing strategy for the realization of an effective BCI which has the capability to improved classification accuracy and communication rate as well. A neural networks training based on sequential extended Kalman filtering analysis for classification of extracted EEG signal is proposed. A statistically significant improvement was achieved with respect to the rates provided by raw data.