A Single-Trial Event-Related Potential Estimation Based on Independent Component Analysis and Kalman Smoother

摘要

Event-related potentials have been widely employed to develop brain-machine interface (BMI) systems. To improve the performance of such kind of BMI systems, how to extract P300 wave in a single trial has become an important research question in this field. In this paper, we propose a new approach for extracting P300 wave in a single trial by combining the independent component analysis (ICA) with Kalman smoother. The analysis results from two datasets show that the proposed approach can significantly improve the signal-to-noise ratio (SNR) of the P300 wave and performs better in P300 wave extraction in a single trial than a recursive least squares (RLS) filter.