Novel Method for Single-Trial Estimation of Event-Related Potential

This paper presents a novel method for single-trial estimation of event-related potential based on discrete stationary wavelet transform and independent component analysis (ICA). The discrete stationary wavelet transform could filter high-frequency noise in terms of frequency space subdivision. All wavelet coefficients were concatenated as the input of independent component analysis. FastICA algorithm based on negative entropy was adopted for rapid trial estimation of event-related potential. The results of simulation prove the good feasibility and effectiveness of the proposed method.