An identification and removal method of background EEG rhythmic activities for ERP analyses

摘要

Event-related potentials (ERPs) measured in a single trial often have a low signal-to-noise ratio (SNR) because of relatively large background rhythmic electroencephalographic noise. This study proposed a novel method to improve the SNR of ERPs by combining principal component analysis with multivariate empirical mode decomposition. The performance of the method was evaluated with measured ERPs.