Analysis of single-trial EEG data using a combined robust pre-whitening technique and ICA approach

In electroencephalograph (EEG) experiment, the measure result are usually influenced by different kinds of noise with high power. Treating an averaged event-related potential (ERP) data is a main approach in recent topics of applying independent component analysis (ICA) to EEG signal processing. By taking the average, the signal-noise ratio (SNR) is increased, however, some important information such as the trial-by-trial variation of the amplitude is lost during the averaging. On the other hand, the analysis of single-trial data is generally hard in the low SNR. This paper presents a robust pre-whitening technique with ICA approach for the unaveraged single-trial EEG data. Our approach is based upon the two techniques: decorrelation with a high-level additive noise reduction and decomposition of individual source components. The results on the unaveraged visual evoked potential (VEP) single-trial data analysis illustrate that not only the behavior and location but also the activity strength (amplitude) and dynamics of the individual evoked response can be visualized by the proposed method.