Single-trial evoked potentials study by combining wavelet denoising and principal component analysis methods.

The authors have developed a new approach by combining the wavelet denoising and principal component analysis methods to reduce the number of required trials for efficient extraction of brain evoked-related potentials (ERPs). Evoked-related potentials were initially extracted using wavelet denoising to enhance the signal-to-noise ratio of raw EEG measurements. Principal components of ERPs accounting for 80% of the total variance were extracted as part of the subspace of the ERPs. Finally, the ERPs were reconstructed from the selected principal components. Computer simulation results showed that the combined approach provided estimations with higher signal-to-noise ratio and lower root mean squared error than each of them alone. The authors further tested this proposed approach in single-trial ERPs extraction during an emotional process and brain responses analysis to emotional stimuli. The experimental results also demonstrated the effectiveness of this combined approach in ERPs extraction and further supported the view that emotional stimuli are processed more intensely.