Empirical mode decomposition-based approach for intertrial analysis of olfactory event-related potential features

Abstract

This study presents an empirical mode decomposition (EMD)-based method to study the intertrial variability of olfactory event-related potential (OERP) features. The olfactory stimulus in this study was a mixture of 60 % humidity air and 40 % phenyl ethanol alcohol generated by a computer-controlled olfactometer with a constant flow rate of 8 L/min. A 32-channel whole-head EEG system was utilized to investigate the olfactory responses in 12 healthy subjects. Each EEG epoch was segmented based on the olfactory stimulus onset and subsequently decomposed into a set of intrinsic mode functions (IMFs) by using EMD. Only IMFs that met both frequency and spatial dual criteria were chosen as OERP-related IMFs for reconstructing the noise-suppressed single-trial activity, and those significant trials with N1/P2 amplitudes lower/greater than the mean minus/plus two times the standard deviations of baseline amplitudes were denoted as single-trial OERP for intertrial variability analysis. The present approach enables the capability to study intertrial OERP features, such as the latencies and amplitudes of N1 and P2 peaks, on trial-by-trial basis, which may be helpful to shed light on future olfactory dysfunction studies.