The Hilbert-Huang Transform as a method and tool to support the analysis of non-linear and non-stationary electroencephalographic signals

The analysis of non-stationary and non-periodic signals has long been restricted to Fourier decomposition in the frequency domain. However, when bringing the frequency content to the time domain (STFT), harmonics appear that do not correspond to reality. A second attempt at temporary decomposition of the signal is by means of the Wavelet transform, which does not fully adapt to the interpretation of the original signal. In the two previous cases, both transforms start from a normative base of sinuses and cosines in Fourier and wavelets in Wavelet. As an alternative to these methods, an adaptive method is presented, the Hilbert Huang Transform that better decomposes, in the frequency and temporal domains, non-periodic and non-stationary signals, in various situations. This study emphasizes the rapid and effective analysis of electroencephalographic signals and presents a method for filtering EEG signals in subjects undergoing a cognitive activities experiment.