Basic Taste Intensity Recognition based on sEMG and EEG Signals

Abstract

In recent years, brain-computer interfaces based on sEMG and EEG have made progress in taste sensation recognition, but mainly focus on the classification of different tastes. This paper designs and carries out experiments to record the sEMG and EEG signals of different intensities of the five basic tastes stimulus: sour, sweet, bitter, salty and umami. Then time and frequency domain features are extracted for sEMG and wavelet features are extracted for EEG. Various machine learning algorithms are used to regress the intensities of each taste and compare their performance. The results show that different intensities of sour, bitter and salty are easier to be distinguished, while those of sweet and umami can hardly be distinguished. We discuss the possible physiological mechanisms behind this and it may give more researchers a reference.