Emotion Recognition from EEG Using Mutual Information Based Feature Map and CNN

Abstract

Emotion is the fundamental trait of human beings, and brain signal is the most prospectus for emotion recognition. Electroencephalography (EEG) is a preferable brain signal for recognizing emotions. Extracting features from EEG signals is an essential part of this emotion recognition. Recently, EEG channel connectivity features have been widely used in emotion recognition. Automatic emotion recognition (ER) from EEG is a challenging computational intelligence or machine learning task due to the inherited complexity of EEG signals. The aim of the study is to analyze mutual information (MI) based connectivity measures for feature extraction from EEG signals and classify emotions using deep learning. At a glance, this study investigated MI and its two variants, Normalized MI (NMI) and Partial MI (PMI), for connectivity feature extraction from EEG; and then Convolutional Neural Network was employed for emotion classification from these features. Experimental results confirm the effectiveness of PMI based ER method while compared with related methods on a benchmark EEG dataset.