EEG-based Emotion Classification - A Theoretical Perusal of Deep Learning Methods

Human emotions are the basic element in the determination of a human’s cognitive abilities. These are the reaction of the human brain to internal or external events. Recognizing of different human emotions using a brain-machine interface is a crucial in understanding the emotional well-being of a patient who is completely paralyzed. One of the modalities for capturing neural activities in a brain-machine interface for recognizing human emotions is electroencephalogram (EEG). EEG-based systems for recognizing human emotions could offer a quick, simple, accessible, and user-friendly method for determining emotions. In the last few years emotion identification based on EEG has achieved significant attention. Many machine learning involved technologies have been developed for the classification of emotions of humans. Here we review the modern developments in the research area of emotion identification using deep learning methods, focusing on both feature extraction and classification methods. We also present the comparison of the accuracies obtained by different models.