Review on Emotion Recognition Using EEG Signals Based on Brain-Computer Interface System

Abstract

Deep learning is closely related to theories of brain development. Brain-Computer Interface (BCI) is the latest development in human–computer interaction (HCI). The BCI reads brain signals from different areas of the human brain and translates these signals into commands that can be controlled through the computer applications. BCI technology is effective in the field of human emotions recognition, with high accuracy using EEG signals. When the brain signals are collected and analyzed using deep learning algorithms, it helps in diagnosing diseases and in distinguishing between physical and psychological diseases, which is helpful in making a correct medical decision. The combination of feature selection methods and classification algorithms serves to recognize emotion more accurately from EEG signals. Each of these algorithms has degree of accuracy and unique characteristics. In this paper, we have reviewed and discussed the related studies on BCI technology that are most concerned with classification of emotions through EEG signals. In addition, we have reviewed the methods of collecting signals and feature extraction from EEG datasets. The paper also discusses the main challenges faced in emotion recognition using EEG. We have reviewed several recent studies are classified based on the techniques used in the emotion recognition process. The results show a clear increase in research related to emotion recognition as an important area of investigation, and a diversity of techniques being used to extract and classify features. After discussing the challenges, we found that given the state of technological development, the interconnection between technology and medicine will generate a tremendous volume of applied solutions in future, contributing to the development of research in health informatics systems. A comparison of the recent studies in this field has been conducted, and we deduce the wide variety of techniques used to detect emotion and the increasingly accurate results.