An analysis of multimodal emotional braincomputer interface

Abstract

Emotional brain computer interface is a branch of emotional intelligence, which can be divided into emotional recognition brain computer interface and emotional regulation brain computer interface. At present, the main research direction of emotional brain computer interface is to accurately identify emotions and realize human-computer interaction. Because physiological signals can more intuitively and accurately express people’s real emotional state, researchers often use the features extracted from EEG to train emotion recognition classifier. However, recent studies have shown that single-mode signals alone can not comprehensively and objectively evaluate emotions, while multi-mode signals have good complementary characteristics, which can improve the classification accuracy. In practical application, the impact of defects of single signals on classification results can be reduced by adjusting weight parameters, Therefore, multimodal brain computer interface has gradually become one of the hot research directions of emotional intelligence. The difficulty of multimodal brain computer interface is how to select the correct fusion method to improve the performance of emotion recognition. The decision level fusion method is widely used in the research. The research of emotion recognition has very important theoretical and practical significance, and plays an irreplaceable role in human-computer interaction system. This paper describes emotion recognition, fusion methods, research results and application prospects around multimodal brain computer interface.