A Comprehensive Survey on EEG-based Emotion Recognition: A Graph-based Perspective

Compared to other modalities, EEG-based emotion recognition can respond to the inner emotional patterns in the brain and, therefore, has become one of the most concerning tasks of brain-computer interfaces and affective computing. Brain activity under emotional states involves widespread activation across multiple regions, with a significant research trend being the use of graph structures to model these complex emotional patterns. However, in EEG-based emotion recognition, there is currently neither a comprehensive summary and categorization of graph-related approaches nor an in-depth exploration of the neuroscientific knowledge related to emotion guiding these methods. In this review, we propose a unified framework grounded in neuroscience to systematically elucidate the commonalities and distinctions among existing graph-related methods. The framework categorizes and analyzes graph-related approaches across three stages, offering a neuroscientific foundation and interpretative insights for their application in EEG-based emotion recognition. In addition, we discuss further applications of graphs in the field.