EEG-based Emotion Classification using Deep Learning: Approaches, Trends and Bibliometrics

Abstract

Emotion classification has emerged as a critical area of research, holding immense significance in the understanding of human behaviour, mental health, and social interactions. The increasing recognition of emotional well-being's crucial role in various domains, such as healthcare, psychology, and human-computer interaction, has driven substantial attention toward accurately classifying and analysing emotions. In this study, we conducted a comprehensive bibliometric analysis to unravel the scientific production and temporal evolution of research related to emotion classification. Leveraging the extensive Scopus database, we meticulously collected and meticulously analysed a diverse range of 440 articles on emotion classification from its inception to the present day. The application of advanced bibliometric measures has yielded vital insights into current trends, patterns, and characteristics in this field of study. Our data indicated an unexpected trend: an increase in research activity, especially after 2018. The understanding of how emotions impact human experiences and behaviour has advanced significantly. Researchers from several fields have emphasised the need of better understanding and describing emotions, resulting in a large rise in study output. However, there is still need for improvement in terms of agreement on emotion categorization assessment approaches and standardisation processes. It is difficult to compare and duplicate study findings due to a lack of precise assessment criteria. To address this problem, it's crucial for researchers to collaborate and develop a common knowledge. The aim of the paper is to widen our knowledge of emotions so that it can eventually result in policies being formed to improve our overall health. This knowledge could be implemented in psychological counselling and health promotion resulting in the development of closer social bonds.