Emerging trends and clinical challenges in AI-enhanced emotion diagnosis using physiological data.

This review explores the relationships between physiological parameters and emotions, as well as the potential value and applications of the use of machine learning to facilitate emotion recognition. First, the relationships between physiological parameters (such as heart rate, respiration, blood pressure, galvanic skin response, electroencephalography, and heart rate variability [HRV]) and emotions are discussed. The impacts of emotional states on these physiological parameters represent a crucial aspect of emotion research. For example, the increased heart rates and faster breathing resulting from excitement or anxiety are physiological changes that cannot be ignored. Subsequently, models used for emotion recognition are introduced. These models employ techniques such as machine learning or deep learning and are trained to detect emotional states on the basis of changes in physiological parameters. These techniques have important applications in clinical psychology, including by helping doctors assess patients' status, diagnose emotional disorders, and guide treatment. In the context of managing emotional disorders such as depression, anxiety, bipolar disorder, and borderline personality disorder, emotion recognition technologies can facilitate accurate emotional monitoring and early intervention, thereby reducing the risk of disease recurrence. These models can be used in the contexts of emotion management and health monitoring, thus helping individuals understand and cope with emotional changes more effectively and improving their quality of life. This paper identifies HRV, which reflects an individual's ability to adapt to stress, emotions, and physical conditions, as a key indicator that can be used in the contexts of emotion recognition and physiological parameter analysis. By incorporating HRV parameters into relevant models, emotional changes can be analyzed more precisely, thereby providing more effective emotion management and health monitoring tools, which can enhance individuals' quality of life. However, the use of these physiological parameters entails many challenges, including those pertaining to the collection of physiological data, privacy and security concerns, and the need for personalized adjustments as a result of the variability observed among individuals in this context. These challenges require continuous efforts on the part of technical experts and researchers to advance the development and application of emotion recognition technologies. Finally, this paper presents an in-depth investigation of the associations between physiological parameters and emotions, and it explores the potential value and challenges associated with the use of machine learning to facilitate emotion recognition. The results of these studies suggest that emotion recognition technology can be used more widely in the contexts of mental health, emotional management, and health monitoring to provide individuals with better emotional support and care.