Assessing Occupational Burnout in Psychiatric Professionals Using Multimodal Emotion Recognition Methods

Occupational burnout among psychiatric professionals has a significant impact on both employee wellbeing and the quality of patient care. Traditional assessment methods are limited in their accuracy and inability to provide real-time monitoring. This study proposes a multimodal framework that combines physiological indicators (ECG, PPG) with behavioral cues (facial expressions) for the early detection of burnout. Data collected from 93 psychiatric nurses were used to evaluate the performance of this system, which demonstrated superior results through multimodal fusion (accuracy =0.57), outperforming unimodal methods. Among the individual modalities, facial data achieved the highest accuracy (accuracy = 0.46). For feature extraction, statistical measures such as the mean and standard deviation were utilized. Our findings indicate that the mean value of features played a significant role in enhancing overall performance, while the standard deviation contributed minimally. Future research will focus on refining algorithms for small-sample datasets and developing robust, personalized frameworks to further improve accuracy and applicability.