RheumaPain: a multimodal biosignal dataset for objective pain assessment in pediatric rheumatic diseases

Pain assessment continues to pose a considerable challenge owing to the subjective characteristics of self-reports and the fluctuating intensity of pain. We introduce RheumaPain, an innovative multimodal dataset aimed at reconciling subjective feedback with physiological responses. The dataset contains synchronized physiological recordings from 42 pediatric patients diagnosed with rheumatic diseases (31 female and 11 male; mean age: 13 years) during physiotherapy sessions. We measured Electrodermal Activity (EDA), Blood Volume Pulse (BVP), triaxial Acceleration (ACC), and Skin Temperature (TEMP) with a wristworn wearable device. Signals are in sync with pain intensity scores of the Wong-Baker Faces Pain Scale which shows the difference between resting and exercising phases. The RheumaPain dataset comprises approximately 5.5 h of synchronized multimodal physiological recordings. RheumaPain also has demographic metadata and diagnostic categories of various rheumatic diseases like JIA, SLE, and JDM. It provides a unique basis for personalized pain management and the creation of objective clinical tools in pediatric care.