Abstract
Stress negatively impacts health, contributing to hypertension, cardiovascular diseases, and immune dysfunction. While conventional diagnostic methods, such as self-reported questionnaires and basic physiological measurements, often lack the objectivity and precision needed for effective stress management, wearable devices present a promising avenue for early stress detection and management. This study conducts a three-stage validation of wearable technology for stress monitoring, transitioning from controlled experimental data to real-life scenarios. Using the controlled WESAD dataset, binary and five-class classification models were developed, achieving maximum accuracies of 99.78 %±0.15 % and 99.61 %±0.32 %, respectively. Electrocardiogram (ECG), Electrodermal Activity (EDA), and Respiration (RESP) were identified as reliable stress biomarkers. Validation was extended to the SWEET dataset, representing real-life data, to confirm generalizability and practical applicability. Furthermore, commercially available wearables supporting these modalities were reviewed, providing recommendations for optimal configurations in dynamic, real-world conditions. These findings demonstrate the potential of multimodal wearable devices to bridge the gap between controlled studies and practical applications, advancing early stress detection systems and personalized stress management strategies.•Stress detection methods were validated using multimodal wearable data in controlled (WESAD) and real-life (SWEET) datasets.•Commercial wearable technologies were reviewed, offering insights into their applicability for practical stress monitoring.