Reconstruction and Physiological Basis of Samsung’s Galaxy Watch Stress Score

Abstract

The detection and avoidance of negative stress are worthwhile endeavors from economic, individual, and public health perspectives. At the same time, the proliferation of wearables promises monitoring of physiological characteristics continuously and at a low cost. However, existing stress tracking solutions use proprietary methodology when operationalizing physiological measurements into abstract stress measures and remain vague on the implementation details.We investigated the stress measuring capabilities of the Samsung Galaxy Watch Active2 smartwatch, as it was one of the first wearables to be advertised for its stress score functionality. Forty-five participants were exposed to cognitive load under time pressure in a randomized controlled trial. The smartwatch recorded stress scores during the experiment, and an ECG was used to monitor cardiac activity. The latter allowed us to record individual stress reactions through different heart rate variability measures (HRV). Subjective perception of the workload was reported using the NASA Task Load Index questionnaire. The smartwatch primarily relies on heart rate (bpm) to derive its stress metrics. We were able to reconstruct Samsung’s stress score closely; a random forest regressor based on five features models the smartwatch’s stress score with a significant coefficient of determination. However, the physiological responses indicate a sub-optimal induction of cognitive workload through the chosen mental arithmetic task. Smartphone manufacturers are eager to advertise stress monitoring features but often fail to define what exactly these are doing. Moreover, while the literature provides a body of evidence for measuring short-term mental stress through HRV metrics, the smartwatch we investigated primarily relies on plain heart rate readings to determine the mental state of its users. This could indicate a lack of scientific backing and should be considered when choosing a valid and reliable tool for stress monitoring