Time Segment Analysis of Heart Rate Variability to Evaluate Daily Stress using Wearable Device Technology

Present studies have successfully evaluated psychological properties such as mental health and stress by using physiological data from the cardiovascular system. Most studies established specific interventions and ambiguous heart rate properties according to homeostatic conditions. We proposed a study evaluating mental stress based on daily activities dataset. Twenty-two healthy men were observed in this study. We employed two approaches based on the time segments, while extracting the HRV parameters. We discovered that there was no significant difference between the parameters corresponding to the daily stress score groups (low- and high-stress) when we used whole-day recording in one segment HRV parameter measurement (p > 0.05). However, by extracting the HRV parameters based on multi time segments (phases 1, 2, and 3), we found parameters that were able to properly distinguish the two groups (low- and high-stress). The frequency domain parameters are the most sensitive features, especially the LF and HF (p < 0.01), followed by the total power (p < 0.05). In the time domain measurement, the RMSSD, StdHR, SD1, and SD2 are able to differentiate the participants based on the daily stress scores (p < 0.05). As a result, this study proposed that by continually monitoring biological signals based on time segment and employing the given parameters, it is possible to appropriately and meaningfully measure the daily stress condition for future classification studies.