Pervasive healthcare and wireless health monitoring has been a central area to innovate novel personal and precise medicine for individuals. Many recent wireless health monitoring systems draw upon physiological signals acquired from wearable wristban...
Pervasive healthcare and wireless health monitoring has been a central area to innovate novel personal and precise medicine for individuals. Many recent wireless health monitoring systems draw upon physiological signals acquired from wearable wristbands. However, most studies have paid attention to monitoring physical health states and relatively few efforts have been made to develop pervasive healthcare solutions for mental health. This study investigates a plausibility of the development of a wireless monitoring system for mental health in individuals by collecting and analyzing physiological signals from a wristband in naturalistic daily life environments. In particular, we aim to predict one’s mental stress level using the short-term heart rate variability (HRV) that is estimated from photoplethysmography (PPG) signals. Day-to-day measurements of PPG signals over a week along with a daily log of stress level were conducted three times a day for a week in each participant, in which participants performed measurements by themselves in their own living circumstances. Measurement times were distributed across noon, evening and night. The recorded signals were wirelessly transmitted to a smartphone via the Bluetooth link. The HRV feature of the ratio of high-frequency (0.15Hz - 0.4Hz) power over low-frequency (0.04Hz - 0.15Hz) power was used to predict individual daily stress levels. Prediction with the night measurements showed the highest accuracy compared to the measurements in other times; individual prediction accuracy reached as high as >90% using the night measurements. Across-subject validation of the proposed system demonstrated fair correlations between true and predicted stress scores, implying a possibility to generalize the system for populations. The results of this study may prove the possibility of the development of wireless mental health monitoring system in ambient environments.