Expansion of Smartwatch Use in Daily Life Activities: Reliability in Heart Rate Variabbility Measurement

Kelly Correa Baioco da Silva1, Petronio Cabral Ferreira2, Ana Leticia Gomes dos Santos3, stella tassinari maximo4, Christian Goncalves Sassaki3, Samuel Minucci Camargo5, silvia helena bastos de paula6, Jose L Puglisi7, Daniel Gustavo Goroso3
1Universidade de Mogi das Cruzes, 2Universidade Mogi das Cruzes, 3University of Mogi das Cruzes, 4UniversetyofMogidasCruzesandCollegeSãoLeopoldoMandic, 5UniversidadedeMogidasCruzes, 6Instituto de Saúde/SesSP, 7California North State University


Abstract

Previous studies, our group demonstrated that data collected by a smart watch shows a high correlation with the gold standard -ECG. Our objective is to broaden the application of Smartwatch in everyday activities. Two protocols with three phases each and a total duration of 19 minutes were compared: LTL (lying–treadmill–lying) and STS (sitting–treadmill sitting), with treadmill walking at 3.5 km/h. Seven participants were moni tored using a smartwatch (fs-S = 100 Hz), and heart rate variability (HRV) parameters were analyzed in the time domain (Mean RR, SDNN, RMSSD) and frequency domain (LF, HF, LF/HF). Statistical analysis was performed using the Mann–Whitney U-test (Graph Pad Prism v10.4). Results: Differences between the protocols were observed in the time do main: In phase 1, Mean RR was higher in the LTL protocol (0.86 ± 0.19 ms) compared to STS (0.81 ± 0.15 ms), p < 0.01; SDNN was lower in LTL (0.03 ± 0.01 ms) than in STS (0.04 ± 0.02 ms), p < 0.05; RMSSD (< 0.001) showed no significant difference between protocols, p = 0.66. In phase 2, none of the variables showed significant differences (p > 0.36). In phase 3, Mean RR was again higher in LTL (0.85 ± 0.13 ms) than in STS (0.79 ± 0.14 ms), p = 0.02; SDNN and RMSSD showed no significant differences (p = 0.31 and 0.75, respectively). In the frequency domain, LF, HF, and LF/HF showed no sig nificant differences in any phase. The presented values are consistent with physiological changes in cardiac activity. Furthermore, since the results were obtained under everyday activity conditions, they more faithfully represent physiological values unlike data collected in a controlled laboratory setting, which may induce additional stress. Thus, the smartwatch proves advantageous for remote telemonitoring.