RR-Interval-Derived Respiration From a Polar H10 Chest Sensor: Agreement With Simultaneous Chest Accelerometry During Fixed and Dynamically Changing Breathing

This methodological study evaluated respiratory rate derived from beat-to-beat RR intervals acquired with a Polar H10 chest sensor against simultaneously recorded chest accelerometry.

Sixteen adult volunteers completed 31 recordings using a standardized 210-s controlled-breathing protocol implemented in KardioAS RespLab v2.9.4. The protocol included fixed paced breathing at 6 breaths/min and a dynamic respiratory-rate sweep from 6 to 15 breaths/min followed by 15 to 4 breaths/min.

During stable breathing at 6 breaths/min, participant-level mean absolute error was 0.38 breaths/min, with no significant systematic bias. During dynamically changing breathing, mean absolute error increased to 1.38 breaths/min, with systematic underestimation at higher respiratory rates and a measurable temporal lag.

The results indicate that RR-interval-derived respiration is most reliable during stable or slowly varying breathing, while rapid respiratory transitions are associated with rate-dependent underestimation and reduced temporal responsiveness.