Assessing the Accuracy of Popular Commercial Technologies That Measure Resting Heart Rate and Heart Rate Variability
Stone, Ulman, Tran, Thompson, Halter, Ramadan, Stephenson, Finomore, Galster, Rezai, Hagen · 2021 · Frontiers in Sports and Active Living 3:585870
Observational study · Rang 5 von 9
What this work gives Aellic
Consumer devices deviate from a multi-channel ECG reference to differing degrees; an absolute RMSSD comparison across device boundaries is therefore invalid.
Correction of 25 July 2026
The figure previously recorded here — “PPG underestimates RMSSD by 6–11 %” — was a misreading. The 6.84–9.43 % are MAPE, that is, unsigned error magnitudes of individual devices, not a class property of PPG. Two of the three PPG paths OVERestimated RMSSD; the strongest underestimation came from an ECG chest strap. On top of that, n = 5 only — too thin for a hard rule.
The entry on the register was wrong. It was not replaced but kept — with what holds instead standing next to it.
Where this sits in the engine
This work is on the register, but nothing computes with it today — mostly because the feature it belongs to has not been built yet. It is listed so the numbers on the overview do not claim more than they carry.
Register entry stone2021 · Observational study