HRV: RMSSD vs SDNN, and why your wearables disagree

Your heart does not beat at a perfectly steady pace. The time between one beat and the next (the inter-beat interval) varies from beat to beat, and heart rate variability is the name for that variation. It is summarised with a number, and there are several ways to calculate that number, which is where the confusion starts.

The standard reference for how HRV is measured and reported is the 1996 Task Force report of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology. It separates time-domain measures, which work on the intervals themselves, from frequency-domain measures. RMSSD and SDNN are both time-domain measures.

RMSSD is the root mean square of successive differences: the typical size of the change from one inter-beat interval to the next. Because it looks only at neighbouring beats, it mostly reflects fast changes and is less affected by slow drifts over the recording. That makes it practical for short recordings, which is why many wearables report it from a night's sleep.

SDNN is the standard deviation of the NN intervals, the times between successive heartbeats once irregular beats and artefacts have been removed: how spread out all the intervals in a recording are. It takes in slow changes as well as fast ones, so it depends strongly on how long the recording is. An SDNN from a five-minute recording and one from a full day are different quantities, and published reference values for 24-hour, short-term and ultra-short-term recordings are not interchangeable.

Devices also differ in the sensor they use, when they measure (overnight or in short spot checks), and how they turn many readings into one daily number (an average, a value from one stretch of sleep, a maximum). Two devices worn at the same time can report different HRV for the same night without either being wrong.

• Stay with one device and one metric when you look at your own trend. A change of device is a change of method.
• Do not convert RMSSD to SDNN or back. They are calculated differently and depend on the recording, so there is no reliable conversion.
• Look at your own baseline. HRV depends on age, sex, fitness, time of day, sleep, alcohol and illness, so published ranges describe particular groups measured in a particular way, not you.