SDNN versus RMSSD: why HRV differs between Apple Watch and other apps

SDNN and RMSSD are different ways of calculating HRV

Heart rate variability describes variation in the intervals between heartbeats, but it has more than one calculation method. SDNN and RMSSD are two commonly used measures.

Imagine one record where intervals change gradually and another alternating longer and shorter intervals each beat. Overall spread and successive differences will not be the same. SDNN and RMSSD differ because they focus on different kinds of variation.

• SDNN: how far normal beat intervals spread around their mean across the recording
• RMSSD: how much neighbouring normal beat intervals differ from beat to beat
• Shared features: usually expressed in milliseconds and dependent on the quality of the underlying beat data
• Difference: they need not give the same value because they describe different variation

SDNN describes spread across the recording period

SDNN is the standard deviation of intervals between beats classified as normal. Standard deviation describes how widely numbers spread around their mean. SDNN combines slower and faster changes across the full recording period.

Recording length strongly affects SDNN. Longer recordings more readily include shifts in sleep and wakefulness, posture, activity and time of day. Short recordings, five-minute SDNN and 24-hour SDNN cannot be compared against one standard.

Apple HealthKit saves HRV as individual SDNN samples in milliseconds. Despite the shared name, the measurement differs from a clinical 24-hour ECG.

RMSSD describes differences between neighbouring beats

RMSSD summarises differences between each pair of neighbouring normal beat intervals. The calculation squares the differences, averages them and takes the square root. Despite the long name and equation, it can help to remember that it looks at change from one interval to the next.

Relatively fast interval changes contribute strongly to RMSSD. It is related to breathing-associated beat-interval changes, but breathing speed and depth, posture, heart rate and measurement setting also affect it.

It is inaccurate to treat RMSSD as the amount of parasympathetic activity or recovery capacity itself. The 2026 HRV research guidelines also caution against using HRV as a direct measure of sympathetic activity or sympathovagal balance.

Apple Watch, Oura and WHOOP use different measures and summaries

Apple Health, Oura and WHOOP all display HRV in milliseconds, but their metrics and methods of producing a daily value differ. Features and algorithms change, so check the current explanation for your product and app.

• Apple Watch / HealthKit: saves individual SDNN samples when movement is sufficiently low and the signal is clear
• Oura: describes using RMSSD to show the overnight pattern and nightly average
• WHOOP: uses RMSSD and combines overnight data in a proprietary way for Recovery
• Other apps: some read HealthKit SDNN; others calculate a different measure from their own sensors or beat intervals

Six reasons the same day’s numbers differ

• Different metrics: SDNN, RMSSD, logarithmic values or proprietary recovery scores
• Different sensors: optical wrist or finger sensors versus electrical heart signals from a chest strap, for example
• Different recording lengths: brief automatic samples, a standardised waking measurement or the whole night
• Different times: individual Apple samples and another app’s nightly summary do not necessarily describe the same moment
• Different processing: methods of removing or correcting movement, irregular beats and poor signals
• Different daily summaries: means, medians, weighted averages or the most stable period

There is no universal SDNN-to-RMSSD conversion

Across groups, SDNN and RMSSD sometimes move in similar directions. Even so, there is no universal formula that gives an exact RMSSD from one SDNN value. A summary number does not preserve the original order of beat intervals.

Do not convert “Apple SDNN is 42 ms” into an assumed RMSSD value. Track SDNN against SDNN and RMSSD against RMSSD.

When reading research, check the metric and recording duration too. SDNN from a 24-hour ECG, RMSSD from five minutes of rest and overnight wearable estimates cannot share one comparison standard simply because all are called HRV.

Five checks before comparing

1. Check the metric name

   Is it SDNN, RMSSD, log-transformed lnRMSSD or a proprietary score? If the label only says HRV, read the app’s explanation.

2. Check the data source

   In iPhone Health → Browse/Search → Heart → Heart Rate Variability, check the time of each sample and which device or app wrote it.

3. Choose similar times

   Separate sleep, just after waking, daytime and breathing sessions. Treat automatically collected samples separately from deliberate breathing measurements.

4. Align duration and posture

   Do not mix short recordings with nightly summaries or sitting with lying down. Compare only comparable conditions.

5. Follow the same metric over time

   Choose a device, app and measure and compare with your own range across weeks. It does not need to match another app’s number.

When values are extreme or symptoms occur

Optical sensor errors, movement, irregular beats and beats not properly excluded can affect both SDNN and RMSSD. With frequent irregular beats, extremely high HRV cannot necessarily be interpreted as excellent recovery.

HRV is not a test to detect or rule out arrhythmias such as atrial fibrillation. For chest pain, severe breathlessness, fainting or sudden sustained intense palpitations, seek local urgent medical advice or emergency care before investigating differences between apps.