Mean RMSSD by age, five-minute resting ECG (mean ± SD, ms):
| Age | Women | Men |
|---|---|---|
| 25–34 | 42.9 ± 22.8 | 39.7 ± 19.9 |
| 35–44 | 35.4 ± 18.5 | 32.0 ± 16.5 |
| 45–54 | 26.3 ± 13.6 | 23.0 ± 10.9 |
| 55–64 | 21.4 ± 11.9 | 19.9 ± 11.1 |
| 65–74 | 19.1 ± 11.8 | 19.1 ± 10.7 |
Source: Voss et al. (2015), 1,906 healthy adults, lying down after five to ten minutes of rest. These are not overnight wearable averages. SDNN and more detail: HRV by age. Use the table as orientation, not a target.
The short answer
No population number makes yours “good.” The comparison that holds up is you now versus you last month, measured the same way on the same device.
Why HRV benchmarks are mostly noise
HRV varies widely between people, and even between devices on the same person. Three things conspire to make absolute numbers misleading:
- Different metrics and windows. SDNN (Apple Watch) and RMSSD (Oura, Whoop) summarise the same beats differently. In the five-minute recordings above, SDNN ran only about 1.1 to 1.6 times higher than RMSSD; over longer windows the gap changes. Shaffer and Ginsberg (2017) note that recording length strongly influences time-domain values.
- Different times.The same person produces different numbers at different hours. A “your HRV is 25” reading is meaningless without context.
- Different bodies. Within each decade of the table, the standard deviation is between about half and two-thirds of the mean. Nunan et al. (2010), reviewing 44 studies, reported large interindividual variation. Where you sit against your own history tells you more than where you sit in a population.
“We observe a strong diurnal variation in the heart rate variability.”
Published norms, with caveats
The table at the top comes from one well-described study. It is not a consensus, and it is not wearable data. Natarajan and colleagues did publish benchmark tables from Fitbit users by age and sex, but those were computed in Fitbit’s own pipeline, and there is no reason to expect an Oura, Whoop or Apple Watch number to line up with them.
What the studies agree on is direction: HRV falls with age. The exact number for your decade depends on who measured it, how and when.
The comparison that actually matters
Forget the population. Compare yourself to yourself.
Take a seven-day rolling average and watch it over weeks. If it rises while you practise slow breathing, sleep more or drink less, that is consistent with the change helping. It is also one person, uncontrolled, and plenty else moves the same number.
We don’t publish percentage thresholds for what counts as a meaningful change. We found no validated cut-off for a consumer wearable’s rolling average. Single-day spikes mean little either way.
When to be concerned
A sustained drop is a reason to look for a cause, not a diagnosis. Some causes have direct evidence. In health-care workers wearing Apple Watches, Hirten et al. (2021) found the daily rhythm of SDNN differed in the seven days either side of a COVID-19 diagnosis, with changes visible before the PCR test confirmed it. A meta-analysis of sleep-deprivation studies found RMSSD fell, though SDNN did not change significantly (Zhang et al., 2025). Training load is murkier: in elite endurance athletes, Plews et al. (2013) found both rises and falls in HRV alongside poor adaptation. HRV is not a diagnostic tool. It is a signal to investigate.
Conversely, an upward trend during weeks of slow-breathing practice is in line with the research: a 2022 meta-analysis of 223 studies found increases in vagally mediated HRV after multi-session slow-breathing interventions (Laborde et al., 2022). Your own trend is one uncontrolled case. Read it as consistent with that finding, not as proof.