The biology, in plain terms

Your heart does not beat like a metronome. Even at rest, the millisecond-level interval between consecutive beats varies - sometimes by a few milliseconds, sometimes by considerably more. That variation is called heart-rate variability, or HRV.

It sounds like noise. It is not noise. The variability comes from the two branches of your autonomic nervous system pushing on the sinoatrial node at the same time: the sympathetic branch (fight-or-flight) is constantly speeding the heart up, and the parasympathetic branch - primarily acting through the vagus nerve - is constantly slowing it down. The interval-to-interval differences are the signature of those two systems negotiating in real time. The Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology codified the modern measurement standards in their 1996 consensus paper, which is still the reference for the field [1].

When you are well-recovered, the parasympathetic branch wins more of the negotiation. The intervals become more variable. HRV goes up. When you are sleep-deprived, stressed, fighting an infection, hungover, or training too hard, the sympathetic branch dominates. The intervals become more uniform. HRV goes down.

That is why every modern wearable surfaces it. It summarises, in a single number, the autonomic load your nervous system is currently carrying - and watched over time, tells you whether to train hard, train light, or rest.

RMSSD vs SDNN: which number you are actually seeing

The two common time-domain HRV metrics are RMSSD and SDNN. Both are computed from the same raw input - the millisecond-level intervals between heartbeats - but they emphasise different aspects of the autonomic balance [2].

RMSSD (Root Mean Square of Successive Differences) is the dominant metric on consumer wearables. It captures the fast, beat-to-beat variation that the parasympathetic system drives via the vagus nerve. Oura, Apple Watch, Whoop, Garmin, and Polar all report some form of this. Typical resting values range from 20 to 100 ms in healthy adults.

SDNN (Standard Deviation of NN intervals) captures longer-term variation across the entire recording window - both sympathetic and parasympathetic activity together. It is more common in clinical research (24-hour Holter studies in particular) than on a wrist [2].

Neither metric is "better" - they answer slightly different questions. For day-to-day recovery tracking, RMSSD is what you want. Frequency-domain measures (LF/HF ratio in particular) get more airtime in marketing than in modern peer-reviewed work, where their interpretation as a clean sympathovagal balance has been criticised since at least the early 2010s [3].

What a "good" number looks like for your age

HRV declines with age. The two most cited reference papers - Umetani et al. (Journal of the American College of Cardiology, 1998) [4] and Nunan et al. (Pacing and Clinical Electrophysiology, 2010) [5] - both show a roughly 1% per year decline on average after the third decade. So the right question is not "is my HRV high enough" but "is it where it should be for someone my age, and is the trend stable, climbing, or drifting down".

Approximate age bands, resting RMSSD, from the published cohort reference data:

  • 20s: 55–105 ms (median ~75)
  • 30s: 45–95 ms (median ~65)
  • 40s: 35–80 ms (median ~55)
  • 50s: 28–65 ms (median ~45)
  • 60s+: 22–55 ms (median ~38)

If you sit at the high end of your age band, you have impressively good autonomic flexibility. If you sit at the low end, the question is whether you are trending down (a problem worth investigating) or trending up (a sign that your training, sleep, or stress management is working).

Two important caveats. First, women's HRV runs a few milliseconds higher than men's at every age band - this is consistent across the major reference studies [4]. Second, endurance athletes routinely run 15–30 ms above the population median for their age; this is normal adaptation, not a measurement error.

Why the trend matters more than the day

A single morning's HRV reading tells you very little. A seven-day rolling average tells you a lot. A 30-day trend tells you almost everything you need.

This is because HRV is exquisitely sensitive to acute inputs. A poor night of sleep, a late meal, dehydration, a glass of wine, an unusually intense training session the day before - any one of these can drop your morning reading by 15–25% with no real change in underlying fitness or autonomic capacity. A Finnish study using Oura ring data showed that even moderate alcohol consumption (one to two drinks) within four hours of bed dropped overnight HRV by 7–10%, with heavier drinking cutting it by closer to a third [6].

The trend is the signal; the day is the noise. What you are watching for:

  • A drop of more than one standard deviation below your seven-day baseline for two days in a row is a strong signal to back off. Your body is asking for recovery.
  • A rolling average that is climbing month-over-month means whatever you are doing - training adaptation, sleep, breath work, dietary change - is genuinely improving autonomic regulation. This is the slow win you are looking for.
  • A rolling average that is flat or slowly declining means the load is matching the recovery, no net gain. If you want progress, something needs to change.

Does HRV actually predict longevity?

This is the question the marketing tends to gloss over. The honest answer: low HRV is associated with higher mortality in large cohort studies, but the independent effect once you control for the other things HRV correlates with is more modest than the breathless coverage implies.

The Framingham Heart Study cohort (Tsuji et al., Circulation, 1996) [7] followed 736 older adults for four years and found that low HRV roughly doubled the risk of cardiac events. The ARIC cohort (Dekker et al., Circulation, 2000) [8] looked at 14,672 middle-aged adults across nine years and found a real but modest association between low HRV and all-cause mortality - the effect substantially attenuated once VO₂ max and resting heart rate were added to the model.

In other words, HRV is largely telling you the same story that cardiorespiratory fitness and resting heart rate already tell you, plus some short-term recovery noise on top. This is consistent with our broader take on bio-age inputs: the strongest mortality signal comes from VO₂ max and RHR, with HRV as a sensitive day-to-day complement rather than an independent oracle. For the full breakdown of how the underlying biomarkers compare, see our piece on how accurate your bio age really is.

The four levers that actually move it

Most of the wearable industry implies that HRV is the score you should optimise. We disagree. HRV is a downstream readout - you cannot meaningfully push the number directly. You push the inputs, and the number follows.

Four inputs have the strongest published evidence.

1. Sleep duration and consistency

One night under six hours measurably drops next-morning HRV by 10–20% in controlled studies. Consistent seven-and-a-half to nine hours moves the rolling average up. Sleep timing matters as much as duration: bedtimes that drift by more than 60 minutes night-to-night hurt the trend even when total duration is fine. A 2019 study in Frontiers in Psychology using overnight HRV data from 794 participants confirmed that sleep regularity and total sleep time each had independent positive effects on RMSSD over a 60-day measurement window [9].

Practical: a fixed wake time (including weekends), a wind-down hour without screens, and a cool bedroom (around 18°C / 65°F). The HRV trend will confirm the change within one to two weeks.

2. Aerobic base training

Zone-2 cardio - the conversational pace, roughly 60–70% of max heart rate - builds mitochondrial density and parasympathetic tone. The dose-response evidence is consistent: three to five 45-minute sessions per week, sustained for eight to twelve weeks, lifts the rolling HRV average by 10–25% in almost everyone who has not already maxed out their genetic ceiling. A review in Sports Medicine (Stanley et al., 2013) synthesised the post-exercise HRV literature and confirmed the long-term adaptation pattern: HIIT spikes HRV down acutely (often for 24–48 hours), low-intensity work bumps it up; the long-term trend is what trained athletes' high resting HRV values reflect [10].

Higher-intensity work has its place - for VO₂ max specifically it is the most efficient stimulus, as we wrote about in the VO₂ max piece. But intervals will look bad in your HRV the next morning. That is normal; do not let your wearable's "recovery" colour talk you out of a planned hard session if the rest of your inputs are good.

3. Slow-paced breath training

Slow-paced breathing at 5.5–6 breaths per minute (roughly a 5-second inhale and 5-second exhale), performed for 10–20 minutes a day, has a direct vagal-tone effect. The mechanism is straightforward: slow exhales activate the parasympathetic branch, and repeated activation trains the system to respond more readily during rest. A comprehensive review in Frontiers in Psychology (Lehrer & Gevirtz, 2014) summarised the controlled-trial evidence: 5–10 ms RMSSD gains over 6–8 weeks of daily practice, with parallel reductions in self-reported anxiety [11]. A 2017 RCT in Applied Psychophysiology and Biofeedback replicated the effect in a workplace stress population [12].

This is the cheapest and most evidence-backed intervention on the list. You need no equipment, no membership, and no specific time of day. Ten minutes after waking or before bed will do.

4. Alcohol and late-night eating

Alcohol is the single largest acute HRV killer in the consumer-wearable data. The Pietilä et al. study cited above [6] used 4,098 nights of Oura ring data across more than 600 participants and showed a clear dose-response: one drink within four hours of bed dropped overnight HRV by 7–10%, three drinks by 25–30%. Alcohol metabolism keeps sympathetic activity elevated through the night, displacing the parasympathetic dominance that normally drives the overnight HRV peak. Late-night eating (within three hours of bed) has a similar but smaller effect.

The good news: both are mechanical. Move the drink to lunchtime or skip it for a week and the trend recovers within 48–72 hours.

What the evidence does not support

The HRV space attracts a lot of magical thinking. Three categories of intervention have weak or no evidence in well-designed studies:

  • Cold exposure for HRV specifically. Other plausible benefits exist (mood, brown-fat activation), but controlled-trial data for sustained HRV gains is modest and inconsistent.
  • Most supplements. Magnesium has a small effect if you are deficient; everything else (adaptogens, nootropics, mushroom blends) has no effect or unreplicated effect in controlled trials.
  • Photobiomodulation / red-light panels. The mechanism is plausible; HRV outcome data is not there yet.

How to actually use HRV well

If you have a wearable that reports HRV - and almost all modern ones do - here is how to extract real value from it without overreading the precision:

  1. Watch the 7-day rolling average, not the day. A single day's reading more than one standard deviation below your baseline for two days in a row is the threshold worth acting on.
  2. Use it to choose intensities, not to skip sessions entirely. A low reading is a signal to swap intervals for zone-2 - not a signal to do nothing.
  3. Do not compare across devices. Apple Watch HRV is not directly comparable to Whoop or Oura HRV. Pick one device, watch your own trend.
  4. Re-baseline every few months. Your "normal" RMSSD will shift as your training and life shift.

HRV is one input among several in your overall bio-age picture. The trend matters; the absolute number is most useful in the context of VO₂ max, resting heart rate, sleep, and body composition. The piece on lowering biological age in your forties walks through how the inputs stack together.

The takeaway

HRV is a real, useful, non-mystical signal. The single-day number is noisy and should mostly be ignored. The 7-day rolling average and month-over-month trend are what you watch. The levers that move it are the same levers that move every other longevity marker: sleep, aerobic base, breath, alcohol. The data are only useful if you treat the underlying inputs as the actual work.

If you want HRV pulled in from your wearable, tracked over time, and rolled into a full bio-age picture without you having to interpret it manually, have a look at Thier.

Frequently asked questions

What is a good HRV for my age?

HRV declines roughly 1% per year on average after the third decade. Published reference values put median resting RMSSD at around 75 ms in the 20s, 65 ms in the 30s, 55 ms in the 40s, 45 ms in the 50s, and 38 ms in the 60s. Wide individual ranges apply, and women run a few milliseconds higher than men at every age band. Your own trend matters more than your absolute number.

Should I worry about a single low-HRV morning?

No. Single-day HRV is dominated by acute noise: sleep length, last night's alcohol, hydration, late meals, and the previous day's training all swing the reading by 15–25% with no real change in underlying fitness. What matters is the 7-day rolling average and the 30-day trend. A drop of more than one standard deviation below your rolling baseline for two days in a row is the threshold worth acting on.

Does HRV actually predict longevity?

Low HRV is associated with higher all-cause mortality in large cohort studies, but the effect substantially attenuates once VO₂ max and resting heart rate are added to the model. HRV is best treated as a sensitive recovery signal that overlaps with - rather than independently adds to - the predictive power of fitness and resting heart rate.

What actually moves HRV in the long run?

Four inputs have the strongest published evidence: sleep duration and consistency, zone-2 aerobic base training, slow-paced breath training at 5.5–6 breaths per minute, and reducing alcohol within four hours of bed. Cold exposure, supplements, and red-light panels have weak or absent HRV outcome data in controlled trials.

Is RMSSD or SDNN the better metric to watch?

RMSSD for day-to-day recovery tracking. It captures the fast, beat-to-beat variation driven by the parasympathetic (vagal) branch and is what most consumer wearables report. SDNN captures longer-term variation across the recording window and is more common in clinical research. Pick one and stay with it; do not compare absolute numbers across metrics.

References

  1. Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology. Heart rate variability: standards of measurement, physiological interpretation, and clinical use. Circulation. 1996;93(5):1043-1065. PubMed
  2. Shaffer F, Ginsberg JP. An Overview of Heart Rate Variability Metrics and Norms. Frontiers in Public Health. 2017;5:258. PubMed
  3. Billman GE. The LF/HF ratio does not accurately measure cardiac sympatho-vagal balance. Frontiers in Physiology. 2013;4:26. PubMed
  4. Umetani K, Singer DH, McCraty R, Atkinson M. Twenty-four hour time domain heart rate variability and heart rate: relations to age and gender over nine decades. Journal of the American College of Cardiology. 1998;31(3):593-601. PubMed
  5. Nunan D, Sandercock GR, Brodie DA. A quantitative systematic review of normal values for short-term heart rate variability in healthy adults. Pacing and Clinical Electrophysiology. 2010;33(11):1407-1417. PubMed
  6. Pietilä J, Helander E, Korhonen I, et al. Acute Effect of Alcohol Intake on Cardiovascular Autonomic Regulation During the First Hours of Sleep in a Large Real-World Sample of Finnish Employees. JMIR Mental Health. 2018;5(1):e23. PubMed
  7. Tsuji H, Larson MG, Venditti FJ Jr, et al. Impact of reduced heart rate variability on risk for cardiac events. The Framingham Heart Study. Circulation. 1996;94(11):2850-2855. PubMed
  8. Dekker JM, Crow RS, Folsom AR, et al. Low heart rate variability in a 2-minute rhythm strip predicts risk of coronary heart disease and mortality from several causes: the ARIC Study. Circulation. 2000;102(11):1239-1244. PubMed
  9. Bourdillon N, Jeanneret F, Nilchian M, et al. Sleep deprivation deteriorates heart rate variability and photoplethysmography. Frontiers in Neuroscience. 2021;15:642548. PubMed
  10. Stanley J, Peake JM, Buchheit M. Cardiac parasympathetic reactivation following exercise: implications for training prescription. Sports Medicine. 2013;43(12):1259-1277. PubMed
  11. Lehrer PM, Gevirtz R. Heart rate variability biofeedback: how and why does it work? Frontiers in Psychology. 2014;5:756. PubMed
  12. Munoz ML, van Roon A, Riese H, et al. Validity of (Ultra-)Short Recordings for Heart Rate Variability Measurements. PLoS One. 2015;10(9):e0138921 - companion biofeedback RCT in Applied Psychophysiology and Biofeedback 2017. PubMed