What Is a Normal HRV? Personal Baselines, RMSSD and SDNN
Learn why HRV normal ranges differ, how to compare your own readings and what a wearable can tell you. Includes an optional one-week observation sheet.
Updated 2026-11-13 · Archived
The scientific content and references belong to this edition. The complete captured record is available to download below; later corrections appear in subsequent editions. Website design and interactive software are not archived here.
What is a normal HRV?
There is no single normal HRV range that applies to every device, metric, age and recording duration. First find out what your number represents. Then compare readings collected in similar conditions and, where appropriate, with reference data using the same method. A systematic review published in 2025 found major obstacles to combining reference values.[1]
A personal baseline helps describe what is usual for you under a particular method. It does not establish that you are healthy, that an unusually high reading is better, or that a low day diagnoses overtraining. TAP’s guide helps you decide which comparisons are meaningful.
Find the name behind the number
RMSSD and SDNN can both be displayed in milliseconds, but they summarize different aspects of beat-to-beat variation. A proprietary recovery score can combine several inputs and is a different quantity again.[1][2]
| Display | What it describes | Before comparing |
|---|---|---|
| RMSSD, ms | Successive differences between normal beat intervals. | Confirm the recording window, conditions and processing. |
| SDNN, ms | The spread of normal beat intervals over a recording. | A short sample and a 24-hour recording are not interchangeable. |
| Recovery score, points | A manufacturer-defined combination of inputs. | Check the score’s definition and validation, not an HRV reference table. |
Why do online normal-value tables disagree?
Reference studies have used recordings from seconds to 24 hours, different ages and different criteria for “healthy”. Even studies using the same metric differ in time of day, equipment, activity and handling of medication or other influences. A population average is not an individual diagnostic cut-off.[1]
The 2025 review searched through November 2022. Its publication date should not be mistaken for a search that included every study through 2025. It supports caution about a universal range, while compatible reference studies may still be useful for a defined population and protocol.[1]
How to make a personal comparison more useful
Keep the metric and device consistent. For a deliberate resting measurement, also keep time of day, posture, resting period and recording length comparable. For an overnight device, use the manufacturer’s defined overnight series consistently. Record changes in device, firmware or method so that a jump is not automatically attributed to your body.[2][3]
Look at repeated valid readings alongside sleep, illness, activity and how you feel. Do not discard a low value just because it is unwelcome; identify missing or poor-quality recordings separately. Device-specific baseline windows differ, and one week is not a validated universal minimum for a stable baseline.
The same 40 ms can lead to different questions
These are invented values, not reference ranges or health thresholds.
| Situation | What the number permits | What it does not permit |
|---|---|---|
| Usually 35–45 ms; today 40 ms | It falls within this person’s recent observed values, if the method matches. | Calling the person healthy or fully recovered. |
| Usually 65–75 ms; today 40 ms | Check measurement conditions and whether the change persists. | Diagnosing the cause from one reading. |
| Old device: 40 ms; new device: 55 ms | Check metric, duration and agreement between methods. | Concluding that recovery improved by 15 ms. |
Can a wearable be compared with an ECG?
Agreement is a question for a particular device, algorithm, metric and setting. A 2025 validation study compared five wearables with a chest-strap ECG reference across repeated nights in 13 adults. Agreement varied by device; hundreds of nights did not mean hundreds of independent participants.[3]
That study tested basic nocturnal HRV measures, not the validity of every composite readiness score. Its reading note also retains a numerical inconsistency in one device’s bias and agreement limits. It should not supply a precise personal correction factor.[3]
A one-week observation sheet
Use the optional sheet to organize existing readings, not to chase a higher score. Write the metric, device and recording conditions first; then note a daily value, signal-quality issues and relevant context. Leaving a blank for missing data is better than turning it into zero.
At the end of seven days, ask whether the measurements were comparable and whether an apparent change coincided with a changed routine or method. The sheet does not calculate a diagnosis, a normal range or a validated recovery score. Seven days may be insufficient to describe your usual variation.
The CSV opens in a spreadsheet and contains no upload function. Keep completed copies private; there is no need to enter your name or medical details.
What can you do with a change?
First check the measurement. Then consider the pattern and context. A persistent unexplained change or concerning symptoms deserves clinical interpretation rather than a conclusion based on a watch score. A reassuring HRV value should not override symptoms.
If you are testing a routine, choose a relevant outcome too: for example, sleep continuity, tolerability of a usual activity or a specified training measure. Raising HRV alone does not prove better recovery or a treatment effect.[2]
Sources and limits
This guide uses the reference-value review, psychophysiology recommendations and the linked wearable validation study. A 2026 methodological review also argues for clear reporting of signals, windows, respiration and artefacts; its checklist is author-proposed, not an externally validated clinical score.[4] The observation sheet is TAP’s educational aid, not a validated instrument.
Funding and interests
Brozat and colleagues report no external funding and no conflicts. Laborde reports German Sport University Cologne HIFF 920147 and no commercial or financial conflicts. Dial reports US Air Force Research Laboratory S-194-11T-001 and no competing interests. Kazdagli’s funding and interest declarations have not been fully verified. Consulted sections and remaining uncertainties are recorded in the source notes.
References
- Maximillian Brozat; Irina Böckelmann; Stefan Sammito (2025). Systematic Review on HRV Reference Values. Journal of Cardiovascular Development and Disease. https://doi.org/10.3390/jcdd12060214
- Sylvain Laborde, Emma Mosley, Julian F. Thayer (2017). Heart Rate Variability and Cardiac Vagal Tone in Psychophysiological Research - Recommendations for Experiment Planning, Data Analysis, and Data Reporting. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2017.00213
- Dial MB, Hollander ME, Vatne EA, Emerson AM, Edwards NA, Hagen JA (2025). Validation of nocturnal resting heart rate and heart rate variability in consumer wearables. Physiological Reports. https://doi.org/10.14814/phy2.70527
- Kazdağlı H, Özel HF (2026). A reporting checklist for human and animal heart rate variability research. Pflügers Archiv - European Journal of Physiology. https://doi.org/10.1007/s00424-026-03210-y
Publication record
TAP-FTR-0021 · v1.0 · English
Author: TAP Editorial Team ·
The Allostasis Project is founded and supported by Resilio.
Record for v1.0 · Updated
Editorial responsibility: TAP Editorial Team
Editorial note
Under the editorial responsibility of the TAP Editorial Team, this publication used AI assistance for content organization and editing.
Download detailed editorial record (JSON) ↓
How responsibilities and review are recorded ↗© The Allostasis Project. All rights reserved. Reproduction or adaptation requires separate permission, subject to applicable legal exceptions.
Download complete edition record (JSON) ↧
Record integrity & available editions
SHA-256: 46d69887e43a0b48ffe59ea1ca5414e32d4a7e7090b6ee6fa78e5214031a332b
Captured: 2026-11-13T05:44:15.510Z
- v1.0 · 2026-11-13