Alicia sees a resting heart rate of sixty beats per minute and imagines one beat exactly every second. Tricia plots the actual intervals and finds 0.92 seconds, 1.04, 0.97, 1.08 and 0.99. Kai Kai asks whether the variation is noise that should be averaged away or information about how the sinoatrial node is being regulated.
Heart rate variability, HRV, is the variation in the time intervals between successive heartbeats over a defined recording period. In normal sinus rhythm, those intervals change because the sinoatrial node is continually modulated by parasympathetic activity, sympathetic influences, respiration, baroreflexes, thermoregulation, hormones, posture, activity and intrinsic pacemaker dynamics. HRV is therefore a property of a time series, not one heart-rate number.
This article supports How the Heart Works. How Autonomic Control of the Heart Works owns the efferent neural mechanisms. How the Baroreflex Works owns rapid pressure feedback. Here the reader job is measurement and dynamics: how beat-to-beat timing varies, how that variation is quantified and what HRV does not uniquely prove.
This is educational physiology, not a method for diagnosing stress, heart disease or autonomic disorders from a watch or one HRV score.
1. Average heart rate hides the interval sequence
A heart rate of 60 beats/min can arise from sixty identical 1.00-second intervals or from a mixture of shorter and longer intervals averaging one second.
The average rate is the same in both cases, but the time-series structure differs.
HRV exists because the timing of individual cycles contains information that disappears when all intervals are compressed into one average.
2. HRV begins with correctly identified beats
On an ECG, beat timing is commonly measured between successive R-wave peaks, creating RR intervals.
When analysis is intended to describe normal sinus-node modulation, intervals judged to arise from normal-to-normal beats are often called NN intervals.
The distinction matters because premature beats, missed detections or artifacts can create large interval changes that do not represent the sinus-node dynamics the analyst intended to study.
3. Heart rate and interval are reciprocals
Instantaneous heart rate is approximately 60 divided by the interval in seconds. A 1.0-second interval corresponds to 60 beats/min; 0.5 seconds corresponds to 120; 1.2 seconds corresponds to 50.
This reciprocal relationship is nonlinear. A 100 ms change means a larger rate change when intervals are short than when intervals are long.
HRV metrics are therefore usually calculated from intervals rather than simply from differences between rounded beats-per-minute values.
4. The sinoatrial node is the timing source
Pacemaker cells gradually depolarise until the next action potential begins. Their timing depends on HCN currents, calcium currents, intracellular calcium cycling and autonomic receptor signalling.
Parasympathetic acetylcholine can slow the approach to the next threshold; beta-adrenergic signalling can accelerate it.
Beat-to-beat interval variation therefore emerges partly because the biological oscillator itself is being continually retuned.
5. Vagal modulation can change beat timing quickly
Parasympathetic vagal effects on the sinoatrial node can change rapidly because acetylcholine acts through fast receptor and ion-channel pathways and is rapidly broken down.
This rapid timescale makes high-frequency beat-to-beat variation particularly sensitive to respiratory modulation of cardiac vagal activity under controlled conditions.
The 2026 guidelines Guidelines for rigor and reproducibility of heart rate variability within human cardiovascular research emphasise that HRV can reflect respiratory modulation of parasympathetic cardiac activity but should not be overinterpreted as a universal direct measure of vagal tone.
6. Sympathetic influence has a slower and more complex signature
Sympathetic signalling also changes sinoatrial rate, but its effects develop and decay on different timescales and interact with baroreflex and vascular changes.
There is no single HRV number that directly reports cardiac sympathetic nerve firing.
The 2026 guidelines explicitly caution that HRV is not appropriate as a specific marker of cardiac sympathetic outflow or a direct sympathovagal-balance meter.
7. Respiration produces structured variability
During ordinary breathing, heart rate often increases during inspiration and decreases during expiration, a pattern called respiratory sinus arrhythmia.
The mechanism includes respiratory modulation of vagal outflow, central respiratory-autonomic interactions and mechanical changes in venous return and baroreflex input.
The variation is therefore not random noise. It is organised around a second biological oscillator: breathing.
8. Breathing rate changes the frequency where respiratory HRV appears
A person breathing 12 times per minute has a respiratory frequency of 0.2 Hz. Breathing 6 times per minute corresponds to 0.1 Hz.
Because respiratory cardiac modulation follows the breathing rhythm, changing breathing frequency shifts the spectral location of that variability.
This is why frequency-domain HRV cannot be interpreted intelligently without knowing the respiratory pattern.
9. The baroreflex adds pressure-linked timing changes
Arterial pressure fluctuates from beat to beat. Baroreceptor stretch signals influence autonomic output, which can alter the timing of subsequent sinoatrial firing.
This creates closed-loop interaction: RR intervals influence cardiac output and pressure; pressure influences baroreceptors; baroreflex output influences later RR intervals.
HRV therefore contains contributions from feedback dynamics, not just direct central autonomic commands.
10. Time-domain metrics summarise interval variation directly
Time-domain measures calculate statistical properties of the interval series itself without first decomposing it into frequencies.
Common examples include SDNN, the standard deviation of NN intervals, and RMSSD, the root mean square of successive interval differences.
The metrics answer different questions. SDNN reflects total variability over the chosen recording duration; RMSSD emphasises short-term beat-to-beat change.
11. RMSSD emphasises successive differences
To calculate RMSSD, subtract each NN interval from the next, square those differences, average the squares and take the square root.
Large slow changes across an entire recording contribute less directly than they do to SDNN because RMSSD focuses on adjacent beats.
Under controlled resting conditions, RMSSD is commonly used as a short-term index sensitive to parasympathetic respiratory modulation, but it is not a direct nerve recording.
12. SDNN depends strongly on recording duration
A five-minute recording captures a narrower range of physiological cycles than a 24-hour recording containing sleep, waking, movement, meals and posture changes.
SDNN from those two recording lengths therefore has different meaning and magnitude.
Comparing HRV metrics requires matching time window as well as mathematical definition.
13. Frequency-domain analysis asks which oscillation rates contain variance
Spectral methods decompose variability into oscillatory components across frequency. Power represents how much variance is associated with a given frequency region.
In conventional short-term analysis, frequency bands historically labelled high frequency and low frequency are often reported.
Those labels are analysis bands, not anatomical channels. A frequency band can contain contributions from several interacting physiological processes.
14. High-frequency HRV is strongly linked to respiratory vagal modulation
When breathing frequency lies inside the conventional high-frequency range, respiratory sinus arrhythmia contributes substantial power there.
This makes high-frequency HRV useful for studying respiratory cardiac vagal modulation under controlled conditions.
It should not be described as a pure measure of total parasympathetic nerve activity independent of breathing, posture, age and recording method.
15. Low-frequency HRV is not a pure sympathetic band
Low-frequency variability contains baroreflex-related oscillations and mixed autonomic influences. It cannot be assigned uniquely to sympathetic cardiac activity.
At slow breathing rates, respiratory modulation itself can shift into the low-frequency range.
This is a central reason the simple interpretation “LF = sympathetic” is scientifically unreliable.
16. LF/HF is not a validated universal sympathovagal-balance gauge
Dividing low-frequency power by high-frequency power produces the LF/HF ratio. Historically, this was often described as sympathovagal balance.
That interpretation assumes LF maps uniquely to sympathetic activity and HF uniquely to parasympathetic activity—assumptions that do not hold generally.
Current 2026 HRV guidance explicitly cautions against using HRV as a specific measure of sympathetic outflow or sympathovagal balance. The ratio is a mathematical relation between two spectral bands, not a direct meter with sympathetic units on one side and vagal units on the other.
17. Nonlinear metrics examine structure beyond variance
Beat-to-beat timing can be analysed using Poincaré plots, entropy measures and other nonlinear methods.
A Poincaré plot graphs each interval against the next, revealing the geometry of short- and longer-term variation. Entropy approaches attempt to quantify irregularity or complexity.
These metrics can expose features not contained in one variance number, but their physiological interpretation still depends on recording quality, duration and state.
18. Stationarity matters
Many spectral and statistical methods assume the process is reasonably stable over the analysis window. If the person stands up, begins exercising or changes breathing pattern halfway through, the underlying generating process changes.
The combined HRV may then mix two states rather than describe one stationary condition.
Segment selection is therefore part of the scientific question. More data is not always better if the extra data belong to a different physiological state.
19. Ectopic beats and detection errors can dominate HRV metrics
A missed R wave can make one interval appear twice as long. A false detection can create two artificially short intervals. One premature beat can create a short-long interval pair much larger than ordinary sinus variability.
Those events can inflate RMSSD, SDNN and spectral power if they are not handled appropriately for the analysis purpose.
HRV therefore begins with signal-quality control, beat classification and transparent preprocessing rather than with a final score.
20. ECG and optical pulse intervals are related but not identical
An ECG measures the electrical R-wave timing of ventricular depolarisation. A photoplethysmographic wearable detects a peripheral pulse wave after ejection and vascular propagation.
The interval between pulse arrivals—pulse-rate variability—often tracks RR variability at rest, but electromechanical delay and pulse-transit time can vary.
Wearable HRV therefore depends on both beat detection and the physical signal used. Device algorithms, motion and skin contact become part of the measurement chain.
21. Posture changes HRV through autonomic and mechanical pathways
Moving from lying to standing redistributes blood, reduces central filling transiently and activates baroreflex responses.
Heart rate rises and parasympathetic respiratory modulation commonly decreases while sympathetic vascular support increases.
An HRV value obtained supine therefore should not be compared casually with one obtained standing as if posture were irrelevant.
22. Exercise changes the meaning of the interval series
As exercise intensity rises, heart rate increases and intervals shorten. Vagal withdrawal reduces much of the high-frequency respiratory modulation present at rest.
Motion artifact and breathing changes also become larger measurement problems.
Resting HRV metrics therefore should not simply be carried unchanged into vigorous exercise interpretation. The operating state and signal quality have changed.
23. Sleep produces large state-dependent changes
Autonomic state, breathing, movement and arousal vary across sleep stages and across the night.
Long overnight recordings therefore contain structured changes rather than one homogeneous stationary signal.
This is why nightly wearable HRV can be useful for longitudinal comparison only when measurement timing, device method and context are reasonably consistent.
24. A higher HRV number is not universally better
In many controlled contexts, reduced HRV is associated statistically with age, illness or reduced autonomic flexibility, and certain HRV measures have prognostic value in specific populations.
But extremely irregular rhythms, ectopic beats or signal artifacts can also produce large variability.
HRV must therefore be interpreted as a defined measurement of a defined rhythm under defined conditions, not a universal wellness score where larger always means healthier.
25. Worked problem: same average rate, different RMSSD
Series A alternates intervals of 0.9 and 1.1 seconds. Series B contains intervals clustered near 1.0 second. Both can average approximately 60 beats/min.
Series A has much larger successive differences and therefore a larger RMSSD.
The mean rate did not change, but the beat-to-beat timing structure did.
26. Worked problem: slow breathing moves respiratory power
A subject breathes at 12 breaths/min, or 0.20 Hz, during one recording. During another, breathing slows to 6/min, or 0.10 Hz.
Respiratory-linked HRV therefore shifts from around 0.20 Hz toward 0.10 Hz.
If an analyst interpreted all 0.10 Hz power as sympathetic simply because it falls in a conventional low-frequency band, the breathing change would be misclassified.
27. Worked problem: one artifact overwhelms short-term HRV
A clean five-minute series has successive NN differences mostly around 20–50 ms. One missed beat creates an apparent interval that is roughly twice normal followed by an abnormally short correction interval.
Squaring those huge successive differences can substantially inflate RMSSD.
The final number may be mathematically correct for the corrupted series yet physiologically meaningless for sinus HRV.
28. Worked problem: equal RMSSD, different long-term structure
Two recordings can have similar average successive beat differences and therefore similar RMSSD, while one slowly drifts in rate across the session and the other remains centred around a stable mean.
Their longer-term variance and spectral content can differ despite matching RMSSD.
No single HRV metric captures every timescale of the interval series.
29. The HRV mechanism in one causal chain
The sinoatrial node generates intrinsically variable pacemaker intervals. Vagal and sympathetic signals alter the rate at which pacemaker cells approach the next action potential. Respiration modulates vagal output and venous return. Arterial pressure fluctuates and feeds back through the baroreflex. Hormones, posture, temperature, sleep, activity and emotional state alter the control network. The resulting sequence of normal beat intervals contains variation across several timescales. Time-domain metrics summarise interval statistics; spectral methods distribute variance across frequencies; nonlinear methods examine other structures. Every metric remains dependent on beat detection, recording duration, breathing and physiological state.
Alicia stops treating sixty beats per minute as sixty identical seconds. Tricia labels the interval sequence rather than only the average. Kai Kai adds respiration and pressure oscillations, then crosses out the simplistic equation “LF/HF = sympathetic/parasympathetic.”
The deeper lesson is that variability can be information, but only when its source, timescale and measurement method are understood. HRV is a window into cardiovascular dynamics, not a direct meter of one autonomic nerve.
Evidence trail and connected reading
The most current methodological anchor is the 2026 paper Guidelines for rigor and reproducibility of heart rate variability within human cardiovascular research, which specifically cautions against treating HRV as a direct measure of cardiac sympathetic outflow or sympathovagal balance. For the autonomic mechanism beneath the interval series, return to the dedicated eduKateSG Autonomic Control and Baroreflex pillars.
Return to the parent: How the Heart Works. Continue to How Autonomic Control of the Heart Works, How the Baroreflex Works, and How the ECG Works.