What Does Heart Rate Variability Actually Measure?

Open a heart-rate-variability app in the morning and it hands you a single number, sometimes with a color and a one-word verdict stapled to it: Balanced. Recovering. Low. It reads like a lab result. It isn’t quite one, and the gap between the confidence of that display and the messiness of what’s actually being measured is worth taking apart properly.

None of which means the number is fake. Heart rate variability is a genuine, measurable property of a genuine physiological system. The problem is narrower and more specific: the version on your wrist, the version a cardiology paper reports, and the version your friend’s different app shows her are three different measurements wearing one name.

The gap the number is actually counting

Your heart doesn’t beat like a metronome, even when it’s perfectly healthy. Sit still and measure the exact interval between each beat and you’ll find those intervals drifting slightly, beat to beat — a little shorter, a little longer, constantly adjusting. Heart rate variability is the size of that drift.

The adjusting is done mostly by two branches of the autonomic nervous system pulling in opposite directions: the sympathetic branch, which speeds the heart’s own pacemaker up, and the parasympathetic or vagal branch, which slows it down. Their back-and-forth, moment to moment, is what produces the variation. A heart with more room to swing between the two has, in that specific sense, more adjusting capacity available to it.

This is why heart rate variability is a different quantity from heart rate, even though the two get conflated constantly. Heart rate is an average — beats per minute. Two people can share an identical average of 60 beats per minute while one heart is ticking with near-metronome regularity and the other is swinging widely around that same average. Same rate. Very different variability. The number your tracker calls HRV is trying to capture the second thing, not the first.

Two different ways of catching the same drift

An ECG reads the heart’s electrical activity directly, which is why it remains the reference standard. A wearable instead reads a pulse wave through your skin — an optical proxy for the same timing — and has to infer the same beat-to-beat gaps from a noisier signal.

That gap between methods isn’t uniform across every way of summarizing it, either. Comparisons against ECG have found some time-domain metrics tracking closely — SDNN and a related measure called SD2 sit within roughly 2 to 2.5 percent of the ECG-derived value — while another common metric, pNN50, can be off by around 30 percent under the same conditions. Two apps reporting “your HRV” can be reporting numbers with genuinely different reliability, and most interfaces never tell you which metric is underneath the score.

Why the metric matters as much as the device

RMSSD and SDNN are the two you’ll actually encounter, usually without the name attached. RMSSD looks at the differences between consecutive beat-to-beat intervals and is generally preferred in this kind of tracking because it’s more sensitive to parasympathetic activity specifically and less thrown off by breathing rate or how long the recording ran. SDNN is a broader measure of overall variation, and its value is more directly tied to recording length — a five-minute SDNN and a 24-hour SDNN from the same person are not the same figure, and shouldn’t be read as if they were.

That duration-dependence is a genuinely underappreciated source of “why did my number change” confusion. If your app quietly altered its recording window in an update, or if last night’s sleep tracking ran shorter than usual, the resulting number can shift for a reason that has nothing to do with your physiology.

Why your own number swings around

Set the measurement issues aside and there’s a second, entirely legitimate reason your HRV moves: the system it’s measuring is supposed to respond to your life. Research on the metric describes it declining with age, with an extended sedentary stretch, and possibly with overtraining, while rising with sustained aerobic exercise as resting vagal tone strengthens. It also drops acutely under mental load — a period of complex decision-making or public speaking is a specifically documented trigger — which is exactly the kind of ordinary stress most people experience on an unremarkable Tuesday.

Breathing rate adds its own distortion on top of all that. Very slow breathing in particular can inflate certain frequency-based readings independent of anything else about your autonomic state changing, which is one reason a single night’s number is a weak thing to react to on its own.

None of this is a reason to distrust the concept. It’s a reason to expect the graph to be noisy, and to stop treating one low morning as an alarm rather than as data.

Is a higher number always the goal?

Inside a normal, stable heart rhythm, yes — broadly, more variability tracks with more parasympathetic capacity to modulate the heart, and that’s the physiology behind most “raise your HRV” advice, including the genuine link to aerobic training.

But that rule has a real edge, and it’s worth stating plainly rather than glossing over: an optical sensor cannot distinguish healthy autonomic swing from a rhythm that has simply gone irregular for a different reason. Atrial fibrillation is one of the most common heart-rhythm disorders, and it causes the heart to beat irregularly — sometimes faster than normal, sometimes with no noticeable symptoms at all. To a wrist sensor counting the gaps between beats, that irregularity can look like a variability signal. It isn’t the same thing being measured, and it isn’t good news dressed up as a high score.

The two get told apart the way they’ve always been told apart: with an EKG, which reads the heart’s electrical activity directly, or a longer recording like a Holter or event monitor if the pattern needs to be caught over a full day rather than a single reading. A wellness score, however confident the app is about it, isn’t built to make that distinction, and isn’t trying to.

Why this is graded moderate

The measurement science here is solid. How ECG and optical sensors differ, which metrics hold up under that difference, what moves the number physiologically — that’s documented, methods-focused research, not a marketing claim dressed up as one.

What keeps this at moderate rather than strong is the leap from that measurement science to the “raise your number, improve your health” framing most consumer apps sell. The underlying physiology supports a general association between higher variability and greater autonomic flexibility. It does not license treating a single day’s app score as a diagnostic reading, in either direction, and the research this guide draws on is explicitly about method and mechanism rather than about validating any specific product’s wellness score.

What’s actually worth doing with the number

Read it as a trend, not a snapshot. The same device, at roughly the same time of day, under roughly similar conditions, tracked across weeks rather than mornings, is the version of this number that can tell you something — whether your recent training load, sleep, or stress load is trending in a direction worth noticing.

Don’t compare it across people, and be cautious comparing it across apps, because the metric, the algorithm, and the recording window all move the absolute value independently of your physiology. And don’t ask the number to do a job it was never built for: if your pulse itself ever feels irregular rather than simply fast, slow, or noisy on a graph, that’s a conversation for a clinician and an ECG, not for a trend line.