Acute Stress Marker · RMSSD explained
The HRV number that moves first when something's off. What RMSSD actually measures, how your Apple Watch records it, and why Body Insights reads it as the acute stress marker - not a score to chase.
Three nights before a cold lands, before a flare you couldn't yet name - something in the rhythm of your heartbeat narrows. The intervals between beats stop varying the way they used to. The body has registered the strain. You haven't.
That narrowing has a name. RMSSD - the root mean square of successive differences between heartbeats - is the HRV metric that captures it earliest. Of the standard short-term HRV measures, it responds most quickly to acute stress, which is why Body Insights reads it as your acute stress marker.
This page is the long-form version. For the daily read, see stress monitoring. To understand what's underneath it, keep going.
Your heart doesn't beat like a metronome. The interval between one beat and the next changes slightly with every breath, every shift in posture, every passing thought. That variation is heart rate variability, and the size of the beat-to-beat variation is what RMSSD captures.
The calculation starts with each pair of consecutive heartbeats. It finds the difference between them and squares it. Then it averages those results and takes the square root. The result is reported in milliseconds. A higher RMSSD means more beat-to-beat variation. A lower RMSSD means the intervals are clustering close together - the heart is being held on a shorter leash.
That leash is the vagus nerve. RMSSD is the cleanest non-invasive read on parasympathetic (vagal) tone we have. When the parasympathetic system is engaged - rest, digestion, recovery - RMSSD goes up. When it pulls back - illness, late dinner, alcohol, real stress, hormonal shifts - RMSSD comes down.
RMSSD isn't a wearable-era invention. A joint cardiology Task Force defined it as a standard short-term HRV measure in 1996. The group included European and North American societies for cardiology and heart pacing. Their paper is still a reference for HRV measurement and clinical use.
The Task Force named three time-domain HRV measures worth tracking: SDNN, RMSSD, and pNN50. RMSSD responds fastest to short-term parasympathetic changes. It is also robust in short recording windows. That makes it suitable for overnight wearable data.
That's why Body Insights anchors on it. The math has been settled for three decades. The wrist is new; the metric isn't.
When the immune system activates - for any reason, viral, bacterial, inflammatory - one of the earliest responses is a suppression of vagal tone. The inflammatory cascade releases cytokines. They act on the brainstem and the heart, which reduces parasympathetic activity. The heart's variability narrows.
This happens before the fever. Before the sore throat. Often a night or two before any conscious symptom. Which is why a sustained drop in RMSSD against your own baseline is one of the most useful early signals a wearable can surface.
The same pattern shows up in non-illness contexts. A poor night of sleep. A heavy training day. A perimenopausal hot-flash night. A flare day in a chronic condition. Each one pushes RMSSD down, because each one calls on the parasympathetic system to back off.
That's why we treat RMSSD as the acute marker. It moves on a timescale of days. The slower trends - fitness, age, structural change - show up in SDNN and in your longer baseline. The thing that changed this week shows up here first.
Free for core features. No card needed.
Everything below is what an honest write-up of RMSSD on a wrist wearable should tell you - including where the wrist falls short.
Body Insights doesn't measure your heart. Your Apple Watch does. We read what it wrote to Apple Health and translate it.
Apple Watch records beat intervals during sleep and quiet periods. Apple Health stores this stream as HKHeartbeatSeries. We read that stream rather than Apple's default SDNN summary, which means the RMSSD you see is computed from raw data - not derived from another statistic.
The cleanest read is during sleep, when posture and breathing are stable and the parasympathetic system is doing its primary work. We compute RMSSD across the overnight window for consistency night to night.
Wrist photoplethysmography can record intervals that are not real heartbeats. Motion artefacts and signal gaps can cause this. We discard intervals outside a physiologically plausible range before the calculation runs.
After two to three weeks of nights, we have enough of your own data to build a personal baseline distribution. Every subsequent night is read against that distribution - not against a textbook number.
Personal HRV distributions are often skewed, especially in chronic illness. We use percentile rank against your own history and z-scores. This flags an unusual night even when your distribution is not bell-shaped.
You don't have to read the number. The morning insight is a sentence: "Your acute stress marker dropped against your baseline overnight. Today, take it slow." The score is there if you want it.
A chest-strap ECG is the consumer reference standard for HRV. It reads the heart's electrical signal directly. In a controlled short test, it is more accurate than any wrist device on the market. That includes Apple Watch.
The wrist trades that single-reading precision for two things the chest strap can't give you: it stays on all night, and it stays on every night, without a ritual. Recent validation work compares Apple Watch overnight RMSSD with chest-strap reference devices. It uses the same raw heartbeat-series data that we read. The results show good agreement in trends across weeks. Single-night agreement is reasonable but noisier - especially when wrist contact is poor or movement is high.
The practical consequence: if you want a competition-grade measurement of one morning's RMSSD, use a chest strap and the protocol that comes with it. If you want a sustained signal that catches an illness coming three nights early, or a flare building over a week, the wrist wins - because it actually shows up.
Body Insights leans into that trade. We read RMSSD as a pattern across nights, not as a verdict on any single one.
Three different things often called "HRV." Worth distinguishing, because they answer different questions.
| RMSSD | SDNN | Apple HRV (default) | |
|---|---|---|---|
| What it captures | Beat-to-beat variation | Overall variance across window | SDNN over short reads |
| Reflects | Parasympathetic (vagal) tone | Mixed autonomic activity | Mixed autonomic activity |
| Best for | Acute / short-term change | Longer-term structural change | Casual trend chart |
| Window needed | Short (minutes) | Longer (ideally 24 hr) | Whatever the watch sampled |
| In Body Insights | Acute stress marker | Background context | Available but not the primary read |
The Task Force named both RMSSD and SDNN as standard time-domain measures. They're not redundant - they answer different questions. RMSSD is what we surface when something's changed this week.
Useful to know before you read a low number and worry.
Poor sleep, alcohol, late dinners, infection, and intense training can lower RMSSD. So can perimenopausal hot flashes, anxiety, dehydration, or autonomic dysfunction.
Consistent sleep and aerobic fitness can raise RMSSD. So can slow breathing before bed, hydration, or a restorative night. Some of these are choices; some aren't. None of them are an instruction.
RMSSD is not a measure of how hard you're working, how fit you are in absolute terms, your VO2max, or your readiness as a single concept. It's a window onto vagal tone. Other inputs carry the rest.
Single-night variation is real. One night below baseline often means nothing - you slept poorly, you had wine. A sustained drop across multiple nights is the signal worth reading.
You will not find anywhere on this page the phrase "boost your RMSSD." That framing is wrong on two counts.
First, RMSSD is a window onto your nervous system, not a muscle to train. The things that move it sustainably are sleep, fitness, and not being chronically ill - none of which are gameable. Chasing the number creates a false sense of discipline. Wearable culture has made that habit common. It does not survive contact with a chronic condition.
Second, and more practical: when RMSSD drops, the appropriate response is usually to do less, not to drill protocols designed to push it back up. The drop is the body telling you it's busy. The intervention is rest, food, and patience. The number recovers when the underlying state recovers - not before.
So we surface the signal. We tell you what changed. We don't hand you a score to optimise.
The root mean square of successive differences between heartbeats. A time-domain HRV measure that reflects vagal (parasympathetic) tone, named as a standard by the Task Force in 1996.
No. Apple Watch reports SDNN by default. Body Insights computes RMSSD from the raw beat-to-beat data Apple Watch records overnight.
For trends across one to four weeks, good. A single reading is less precise than a chest-strap ECG. Wrist light sensors are noisier than a chest electrical signal, especially during movement.
There isn't one. Healthy adults range roughly 20-80 ms, falling with age and rising with fitness. Chronic-illness baselines often run lower. Your own baseline is what matters.
Inflammatory signalling suppresses vagal tone early in immune activation - often a night or two before symptoms. The drop is a leading indicator, not a diagnosis.
Your morning insight names it when it changes. Detail views show the overnight trend against your own baseline. For the daily read, see stress monitoring.
On-device. No account. No subscription required. No data sold. Ever.
Free for core features. No card needed.