Should I use RMSSD or SDNN?

For day-to-day personal tracking, RMSSD is more responsive and is what your wearable reports. SDNN is a clinical 24-hour Holter metric used in risk stratification.

RMSSD vs SDNN: which HRV number means what

HRV has dozens of metrics. For consumer use the two you'll see most often are RMSSD and SDNN. They measure related but different things.

At a glance#

Metric	What it captures	Typical range (adult, rest)	Dominant input	Where it's used
RMSSD	Beat-to-beat short-term variability	20–80 ms	Parasympathetic (vagal)	Consumer wearables (sleep HRV)
SDNN	Total variability across a window	100–200 ms over 24 h	Mixed (vagal + sympathetic + baroreflex)	Clinical 24-h Holter
Time to stabilise	1–5 minutes clean data	5+ min minimum; 24 h ideal	—	—
Cross-device comparability	Algorithm-dependent	Don't compare brands	—	—

RMSSD#

RMSSD is the root mean square of successive differences between heartbeats. Mathematically, take the difference between each pair of consecutive R-R intervals, square them, average, then take the square root.

What it reflects: short-term, beat-to-beat variability. This is dominated by parasympathetic (vagal) input.

Typical resting values: 20–80 ms in healthy adults, varying with age, fitness, and time of measurement.

Why wearables use it: RMSSD stabilises quickly (1–5 minutes of clean data is enough), is robust to short measurement windows, and tracks parasympathetic activity, which is what most people want to know.

SDNN#

SDNN is the standard deviation of all N-N intervals (normal-to-normal beats) across a measurement window. It captures the total spread of heart-beat intervals over the window.

What it reflects: overall heart-rate variability across the entire recording, including both short-term (vagal) and longer-term (sympathetic, baroreflex, respiration) components.

Typical resting values: Highly window-length dependent. Over a 24-hour recording, healthy adults usually report SDNN in the 100–200 ms range. Over a 5-minute recording, much lower.

Where it's used clinically: 24-hour SDNN is a long-standing prognostic measure post-myocardial-infarction.

Which one your wearable shows#

Oura, Whoop, Garmin, Eight Sleep, Apple Watch: RMSSD (often relabelled or transformed)
24-hour clinical Holter monitors: typically report SDNN
HRV apps for daily morning readings (Elite HRV, HRV4Training): typically RMSSD

Garmin shows the "transformed" case clearly: alongside its HRV Status trend, the same beat-to-beat data feeds a 0–100 Stress score and Body Battery, a composite energy-reserve score, and neither of those is an HRV value in milliseconds.

If a number you're tracking is in the range 20–100 ms, it's almost certainly RMSSD or a transformation thereof. If it's 100–200+ over many hours, it's likely SDNN.

Practical implications#

For monitoring your own state day-to-day, RMSSD is the more useful metric: it responds faster and is dominated by the autonomic balance you can shift with sleep, alcohol, training load, and stress.

SDNN is useful for clinical risk stratification but less actionable as a daily tracker because it requires longer windows.

Common confusions#

Don't compare RMSSD numbers across different devices. Algorithms differ; even with identical inputs the reported number can vary by 10–20%.
Don't compare "HRV score" between brands. Each vendor's score is a different transformation of RMSSD.
Don't over-interpret a single morning reading. RMSSD on a single morning has a wide noise band. Three-day or seven-day rolling averages are the level at which trends become meaningful.

A reasonable read#

Pick one device, stick with it, and look at trends over weeks. The exact number matters less than the direction, and the direction matters more in the context of your sleep, training, and life, not in isolation.