Can I convert SDNN to RMSSD?

No. SDNN SDNN SDNN The standard deviation of all NN intervals in a recording, in milliseconds. It measures overall variability, and its value depends strongly on how long the recording lasts. More about SDNN and RMSSD RMSSD RMSSD Root mean square of successive differences: an HRV metric built from how much each heartbeat interval differs from the next, in milliseconds. It mainly reflects vagal activity. More about RMSSD ask different questions of the same heartbeats, and the ratio between them changes with the rhythm, so no formula turns one into the other. The only conversion in this family is SD1 SD1 SD1 and SD2 Two measures of the spread of dots in a Poincaré plot, in milliseconds. SD1 is the width and tracks short-term variability, like RMSSD; SD2 is the length and tracks overall variability. More about SD1 and SD2 , the width of a Poincaré plot Poincaré plot Poincaré plot A scatter plot of each heartbeat interval against the one after it, in milliseconds on both axes. The shape of the cloud of dots shows short-term and overall variability at a glance. More about Poincaré plot , which is RMSSD divided by the square root of two.

Two questions about the same gaps

Take the gaps between beats, in milliseconds. SDNN is their standard deviation: how far they spread around their average, whether that spread comes from fast swings or slow drifts. RMSSD takes each gap minus the one before, squares those differences, averages them and takes the square root: it only sees the step from one beat to the next. A review of HRV norms puts it plainly: RMSSD is more strongly shaped by the parasympathetic parasympathetic Parasympathetic nervous system The "rest and digest" branch of the autonomic nervous system. It slows the heart, mainly through the vagus nerve, and drives most of the beat-to-beat variation in HRV. More about Parasympathetic nervous system branch than SDNN is.

Watch them come apart

SDNN and RMSSD on the same beats

Both by the 1996 Task Force definitions

One slow breath Slow drift Steady beats One extra beat

Gap before each beat 750 to 1050 ms

SDNN

76.7 ms

RMSSD

45.9 ms

SDNN ÷ RMSSD

1.67

SD1

32.5 ms

Beat-to-beat swings from breathing: both measures see them.

Here SDNN is 76.7 ms and RMSSD is 45.9 ms, a ratio of 1.67. For one slow breath the ratio was 1.67, so this series sits close to it. The ratio changes with the pattern of the beats, which is why no formula converts one into the other. SD1, by contrast, is always RMSSD divided by the square root of two: 32.5 ms here.

Your own gaps, in ms (3 to 60, separated by commas) 900, 840, 800, 790, 820, 880, 950, 1000, 1010, 980, 930, 880

The presets and the text box work when JavaScript is on.

A line chart of 12 gaps between heartbeats with a dashed line at their mean, and readouts for SDNN, RMSSD, their ratio and SD1. Preset buttons load a slow breath, a slow drift, steady beats and one extra beat; a text box lets the reader type their own gaps. For the slow breath, SDNN is 76.7 ms and RMSSD is 45.9 ms. For the slow drift SDNN stays similar while RMSSD falls to a fraction; for the extra beat RMSSD jumps far more than SDNN.

Try the presets. For one slow breath, the gaps swing up and down and both measures see the swing, with SDNN about 1.7 times RMSSD. For a slow drift with a similar overall spread, SDNN stays about the same while RMSSD falls to a fraction, because each step is tiny. For one extra beat, RMSSD jumps far more than SDNN, because one early beat and its pause make two huge steps. Steady beats with jitter put RMSSD above SDNN. A single exchange rate cannot describe all four.

Type your own gaps into the box and the numbers recompute. If you export beat intervals from a chest-strap app, you can paste a short run of them here.

Why recording length makes it worse

SDNN also grows with the length of the recording, because longer recordings include slower swings. The same review states that it is inappropriate to compare SDNN values calculated from epochs of different length, and that 24-hour, short-term and ultra-short-term values are not interchangeable. That is why the site’s reference data keeps 24-hour rows apart from 5-minute rows, and why 24-hour SDNN runs above 100 ms in healthy adults while 5-minute SDNN, pooled across studies, averages about 50 ms.

So “45 ms SDNN from an Apple Watch” and “45 ms RMSSD from a ring” differ in the statistic, the recording length and the time of day. Apple’s documentation describes its HealthKit HRV type as SDNN; a 2025 validation study notes that Oura, WHOOP, Garmin and Polar all report RMSSD from sleep. Why Apple Watch HRV is different goes through the consequences.

The one conversion that works

Ciccone and colleagues showed in 2017, both mathematically and on data, that RMSSD and SD1 are the same metric: SD1 equals RMSSD divided by √2. If your app shows SD1 and another shows RMSSD, multiply SD1 by 1.414 and you have RMSSD exactly. The widget shows SD1 next to the other two so you can see it track RMSSD whatever the preset. The glossary entry on SD1 and SD2 explains the plot.

What to do with your own data

• Track the measure your device gives and compare it only with its own history.
• If you look up norms, match the measure and the recording length. The HRV checker asks for both.
• If you switch from a device that reports SDNN to one that reports RMSSD, start a new series. Building your baseline takes about three weeks.