HRV.Understand your readings.

Heart rate and HRV are different

Heart rate counts beats per minute. Heart rate variability (HRV) describes how much the time between successive beats varies. Even at a steady average heart rate, the intervals need not be identical.

Apple Watch can record HRV in HealthKit as SDNN, the standard deviation of normal heartbeat intervals, in milliseconds. WatchMuse reads those available SDNN records; it does not reconstruct them from occasional heart-rate readings. Apple’s SDNN definition.

Apple describes using its optical heart sensor for HRV. Sensor acquisition and WatchMuse’s comparison are separate steps: the formulas here explain WatchMuse, not Apple’s internal signal-processing algorithm. How Apple Watch measures heart rate.

Why HRV can offer context—not certainty

Researchers study changes in HRV during mental stress because heartbeat timing reflects autonomic regulation. A physiological association does not make one HRV reading a reliable label for how someone feels. A study of HRV during controlled stress tasks.

WatchMuse therefore compares a reading with your own history rather than giving everyone the same “normal” HRV. A lower ratio falls into a higher reference band in this algorithm. That is a rule for presenting a change, not proof that psychological stress caused it.

First, establish your personal baseline

WatchMuse retrieves up to 120 SDNN records from the previous 14 days, newest sample end time first. Values must be finite and greater than 0, no more than 1,000 ms, with valid non-future timestamps less than 14 × 24 hours old. Duplicate timestamps are removed. These are input-validation rules, not medical normal ranges.

There must be at least four valid records with distinct times: the latest reading plus at least three earlier readings. The latest is the current value. The median of the rest becomes the personal baseline. With an even number of values, the median is the mean of the two middle values.

For example, historical readings of 40, 50 and 60 ms give a baseline of 50 ms. The requirement is three earlier records, not three different days. At most 119 records contribute to the baseline, each with equal weight. Rejected records are not replaced by an extra query.

This HRV query retains the actual HealthKit source, but does not restrict input to Apple Watch or exclude manually entered records. It does not separate sleep, daytime or post-exercise baselines. Differences in source and measurement context can therefore affect the comparison.

Then, compare the latest value with the baseline

Let B be the personal median and C the latest valid HRV. Divide C by B to obtain the ratio R. WatchMuse applies these conditions in order:

With B = 50 ms: C = 45 ms gives R = 0.90, classified as relaxed; C = 32.5 ms gives R = 0.65, balanced; C = 32 ms gives R = 0.64, elevated. These are fictional arithmetic examples, not recommended health targets.

The 0.90 and 0.65 boundaries are WatchMuse’s own heuristic rules, not Apple’s stress scale or clinically validated diagnostic cutoffs. The algorithm calculates this reference locally on the watch; it is not an AI emotion-recognition model.

B = median(previous valid HRV readings)
R = latest valid HRV ÷ B

R ≥ 0.90          → relaxed
0.65 ≤ R < 0.90   → balanced
R < 0.65          → elevated

Try an example

Move a slider or enter a number to see which band the rule selects. The labels describe only the example calculation.

Why your HRV measurement time matters

Opening WatchMuse or refreshing its data does not force Apple Watch to take a new HRV measurement. The app checks the records already available. watchOS controls background execution, so checking for new records does not start a new measurement.

A saved estimate shows when the HRV was measured, not when you last opened it. At 30 minutes old, a result is labelled historical; its numerical value and original timestamp are unchanged.

Without enough valid readings, WatchMuse cannot calculate a baseline or a new estimate. It can still display an available HRV reading or a previous valid estimate with its original measurement time. If no records are returned, check both Health access and whether any measurements are available. An empty result does not necessarily mean access was denied.

What the result cannot tell you

Relaxed, balanced and elevated are reference bands, not verified feelings. They cannot tell whether you are happy, anxious or angry, diagnose an illness, or explain why a reading changed.

The algorithm described here has no clinical validation or accuracy score for detecting a person’s emotions. A personal median limits the influence of a few extreme values, but does not make differently timed or differently sourced measurements directly comparable.

Treat the result as context alongside how you actually feel. Do not use it alone for diagnosis, medication decisions or judging an emergency.

About these methods

Algorithm description based on the WatchMuse source dated September 7, 2026; this page was revised September 8, 2026. The source describes the rules but does not establish that every current App Store build has identical behavior.

The Apple links above define the measurement and its system context. The research link illustrates why HRV is studied in relation to stress; it does not validate WatchMuse’s thresholds. This page covers HRV and its stress-reference calculation, not reminders, artwork or other app features.