The sources in parentheses are links to the full references in the research review.
Definition
Heart rate variability (HRV) is the variation in the time between successive heartbeats (RR intervals, called NN intervals once abnormal beats are removed). HRV and heart rate are different things: a heart rate of 60 beats per minute does not mean that every interval is exactly 1.000 second (Task Force, 1996; Shaffer & Ginsberg, 2017).
Beat-to-beat intervals are shaped at the same time by the sympathetic nervous system, the parasympathetic (vagal) nervous system, breathing and the baroreflex that regulates blood pressure. HRV is therefore the sum of several mechanisms, and no single indicator directly represents "autonomic state" (Task Force, 1996; Shaffer & Ginsberg, 2017; Billman, 2013). Psychophysiological research often uses the vagally mediated components, such as RMSSD and HF, to estimate cardiac vagal regulation (Laborde et al., 2017).
The neurovisceral integration model proposes that resting HRV is linked to prefrontal regulation of emotional and autonomic responses, and a meta-analysis of neuroimaging studies has identified brain regions associated with HRV. This evidence is correlational (Thayer & Lane, 2000; Thayer et al., 2009; Thayer et al., 2012).
Measurement methods
| Method | What it measures | Use and limits |
|---|---|---|
| Electrocardiogram (ECG) | RR intervals from the position of R waves in the ECG waveform | The methodological reference. Beats can still be missed or misdetected, so the data need checking (Task Force, 1996) |
| ECG chest strap | RR intervals | In a validation study with 10 participants, the Polar H10 detected heartbeats with about 99.6% accuracy, and about 99% during high-intensity activity (Gilgen-Ammann et al., 2019) |
| Photoplethysmography (PPG): watches, armbands, rings | Pulse-to-pulse intervals; strictly speaking, pulse rate variability | Most studies consider it acceptable at rest. Agreement with ECG drops clearly during physical activity and in some psychological stress conditions (Schäfer & Vagedes, 2013; Georgiou et al., 2018) |
Ectopic beats, missed detections and intervals distorted by movement must be checked and corrected before calculation. Otherwise indicators such as RMSSD can be pushed too high or too low. Reports should also state the device, sensing method, recording length and correction method (Task Force, 1996; Laborde et al., 2017).
Recording length
The 1996 international standards divide HRV into short-term (5-minute) and long-term (24-hour) recordings. The two differ in their values and physiological meaning and cannot be compared with each other. Frequency-domain analysis needs enough data, and 5 minutes is the standard length for short-term analysis (Task Force, 1996). Ultra-short recordings under 5 minutes are not interchangeable with reference values for 5-minute or 24-hour recordings (Shaffer & Ginsberg, 2017).
When comparing the same person "before" and "after", both recordings should ideally have the same length, posture and breathing conditions (Laborde et al., 2017).
Common indicators
| Indicator | Calculation | Usual interpretation | Caveats |
|---|---|---|---|
| SDNN | Standard deviation of all NN intervals over a period | Overall variability from all sources | Depends heavily on recording length; 24-hour and 5-minute SDNN cannot be compared |
| RMSSD | Root mean square of successive differences between NN intervals | Short-term, vagally mediated variation | Usually less affected by breathing rate than HF, and widely used for short recordings; still sensitive to erroneous intervals |
| HF | Power in the 0.15–0.40 Hz band | The respiratory band, mainly reflecting vagal regulation of the heart | When breathing is slower than 9 breaths a minute, the breathing-related oscillation moves into the LF band, so a fall in HF does not mean a fall in vagal activity |
| LF | Power in the 0.04–0.15 Hz band | Influenced by sympathetic and parasympathetic activity and the baroreflex together | Once widely regarded as a sympathetic index, which most of the literature now does not support; rises sharply during slow breathing |
| LF/HF | LF power divided by HF power | Formerly called the "sympathovagal balance" | Debated; see the next section |
Indicator definitions follow (Task Force, 1996; Shaffer & Ginsberg, 2017; Laborde et al., 2017; Billman, 2013).
The LF/HF debate
LF power reflects sympathetic and parasympathetic activity, the baroreflex and breathing all at once, so the LF/HF ratio cannot accurately represent the balance between sympathetic and vagal activity (Billman, 2013). Stress studies often report lower HF and higher LF, but methods for inducing stress and for measurement vary widely between studies, and LF/HF cannot serve as a personal stress score (Kim et al., 2018; Billman, 2013).
When breathing rate changes, the split between LF and HF changes with it. During slow breathing, LF rises and LF/HF increases because the breathing rhythm falls into the LF band, not because sympathetic activity has increased (Lehrer & Gevirtz, 2014; Laborde et al., 2017).
Effects of breathing
The heart beats slightly faster during inhalation and slightly slower during exhalation. This pattern is called respiratory sinus arrhythmia (RSA), and it is the main source of power in the HF band (Zaccaro et al., 2018; Lehrer & Gevirtz, 2014). When breathing slows to about six breaths a minute, the breathing rhythm comes close to the rhythm of the baroreflex and heart rate oscillations become very large. Resonance breathing in HRV biofeedback makes use of this (Lehrer & Gevirtz, 2014).
A meta-analysis of 223 studies found that vagally mediated HRV was higher during slow-paced breathing, after a single session and after multiple training sessions. A large part of the increase during practice, however, comes from the change in breathing rate itself, so it should be interpreted separately from resting measurements taken after practice (Laborde et al., 2022). A small study in Taiwan comparing breathing at 6 and 12 breaths a minute also observed higher total HRV power during slow breathing (Tsai et al., 2015).
What this means for a relaxation session
Within one session, segments where the voice guidance leads slow breathing and segments of natural breathing reflect different breathing conditions. Mandala Hub measures change by comparing ○ from the quiet wait before the session opens with ● during the journey. The best during the session ★ is shown as a level on its own. See Measurement and research.
Factors that affect measurement
| Factor | Notes |
|---|---|
| Age and sex | Baseline values decline with age and differ between men and women, so comparing with yourself is more appropriate (Shaffer & Ginsberg, 2017) |
| Posture, time of day, recent activity | Values differ lying down, sitting and standing; exercise, eating, caffeine, alcohol and lack of sleep all have an effect (Laborde et al., 2017) |
| Body movement and talking | Add erroneous intervals; optical sensing is affected more (Schäfer & Vagedes, 2013; Georgiou et al., 2018) |
| Psychological stress | Changes HRV itself and can also reduce agreement between optical measurement and ECG (Kim et al., 2018; Schäfer & Vagedes, 2013) |
| Sensing method | ECG-based and optical sensors do not measure exactly the same thing (Schäfer & Vagedes, 2013; Gilgen-Ammann et al., 2019) |
State of research
- Most studies observe changes in HRV under stress, most often a fall in HF. HRV can be one objective indicator in stress research, but it is not a diagnostic tool (Kim et al., 2018).
- More contact with green space is associated with higher HF and lower LF. The evidence comes mainly from observational studies (Twohig-Bennett & Jones, 2018).
- Studies of immersive natural scenes often measure HRV at the same time. So far most are small studies of a single exposure (Anderson et al., 2017; Riches et al., 2021).
- Meta-analyses of HRV biofeedback show improvements in self-reported stress and anxiety, but the outcomes are mostly self-reported and the effect is smaller when compared with active control groups (Goessl et al., 2017; Lehrer et al., 2020).
Common misconceptions
| Common claim | A more accurate statement |
|---|---|
| Higher HRV is always better, and you can compare yours with other people's | Individual differences are large and depend on age, sex and measurement conditions. Compare with your own past values under the same conditions (Shaffer & Ginsberg, 2017) |
| LF/HF is a stress index | LF is not a pure sympathetic index, and the meaning of the ratio is debated (Billman, 2013) |
| HRV from a watch is the same as HRV from an ECG | Optical sensors measure pulse rate variability. The two are close at rest and diverge more during activity or stress (Schäfer & Vagedes, 2013; Georgiou et al., 2018) |
| Higher HRV during a breathing exercise means your physical state has already changed | Much of the increase during practice comes from the breathing itself; look at resting measurements taken after practice (Laborde et al., 2022) |
| HRV from different apps or different recording lengths can be compared directly | Recording length, calculation and sensing method differ, so the values are not interchangeable (Task Force, 1996; Shaffer & Ginsberg, 2017) |
| HRV can tell whether you are stressed or ill | Research mostly shows group-level associations, and a single reading is affected by many factors. It cannot be used for diagnosis (Kim et al., 2018; Laborde et al., 2017) |