The Science of Balance

This page separates two things that are easy to confuse: what peer-reviewed research says about heart rate variability (HRV) and what Feelmo's proprietary Balance score displays.

What is HRV?

The time from one heartbeat to the next is not perfectly constant. Heart rate variability (HRV) is the name for variation in those intervals. Researchers calculate HRV in several ways:

• Time-domain measures summarize how much intervals vary over time. RMSSD, for example, emphasizes short-term beat-to-beat differences.
• Frequency-domain measures describe how much variation appears in different frequency bands. High-frequency power is strongly affected by breathing.

The autonomic nervous system contributes to HRV. Its sympathetic branch helps mobilize the body, while its parasympathetic branch supports functions including slowing the heart. Under controlled conditions, RMSSD, high-frequency power, and related measures can reflect respiratory cardiac vagal modulation — the breathing-linked parasympathetic influence on the heart.

This is narrower than saying that HRV measures the whole autonomic nervous system. The 2026 scientific guideline states that HRV should not be used as a specific marker of sympathetic nerve signals sent to the heart ("cardiac sympathetic outflow") or of "sympathovagal balance." It also advises care with "vagal tone," a broader claim about the general level of vagus-nerve activity (Carter et al., 2026). In particular, the low-frequency/high-frequency ratio (LF/HF) should not be read as a simple accelerator-versus-brake gauge.

Why does measurement context matter?

HRV changes with breathing rate and depth, posture, physical activity, time of day, sleep stage, recent exercise, alcohol, medication, illness, age, and many other factors. Recording duration, sensor type, signal processing, missing beats, movement, and irregular rhythms can also alter the result (Task Force, 1996; Laborde et al., 2017; Carter et al., 2026).

Records are easier to interpret together when the metric and measurement conditions are reasonably similar, although this does not make them perfectly comparable. Higher is not automatically better: measurement error and irregular rhythms can also create large apparent variation. A single change rarely has one certain explanation.

What is established research, and what is Feelmo's product display?

Published research describes general HRV methods and group-level findings. Feelmo's Balance display is not a standardized clinical measure from those papers.

Feelmo uses proprietary on-device processing to summarize HRV-derived information available through Apple Watch / HealthKit and patterns in the user's own past. Detailed processing and display rules are not public.

The cited HRV studies did not test Feelmo's exact algorithm. They therefore do not establish that 0–100 is a clinical scale, that a particular score means a particular physiological state, or that using the app improves health. Research activity or collaboration is not the same as completed peer-reviewed validation of the product score.

Why compare with your own baseline?

HRV differs greatly between people and generally changes with age (Umetani et al., 1998). Looking at a person's own past trend can reduce misleading comparisons with other people, but it cannot establish health or a cause.

It does not remove changes in breathing, sleep, medication, illness, sensor conditions, or routine. "Higher than usual" and "lower than usual" are descriptions of the app's trend, not diagnoses of being healthy, recovered, calm, or stressed.

The six companions are product characters. They do not mean that Feelmo has detected the named emotion or condition, and their display rules are not public.

What do stress studies show?

A meta-analysis reported that psychological stress is associated, on average, with lower values of several HRV measures (Kim et al., 2018). Brain-imaging research has also found group-level associations between some vagally mediated HRV measures and regions involved in regulation (Thayer et al., 2012).

These results overlap substantially between individuals and situations. They do not allow an app to infer one person's stress, emotion, or brain activity from one measurement. For more on that distinction, see The Brain–Heart Connection.

Why does context still matter at rest or during sleep?

Optical wrist measurements are especially vulnerable to movement. Sitting quietly or sleeping may reduce large motion, but that does not make every reading accurate or equivalent to an electrocardiogram.

Sleep is not a laboratory-controlled condition, however. Sleep stage, brief awakenings, breathing, posture, alcohol, illness, sensor contact, and the timing and number of Apple Watch samples vary from night to night. Nighttime HRV is related to sleep physiology at group level (Trinder et al., 2001; Stein & Pu, 2012), but it is not a direct measure of sleep quality or recovery.

What does slow-breathing research show?

Slow breathing changes the mechanical and reflex processes that shape heart rate. The baroreflex is the body's short-term reflex for stabilizing blood pressure; at some breathing rates, breathing-related heart-rate oscillations and this reflex can reinforce one another.

A 2022 systematic review and meta-analysis included 223 studies and found average increases in vagally mediated HRV during slow breathing, immediately after a single session, and after multi-session interventions (Laborde et al., 2022). Around six breaths per minute is common in this literature, but the rate that produces the largest oscillation differs between people.

An increase in HRV during paced breathing partly reflects the breathing pattern itself. It does not, by itself, prove relaxation, recovery, better sleep, treatment of a disorder, or effectiveness of Feelmo's breathing sessions.

What about light and exercise?

Light exposure helps set circadian timing — the body's roughly 24-hour rhythm — and is related to sleep and mood (Blume et al., 2019). Exercise interventions have also been associated with changes in resting HRV in meta-analysis (Sandercock et al., 2005).

These are population-level findings, not promises about an individual day. Feelmo may summarize an observational association between habit records you logged and Balance. Proprietary processing details are not public, and the display does not establish a causal effect.

What are the limits of Apple Watch and HealthKit?

Apple Watch uses photoplethysmography (PPG), an optical method that detects pulse-wave changes at the wrist. A clinical electrocardiogram (ECG) records the heart's electrical activity. A study in healthy participants found useful agreement for selected HRV measurements during relaxation and mental-stress conditions (Hernando et al., 2018), but that study does not validate every person, activity, sleep setting, watch generation, or Feelmo's score.

HealthKit data are not a continuous clinical ECG. Apple controls when and how samples are generated, and movement, fit, skin contact, recording length, breathing, hardware, software, and missing or irregular beats can affect comparability. Wearable results must be interpreted in the context of these limitations (Carter et al., 2026).

Comparing your own records: the N-of-1 idea

An N-of-1 approach focuses on one person's own records. Feelmo Premium's habit association patterns summarize observational relationships between habits the person logged and Balance; proprietary processing details are not public.

This is observational. Habits occur in different contexts, and unmeasured factors may affect both a habit and the records shown. The feature summarizes an association; it cannot prove a causal effect.

References

1. Carter JR, et al. Guidelines for rigor and reproducibility of heart rate variability within human cardiovascular research. American Journal of Physiology-Heart and Circulatory Physiology. 2026. doi:10.1152/ajpheart.00041.2026. PMID: 42495990.
2. Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology. Heart rate variability: standards of measurement, physiological interpretation and clinical use. Circulation. 1996;93(5):1043–1065.
3. Laborde S, Mosley E, Thayer JF. Heart Rate Variability and Cardiac Vagal Tone in Psychophysiological Research: Recommendations for Experiment Planning, Data Analysis, and Data Reporting. Frontiers in Psychology. 2017;8:213.
4. Laborde S, Allen MS, Borges U, et al. Effects of voluntary slow breathing on heart rate and heart rate variability: A systematic review and a meta-analysis. Neuroscience & Biobehavioral Reviews. 2022;138:104711. doi:10.1016/j.neubiorev.2022.104711. PMID: 35623448.
5. Trinder J, Kleiman J, Carrington M, et al. Autonomic activity during human sleep as a function of time and sleep stage. Journal of Sleep Research. 2001;10(4):253–264.
6. Stein PK, Pu Y. Heart rate variability, sleep and sleep disorders. Sleep Medicine Reviews. 2012;16(1):47–66.
7. Kim HG, Cheon EJ, Bai DS, Lee YH, Koo BH. Stress and Heart Rate Variability: A Meta-Analysis and Review of the Literature. Psychiatry Investigation. 2018;15(3):235–245.
8. Thayer JF, Åhs F, Fredrikson M, Sollers JJ, Wager TD. A meta-analysis of heart rate variability and neuroimaging studies: implications for heart rate variability as a marker of stress and health. Neuroscience & Biobehavioral Reviews. 2012;36(2):747–756.
9. Umetani K, Singer DH, McCraty R, Atkinson M. Twenty-four hour time domain heart rate variability and heart rate: relations to age and gender over nine decades. Journal of the American College of Cardiology. 1998;31(3):593–601.
10. Hernando D, Roca S, Sancho J, Alesanco Á, Bailón R. Validation of the Apple Watch for Heart Rate Variability Measurements during Relax and Mental Stress in Healthy Subjects. Sensors. 2018;18(8):2619. doi:10.3390/s18082619.
11. Blume C, Garbazza C, Spitschan M. Effects of light on human circadian rhythms, sleep and mood. Somnologie. 2019;23(3):147–156.
12. Sandercock GR, Bromley PD, Brodie DA. Effects of exercise on heart rate variability: inferences from meta-analysis. Medicine & Science in Sports & Exercise. 2005;37(3):433–439.