Heart Rate Variability (HRV): Physiology and Healthcare Applications

From autonomic control of the sinus node to computing metrics and using wearables

Today’s goals

  1. Explain the differences in molecular mechanism and speed between the vagus and sympathetic nerves
  2. Explain the origins of the fluctuations that produce HRV (respiration, baroreflex)
  3. Compute SDNN, RMSSD, LF, and HF and explain what they mean
  4. List the confounders that mislead HRV interpretation
  5. Explain the evidence and limitations of clinical and healthcare applications

1. What is HRV?

The heart is not a metronome

  • RR intervals fluctuate by tens of ms from beat to beat → heart rate variability (HRV)
  • At rest, fluctuating is normal
  • The main source of the fluctuation is beat-to-beat regulation by the autonomic nerves (especially the vagus)
  • An indirect index that provides autonomic information non-invasively and continuously

2. Autonomic control

Two nerves, two speeds

Vagus Sympathetic
Transmitter / receptor ACh → M2 NE → β1
Main mechanism Gβγ → GIRK opening, cAMP↓ cAMP↑ → If↑, PKA
Speed Within 1 beat Seconds to ten-plus seconds
Fluctuations transmitted Up to HF LF and below only
  • Resting heart rate < intrinsic heart rate (about 100 bpm) → vagal dominance at rest

Difference in speed = difference in frequency response

  • The sympathetic nerves act as a low-pass filter → cannot transmit fluctuations at the respiratory period
  • HF ≈ vagus, LF = vagus + sympathetic
  • Atropine greatly reduces HF and LF too

3. Origins of the fluctuations

Bands and origins

Band Frequency Main origin
HF 0.15–0.4 Hz Respiration (RSA)
LF 0.04–0.15 Hz Baroreflex (Mayer waves, period about 10 s)
VLF 0.0033–0.04 Hz Thermoregulation, hormones, etc. (much unknown)
ULF ≤ 0.0033 Hz Circadian rhythm (24-hour recordings)

RSA and resonance

  • RR↓ during inspiration, RR↑ during expiration
  • Main cause: central gating of the vagus by the medullary respiratory center
  • Breathing at about 6 breaths/min → RSA and the baroreflex resonate → maximal HRV (biofeedback)
  • The fluctuation then falls into the LF band

4. Measurement

ECG and PPG

  • ECG: 250–500 Hz or higher, R-wave detection (Pan–Tompkins, etc.)
  • What PPG gives you is PRV ≠ HRV
    • Pulse-transit-time fluctuations are added, tending to overestimate RMSSD and HF
    • Diverges with movement, standing, low perfusion
  • Correction of ectopic beats is essential; report the correction rate
  • Frequency analysis: resampling → detrending → Welch / AR / Lomb–Scargle

5. Metrics

Main metrics

Metric Computation Mainly reflects
SDNN SD of NN Overall fluctuation (depends on recording length)
RMSSD RMS of adjacent differences Vagus
HF Power in 0.15–0.4 Hz Vagus (depends on respiration)
LF Power in 0.04–0.15 Hz Vagal + sympathetic + baroreflex
SD1 / SD2 Poincaré SD1 ≈ RMSSD/√2

LF/HF ≠ autonomic balance

  1. LF has a large vagal contribution
  2. LF does not correlate with direct measures of cardiac sympathetic activity
  3. The two nerves are not in a simple tug-of-war
  4. Merely slowing breathing raises LF/HF greatly

6. Confounders

If HRV changes, check first

  1. Heart rate (RR = 60000/HR → fluctuations look smaller just because heart rate is higher)
  2. Respiration (rate, depth)
  3. Posture, time of day, recording length, pre-measurement conditions
  4. Correction of abnormal beats and artifacts
  5. Age, fitness, drugs (β-blockers, anticholinergics), alcohol

7. Applications

Strength of evidence

Application Evidence
Prognosis after MI (24 h SDNN < 50 ms: relative risk of death about 5.3) Strong
Diabetic autonomic neuropathy (deep breathing, Valsalva, standing) Strong
Neonatal sepsis monitor (mortality↓ in an RCT) Strong
Stress, mental health Moderate (low specificity)
Exercise conditioning (lnRMSSD moving average) Moderate
Wearables Developing

Principles for use in healthcare

  1. Look at within-individual change (do not compare with others or other devices)
  2. Standardize conditions (right after waking, same posture, during sleep)
  3. Use moving averages and individual reference ranges
  4. Combine with heart rate, sleep, and subjective state
  5. Do not use for diagnosis

Summary

  • Fast HRV fluctuations (HF, RMSSD) are mainly vagal in origin
  • LF is not an index of sympathetic activity; be careful with LF/HF
  • Respiration, heart rate, and abnormal beats are the biggest confounders
  • Clinically valuable for prognosis and autonomic neuropathy; in healthcare, the value lies in within-individual trends