The body has one autonomic control surface you can grab with your hands: the breath. Heart rate, blood pressure, peripheral blood flow, and stress-hormone tone are all influenced by it on a beat-to-beat scale. The interesting question is not whether breathwork “works” — that is overdetermined — but which protocols actually shift physiology and by how much. The answer that has converged from a decade of biofeedback research is unflattering to the trendier methods and flattering to the simplest one: slow, deliberate breathing at roughly six breaths a minute.
What “resonance frequency” actually means
Heart rate is not constant between beats. It speeds up slightly with inhalation and slows with exhalation, a phenomenon called respiratory sinus arrhythmia. The amplitude of that oscillation can be measured as one component of heart rate variability (HRV) — and at one specific breathing rate, around 0.1 Hz (six breaths per minute), the respiratory oscillation couples constructively with the baroreflex oscillation, producing the maximum amplitude possible. This is the resonance frequency, and it is roughly the same for most adults.
Lehrer and Gevirtz’s 2014 review in Frontiers in Psychology is the most thorough treatment of the underlying mechanism: paced breathing at resonance frequency stimulates the baroreceptors, increases parasympathetic tone, and over weeks improves baroreflex sensitivity — the body’s ability to buffer blood-pressure perturbations.1
What the trials show
A 2017 meta-analysis pooled 24 randomized studies of HRV biofeedback for stress and anxiety, totaling 484 participants. The pooled effect size was Hedges g ≈ 0.83 — a large effect by behavioral-intervention standards — with consistent direction across populations.2 The intervention typically involved 10–20 minutes daily of paced breathing at resonance frequency, often with a biofeedback display, for 4–12 weeks.
For blood pressure, Joseph and colleagues in 2005 randomized 20 essential hypertensives to a four-week protocol of 15 minutes daily at six breaths per minute or to a frequency-uncontrolled control. The slow-breathing group dropped systolic BP by about 5 mmHg and improved baroreflex sensitivity; the controls did not.3 Effect sizes that small matter at scale — a 5 mmHg systolic reduction across a population is associated with meaningful reductions in stroke and coronary disease.
The most-cited recent paper is Balban and colleagues’ 2023 Cell Reports Medicine trial, which compared three short daily protocols — cyclic sighing (long-exhale-dominant), box breathing (4-4-4-4 cycles), and meditation — over 28 days in 108 adults. All three improved mood and reduced physiological arousal versus baseline, but cyclic sighing produced the largest effects.4 The trial design is unusually clean for the genre and points to exhale-dominant slow breathing as the active ingredient.
The mechanism, briefly
Slow breathing with a long exhale shifts autonomic balance toward parasympathetic dominance through several converging routes. Vagal nerve afferents fire in time with the exhalation; pulmonary stretch receptors signal reduced threat; baroreceptor activity oscillates at a frequency that the brainstem can use to set blood-pressure regulation. A 2017 review in Breathe catalogs the converging mechanisms in detail.5 Over weeks, baseline HRV rises, baroreflex sensitivity improves, and resting blood pressure modestly falls — adaptations that look more like cardiovascular fitness than a relaxation gimmick.
Two protocols worth knowing
The first is the classic resonance-frequency protocol: inhale through the nose for five seconds, exhale through the nose or pursed lips for five seconds, for 10–15 minutes once or twice a day. It is boring. It works.
The second is cyclic sighing, the protocol Balban tested: a normal nasal inhale, a short second inhale on top to fill the lungs, then a slow extended exhale through the mouth. Five to ten minutes a day. The double inhale appears to maximize alveolar recruitment; the long exhale drives the parasympathetic shift.
Both can be paced by an app, a metronome, or — once you have done them long enough — by feel.
HRV as a metric, in context
The dominant HRV metric on consumer wearables is RMSSD — root mean square of successive differences between heartbeats — which reflects short-term parasympathetic activity. Resting RMSSD is broadly higher in fitter, less-stressed adults. But the inter-individual range is large enough that comparing your absolute number to a friend’s is uninformative; tracking your own trend across weeks is the right use.
The honest caveat is that HRV is a noisy measurement. Single-night fluctuations can be substantial. Use it as a directional signal — if it sustains a drop for several days, something has changed and is worth examining — not as a daily verdict.
What it is not
Breathwork is not a cure for clinical anxiety or hypertension; the effect sizes are real but cumulative. It is not all equivalent: Wim Hof–style hyperventilation cycles produce a very different physiology than slow breathing and have their own risk profile, particularly in water (where they have caused fatal shallow-water blackouts). Box breathing is fine but appears less potent than exhale-dominant slow patterns in head-to-head trials. And HRV is not a moral score; chasing higher numbers can become its own anxiety.
The most defensible use is the smallest one: ten minutes a day of slow paced breathing, most days, as a baseline autonomic intervention layered on top of sleep and exercise. Cheap, quick, supported by surprisingly strong trial data.