When a healthy person breathes slowly and regularly, heart rate rises with each inspiration and falls with each expiration, and the swing grows as the breathing rate comes down towards six breaths per minute. The effect is easy to demonstrate with an electrocardiograph and a metronome, which is one reason it appears in so many undergraduate physiology practicals. It also underlies heart rate variability biofeedback, in which people are trained to breathe at the rate that gives them the largest oscillation (Lehrer et al., 2000).

Sedentary young adults are an obvious group in which to study the response. They are free of the cardiovascular disease and medication that complicate work in older people, yet low habitual activity is associated with lower resting vagal indices than are seen in trained peers, so there is room for change. They are also the people most often recruited, as students, into laboratory studies. The literature on voluntary slow breathing is now large enough to have been pooled (Laborde et al., 2022), and reviews agree that the acute rise in variability is robust while its interpretation, and its persistence after the breathing stops, are less certain (Zaccaro et al., 2018).

Much of that uncertainty is methodological. This article therefore proposes a protocol for measuring the acute response in a consistent way. It is a design on paper, built from existing standards, and no recordings made with it are reported.

Two mechanisms combine during slow breathing. The first is respiratory sinus arrhythmia, the vagally mediated fluctuation of heart rate at the breathing frequency, which arises from central coupling of respiratory and cardiac vagal neurones and from reflexes driven by lung inflation and by changes in arterial pressure (Berntson et al., 1997). The second is the baroreflex. The closed loop formed by heart rate, arterial pressure and the baroreceptors has a natural period of roughly ten seconds, so breathing near 0.1 Hz drives it at resonance and the oscillations in heart rate become very large (Vaschillo et al., 2006; Lehrer and Gevirtz, 2014). Slow breathing has also been shown to raise measured baroreflex sensitivity (Bernardi et al., 2001), and it alters tidal volume, gas exchange and venous return as well.

The measurement problem follows directly. By long-standing convention the high-frequency band of the heart rate spectrum runs from 0.15 to 0.40 Hz and the low-frequency band from 0.04 to 0.15 Hz (Task Force, 1996). At six breaths per minute the respiratory peak sits at 0.1 Hz, inside the low-frequency band. Low-frequency power then rises steeply and high-frequency power may fall, although nothing resembling sympathetic activation or vagal withdrawal has taken place. The ratio of the two bands, already a doubtful index of autonomic balance at rest (Billman, 2013), becomes uninterpretable. Nor can a larger respiratory sinus arrhythmia be read simply as greater vagal tone when breathing rate and depth have changed (Grossman and Taylor, 2007). Any protocol has to be built around this.

Four principles shape the proposal. Each participant serves as their own control, because differences between people in resting variability are large compared with the effect of any brief manoeuvre. The comparison condition is paced breathing at the participant's habitual rate, not spontaneous breathing, so that the act of following a pacer is common to both. Respiration is recorded and not assumed. And the outcome measures are chosen in advance, with one primary index, in line with published reporting guidance (Laborde et al., 2017; Quintana et al., 2016). The protocol should be written up in full and registered before the first participant is seen, so that later choices cannot be shaped by the recordings.

The protocol has six parts, described below in the order in which they would be carried out. Table 1 lists the main design choices with their reasons, and Figure 1 shows the sequence of a single recording visit.

The protocol was drafted with ordinary physiology departments in mind. Everything in the core version can be done with a teaching-grade bioamplifier, a respiration belt and free or low-cost software. Consumer chest-strap monitors that export interbeat intervals are a tempting substitute for the electrocardiogram. They may be adequate for time-domain indices at rest, but they give no waveform to inspect, and a laboratory that uses one should first compare it against an electrocardiogram on its own participants under slow breathing.

A single visit of under an hour is enough for the acute response. Laboratories interested in training effects would need a different design, with repeated sessions and a follow-up recording at rest, and the present protocol would serve only as its measurement component. For class practicals the crossover can be kept and the registration and power calculation dropped, as long as the results are presented as a demonstration and not as a study.

The protocol is proposed and has not been validated here. Three questions would need answers before it could be recommended as a standard. One is feasibility: whether untrained sedentary volunteers can follow a pacer at six breaths per minute for five minutes without discomfort or overbreathing. Slow, deep breathing can lower carbon dioxide, which itself affects heart rate and may cause dizziness, and the core protocol does not measure end-tidal carbon dioxide. A pilot with capnography would show whether this matters. The second is reliability: whether the primary outcome is repeatable when the same person is tested on two days. The third is whether the control condition does its job, or whether pacing at the habitual rate itself changes variability compared with spontaneous breathing.

There are also limits that no amount of piloting removes. A fixed rate will be slightly off resonance for some participants and will understate their largest possible response. The protocol measures a reflex response lasting minutes and says nothing about health outcomes; a large oscillation during slow breathing is not evidence of benefit, and the article makes no claim that slow breathing treats any condition. Finally, sedentary status defined by questionnaire remains imprecise even with accelerometry as a check.

The acute heart rate variability response to slow-paced breathing is easy to elicit and hard to measure in a way that others can reproduce. The protocol proposed here fixes the choices that most often differ between studies: who counts as sedentary, what the slow condition is compared with, how breathing is verified, and which index is primary once respiration has moved into the low-frequency band. It is offered as a starting point for laboratories to pilot, criticise and refine, and its worth will depend on reliability data that have yet to be collected.

To be declared by the author(s) before publication.

To be declared by the author(s) before publication.