VELA: a standardised, integration-centred protocol for conscious connected breathwork

Conscious connected breathwork (CCB) produces large, reliable shifts in physiology and subjective state — yet it has stayed at the margins of science. The obstacle is not the phenomenon but the method: in most settings the practice is improvised, so it cannot be reproduced, compared across sessions and facilitators, or audited. We developed the VELA protocol to remove that obstacle — to specify CCB in enough detail that it can be measured, compared, and put to the test.

The VELA protocol specifies every element of a session — the breathing pattern, the acoustic environment, the role of touch and scent, the integration sequence, and the screening procedure — and gives each element an explicit rationale, labelled as evidence-based, theoretical, or hypothesised. This paper describes that protocol, the respiratory, cerebrovascular and autonomic mechanisms that plausibly underlie it, and the methodological problem of evaluating an intervention that cannot be blinded.

Its guiding design principle is unusual for the field. The protocol standardises toward a neutral field: it minimises emotionally directive music, verbal suggestion and spiritual framing, on the grounds that an altered, high-arousal nervous system is unusually open to outside influence. We draw on four bodies of work, chosen because each describes how an aroused or altered nervous system takes in information and how far the resulting conviction can be trusted: predictive-coding accounts of altered states, response-expectancy theory, the misattribution of arousal, and the somatic-processing tradition. Together they motivate a distinction between emergent insight — arising from the participant’s own system - and provoked insight - arising partly in response to external stimuli and, we hypothesise, less durable over time. This is one of the reasons we think the field’s default toward dramatic, externally driven sessions deserves scrutiny. The design choice is testable, and we state plainly where the argument is strong and where it is not.

This is a tested framework, not a demonstrated therapy. The protocol and the hypotheses set out here emerged from initial data collected in our first breathwork teacher-training cohorts; these are uncontrolled observations, which can generate findings and hypotheses but cannot establish efficacy — and no efficacy claim is made on their basis.

Keywords conscious connected breathwork · hypocapnia · cerebral blood flow · autonomic regulation · altered states of consciousness · predictive coding · suggestion · standardisation

Abbreviations CCB — conscious connected breathwork · CO₂ — carbon dioxide · BPM — beats per minute · REBUS — relaxed beliefs under psychedelics · CT — C-tactile (afferents) · PSS-10 — Perceived Stress Scale · PANAS — Positive and Negative Affect Schedule · DASS-21 — Depression Anxiety Stress Scales · RS-14 — 14-item Resilience Scale · BAQ — Body Awareness Questionnaire · PSQI — Pittsburgh Sleep Quality Index

Why VELA

The name is Latin, and it was chosen because the metaphor is exact rather than decorative.

A powerful practice that resists rigorous study

Breathwork works. The difficulty is that, as it is usually delivered, it is very hard to study — so how it works, and how reliably, remain open questions.

Breath has been used to alter consciousness for centuries, from yogic pranayama onward; the more specific technique of continuous breathing without the pause between exhalation and inhalation runs through twentieth-century methods such as Holotropic Breathwork [1] and Rebirthing [2]. The acute effects are not subtle: marked changes in autonomic balance, altered states of consciousness, intense emotional release. Controlled and meta-analytic evidence indicates that structured breathing practices can reduce stress, anxiety and low mood, though effect sizes are modest and the literature is heterogeneous [3, 4]. The physiological effects are measurable too: brief cyclic breathing lowers respiratory rate and physiological arousal [3], and in breast-cancer patients a single CCB session produced a mild respiratory alkalosis alongside a fall in serum cortisol [5].

Yet the field has a structural problem. Most CCB modalities are neither standardised nor reproducible, and the reason is concrete: there is usually no fixed protocol. The breathing rate, the music, the spoken guidance, the session length and the facilitator’s style all vary from one practitioner — and one session — to the next, and almost none of it is documented in a way another practitioner could follow and repeat. Outcomes are rarely measured, and contraindications are applied inconsistently. So although there is real evidence that breathwork helps — the controlled and meta-analytic findings noted above — we understand far less about how it works; and an intervention that varies this much is very hard to study, to compare against alternatives, or to correct when a claim turns out to be wrong. This is the gap the VELA protocol was built to close.

What VELA standardises toward

Standardisation is necessary, but it raises the critical question that most of the field never asks: standardise toward what? A session can be made reproducible and still be built to overwhelm — driven by dramatic, building music, emotional cueing and interpretive language until something dramatic reliably happens. VELA takes the opposite position. It standardises toward a deliberately neutral field, because the altered nervous system is unusually open to outside influence, and we want what emerges to belong to the participant rather than to the room. That argument — the scientific core of the method — is developed in Section 4.

Position among existing methods

VELA is an operationalisation of CCB, not a break from it. Table 1 compares it with several well-known methods. Among intensive breathwork paradigms, the Wim Hof Method is the best studied: controlled work has shown that its trained practitioners can voluntarily activate the sympathetic axis and attenuate an induced inflammatory response [6]. It is, however, a different paradigm — intermittent cyclic hyperventilation with breath retention and cold exposure — and is included as the current strongest available comparator.

Seven elements, each one defined

A VELA session runs in seven elements, each with a physiological target, a defined facilitator action and explicit transition criteria. Including intake and the post-session dialogue, a full session lasts approximately two hours (around 120 minutes): roughly 90 minutes of breathing and acoustic sequence — the integration phase included — framed by a 10–15-minute intake beforehand and a 15–30-minute dialogue afterwards. Two defined terms recur: Automated Diaphragmatic Expansion (ADE), a brief guided priming sequence, and Continuous Diaphragmatic Breathing (CDB), the active connected-breathing phase.

This paper specifies the protocol at the level of its elements, parameters and rationale — enough to evaluate the design and, in principle, to reproduce it — rather than reproducing the protocol in full. The complete operational detail — exact cueing, phase-transition scripts and decision rules — is set out in the VELA Facilitator Handbook, which is issued only to facilitators who complete certification. Delivering that detail through supervised training rather than publishing it is how the standard is kept consistent across facilitators, cohorts and sites.

2.1 Diaphragmatic priming and body position

The session opens with a short priming sequence — Automated Diaphragmatic Expansion (ADE) — before the continuous phase begins. ADE is built on the physiological sigh: a double inhale (a full breath topped by a second, shorter one) followed by a long, complete exhale. The second inhale re-inflates alveoli that collapse under shallow, stressed breathing, restoring the surface area available for gas exchange, while the extended exhale raises vagal tone and lowers arousal within seconds — the sigh is the body’s own reset for respiratory and autonomic state, generated by a dedicated brainstem circuit that turns ordinary breaths into sighs in response to physiological and emotional stress [7]. Performed deliberately and repeatedly, this pattern is among the fastest-acting of the breathing practices tested in controlled work: five minutes a day of cyclic sighing produced greater improvements in mood and larger reductions in physiological arousal than mindfulness or other paced-breathing conditions [3]. In VELA, ADE uses that mechanism to settle the nervous system and establish diaphragmatic engagement before the demands of the continuous phase — priming the body into a regulated baseline rather than starting from wherever a participant happens to arrive.

Priming is performed seated, in contrast to the supine position used thereafter. The rationale is autonomic: supine posture promotes parasympathetic predominance through baroreflex and postural mechanisms, which, applied before the active phase, tends to lower arousal at onset. A seated position preserves light sympathetic tone and a clear somatic boundary between preparation and process. We note that this rationale rests on facilitation experience and physiological plausibility, not on a controlled comparison.

2.2 Adaptive breathing rate

A single fixed ratio is poorly suited to a heterogeneous group: lung volume, baseline CO₂ tolerance and autonomic state at onset vary widely, and a fixed rate will be sub-optimal or dysregulating for some. VELA therefore defines a standard rate and a single, optional intensification:

• Standard — 3:3. A three-count inhale and exhale, used as the session baseline. It allows the breath to become self-sustaining and limits an early hyperventilation response in low-tolerance participants.
• Intensification — 2:2. Applied during the peak phases, on facilitator assessment, for participants who are stable at the standard rate. Transition criteria — respiratory stability, extremity temperature, facial tone, responsiveness — are specified in the Facilitator Handbook.

The 2:2 intensification is deliberately the upper limit of the protocol. A faster, more aggressive rate — such as a shortened 2:1 exhale — accelerates CO₂ off-loading in a way that is harder to titrate safely, and in practice pushes a substantial share of participants past the point at which they can stay present, producing overwhelm and dysregulation rather than material that can be integrated. The 2:2 rate, by contrast, opens access to a deeper state while keeping the experience within reach: intense enough to matter, restrained enough to remain workable — which is why it, and nothing faster, is the method’s most intense setting.

2.3 Acoustic environment

Music functions primarily as a pacing and co-regulation stimulus rather than as an emotional amplifier - a distinction given its full scientific weight in Section 4. We put this carefully, because no music is affectively neutral: even without words, harmonic mode, timbre, melodic contour and dynamic range all carry feeling, and a clearly traceable melodic line can set up anticipation — the listener senses where the music is heading and is carried along by it — which is itself a gentle form of external direction. The aim is therefore to minimise emotionally directive influence, not to claim it can be removed altogether; pacing and emotional colour are never fully separable, even in a tightly controlled acoustic field. Three constraints serve that aim on every playlist: no lyrics; no prominent solo voice; and a tempo ceiling of ≤105 BPM at the sustained peak, with a working band of roughly 90–100 BPM. The voice constraint goes beyond the absence of words: the human voice is processed by dedicated voice-selective regions of auditory cortex that respond preferentially to vocal sound whether or not it carries language [8], so even a wordless voice acts as a socially salient signal that pulls attention toward an external “other” — the orientation a neutral field is designed to avoid. Within these limits we also favour stable, slowly evolving textures over strong directional melodies that invite the listener to anticipate where the sound is going. The ceiling is a conservative operational parameter, not a physiological threshold (see 3.2).

A reasonable objection follows from this: if the aim is to minimise external direction, why retain music at all rather than work in silence? The answer is that silence is not the absence of an external condition but a different one, and often a stronger. Removing the acoustic field does not remove influence; it shifts the burden of pacing and phase transitions onto the facilitator’s voice, and spoken instruction is far more semantically directive than a wordless, slowly evolving texture — it reintroduces precisely the content-laden suggestion the design is trying to reduce. Music also does measurable regulatory work: across 104 randomised trials, music listening produces small but reliable reductions in physiological and psychological stress markers [9], the co-regulation role the tempo constraints are built to preserve. In a group the case is stronger still. An unmasked room transmits every cough, sob and shift of the person nearby, and affect spreads between people automatically through exactly these vocal and postural cues — the well-documented phenomenon of emotional contagion [10] — so silence yields not a private, endogenous field but a socially coupled one, in which one participant’s catharsis can entrain the room. A neutral, low-information acoustic bed masks this, giving each person an individual container. None of this makes silence wrong in principle; whether it would produce more durable, self-generated insight than neutral pacing is an empirical question, testable within the same dismantling design described in Section 6.

The acoustic tempo across a session is shown in Figure 1. Breathing runs at the 3:3 standard, moves to an optional 2:2 intensification through the peak phases, and returns to natural breathing for the descent and extended integration. The vertical peak is a brief drop to a near-still pulse; the horizontal peak is the longer sustained segment, held below the 105 BPM ceiling. Values are specification targets, subject to facilitator adjustment — orientation, not a fixed prescription.

2.4 Enhancing the protocol: VELA Scent and VELA Touch

We will enhance the protocol through the planned addition of two further modules — VELA Scent and VELA Touch — that extend it into other sensory channels. Both are in active development; the outlines below are deliberately preliminary, and their full specifications will be published separately.

VELA Scent maps olfactory cues to the session phases, exploiting the olfactory system’s unusually direct projection to the limbic structures that govern emotion and autonomic state. Aromatics are chosen for documented effects on arousal and parasympathetic tone and for fit to the arrival, peak and integration phases; used consistently, a fixed scent can also act as a conditioning anchor for the state across sessions.

VELA Touch uses calibrated, consent-based facilitator contact as a non-verbal regulation tool. It is grounded in the affective-touch system: slow, gentle stroking of hairy skin activates C-tactile (CT) afferents projecting to the insula, supporting vagal engagement and lowered cortisol. Touch is not a neutral intervention — in the high-arousal CCB state its meaning shifts, and the same contact can be experienced as grounding or as intrusive — therefore it is indicated only in defined phases and never before consent or during the closing dialogue.

Both modules follow the principles of the core protocol: an explicit physiological rationale, phase-specific indication, defined contraindications, and no claim beyond the evidence. Each will be published in full separately.

What the breath plausibly does

The mechanisms below are hypotheses. They rest on established physiology and on the emerging neuroscience of high-ventilation breathwork [11]; none has yet been measured in this protocol specifically. Section 6 describes how they would be tested.

3.1 Respiratory and cerebrovascular effects

Continuous breathing without the inter-breath pause increases alveolar ventilation and lowers the arterial partial pressure of carbon dioxide (hypocapnia); a fall in CO₂ during circular breathwork has recently been reported and linked to the emergence of altered states [12], and measured directly as respiratory alkalosis — a fall in pCO₂ and a rise in pH — during CCB sessions in a clinical population [5]. Hypocapnia causes cerebral vasoconstriction and a reduction in cerebral blood flow — a relationship first quantified by Kety and Schmidt (1948) [13] and reproduced many times since; across the physiological range the reduction in flow is roughly proportional to the fall in CO₂, on the order of a few per cent per millimetre of mercury [14]. At the same time, the respiratory alkalosis that accompanies hypocapnia shifts the oxyhaemoglobin dissociation curve leftward (the Bohr effect), raising haemoglobin’s affinity for oxygen and reducing its release to tissue despite high arterial saturation. These changes are dose-dependent and reversible, and their magnitude tracks breathing rate and depth — which is precisely why the rate is specified rather than left to chance. Figure 2 summarises the cascade: lowered CO₂ is the central event, branching into the Bohr effect (reduced tissue-oxygen offloading) and cerebral vasoconstriction (reduced perfusion), with acoustic pacing modulating the autonomic trajectory in parallel (3.2). Each step is established physiology; their integration in this protocol is hypothesised, not yet measured.

3.2 Autonomic modulation and acoustic tempo

The tempo ceiling prevents sustained sympathetic dominance. Controlled studies of musical tempo observed that slower tempi and silence reliably favour a shift toward parasympathetic recovery, while faster tempi are associated with increased sympathetic arousal — heart rate, blood pressure, ventilation, skin conductance — though this second effect is markedly more variable and strongly shaped by individual differences [15, 16, 17]. Because the arousal response is graded and inconsistent rather than switch-like at any particular value, the ≤105 BPM ceiling is therefore a conservative parameter placed below the band where arousal tends to take over, not a threshold with independent physiological meaning. The descent into ambient sound and silence is aimed at returning the system toward parasympathetic predominance, a shift associated with greater vagally mediated activity in the slow-breathing literature [18].

There is a second reason to keep arousal in check. Sustained sympathetic dominance does not only raise heart rate and ventilation; it changes how the brain handles what is happening. Under stress, large releases of noradrenaline and dopamine weaken the prefrontal cortex that supports slow, reflective, top-down processing, while strengthening faster, more reactive circuits in the amygdala and striatum [19]. The adaptive result is a nervous system tuned outward — toward rapid defensive responding rather than inward reflection — which is useful in real danger but is the opposite of the receptive, interoceptive mode in which emotional material can be noticed and worked through. That same shift therefore does double damage to the work of a session: it pulls attention toward the environment, and it makes the quiet, integrative processing on which durable insight depends harder to sustain. (Whether high arousal specifically reduces interoceptive accuracy is still debated; what is well established is the prefrontal-to-subcortical shift itself.) It is part of why the protocol caps tempo and devotes a long, low-arousal phase to integration rather than maximising peak intensity.

3.3 Altered states, cortical activity and perfusion

Hypocapnia-related changes in cortical excitability, combined with sustained rhythmic input, are associated with altered states whose subjective features — reduced self-referential processing, altered time perception, heightened emotional availability — resemble, without being identical to, those of psychedelic and deep meditative states [20, 21]. Recent neuroimaging of high-ventilation and circular breathwork ties these states to measurable changes in cerebral perfusion, cortical activity and autonomic tone [12, 22, 23]. In an arterial-spin-labelling study the intensity of the altered state tracked both reduced cerebral perfusion in interoceptive and affective regions and a rise in sympathetic activation [22]. It is worth being explicit about which similarities are which: the overlap at the level of subjective experience — altered time perception, reduced self-referential processing, heightened emotional availability — is reasonably well documented by self-report, and some overlap in cerebral perfusion and cortical activity has now been measured for breathwork specifically [12, 22, 23]; the stronger claim of a shared underlying neural mechanism, however, remains a conceptual and phenomenological parallel rather than an established finding. Involvement of the default mode network is frequently proposed by analogy with the psychedelic literature, but it has not been characterised for CCB, and we treat it as a hypothesis rather than a result.

3.4 Integration and bottom-up processing

The debrief order in element 7 — sensation, then emotion, then narrative — follows clinical models of trauma processing that prioritise interoceptive and somatic awareness before cognitive narration [24, 25]. The rationale is that leading with a narrative recruits prefrontal regulation prematurely and forecloses access to the somatic material the session opens. These are clinical syntheses rather than controlled mechanism studies, and we mark them as such; the sequencing claim is, in our own terms, a hypothesis.

Emergent versus provoked insight

Two participants can leave a session each convinced they have understood something essential about their lives. The problem is that, in the moment, the two are indistinguishable — and the difference only becomes visible weeks later, in whether or not the insight holds.

This is the question that decides what a breathwork method should standardise toward, and most of the field settles it by default rather than with intention. We define the distinction precisely and then defend it.

• Emergent insight arises from the participant’s own nervous system once the physiological conditions are in place and external input stays neutral. Its content is generated internally.
• Provoked insight arises as a response to external stimulus — emotionally directive music, verbal guidance, spiritual framing. It can feel entirely like one’s own realisation. It is not, or not wholly.

In the altered state the threshold for suggestion is markedly lowered: what would ordinarily be recognised as an outside signal can be encoded as an inner truth.VELA acoustic-architecture rationale

Four independent bodies of work — the predictive-coding account of altered states, response-expectancy theory, the misattribution of arousal, and the somatic-processing literature — converge on the same conclusion: the deep breathwork state is the condition under which external input is most likely to be mistaken for internal revelation. None of these literatures was developed for breathwork, and we are careful below about how far each can be carried. The premise is not ours alone: across the best-studied altered states — hypnosis and the classic psychedelics — the capacity to respond to suggestion is reliably elevated, and verbal suggestion alone can shape what the experience becomes [26]. Whether high-ventilation breathwork raises suggestibility to the same degree has not been measured directly; we treat the parallel as a well-motivated extrapolation, not an established fact about breathwork.

4.1 Why the altered state is uniquely open to suggestion

The predictive-coding account (REBUS)

We start from a general idea about altered states, then borrow a specific model. The general idea is that many altered states — however they are produced — share an information-processing signature: the high-level priors that normally filter perception (a person’s settled expectations about self and world) lose precision, so the system leans more heavily on incoming, bottom-up information. When that happens, external input acquires disproportionate influence, and the brain is most open to outside influence at the very moment its usual filters are down. The most developed account of this is the REBUS model — relaxed beliefs under psychedelics — which integrates the free-energy principle with the entropic-brain hypothesis [27]. We borrow REBUS as a predictive-coding lens, not as a claim of shared biology, and the distinction matters: REBUS was formulated specifically for serotonergic psychedelics, where the trigger is agonism at the 5-HT2A receptor [27], whereas a breathwork state is driven by hypocapnia and the cerebrovascular and metabolic changes that follow [5, 12]. Two states can resemble each other from the inside without sharing a mechanism, and REBUS may not transfer in its full theoretical structure — a caveat the reader should carry from the outset rather than meet only as a later disclaimer (we return to it in 4.6). What makes the extension worth making at all is that CCB produces an altered, lowered-filter state of the same broad family without the after-effects of psychedelics, and predictive coding is the most plausible level at which the two might correspond.

Response expectancy and suggestion

The hypnosis and placebo literature supplies the mechanism’s behavioural counterpart. Response-expectancy theory shows that an expected non-volitional response tends to generate the corresponding subjective experience, and that these experiences are genuinely felt — verifiable in behaviour and physiology — rather than merely reported [28]. Suggestion can produce profound, authentic-feeling alterations of experience, including dissociative ones; crucially, the strength of subjective conviction is not evidence of the experience’s origin. Notably, response-expectancy theory has itself been reframed in predictive-coding terms, which is why it dovetails with REBUS rather than competing with it. The implication for breathwork is direct: a facilitator’s framing, or a track engineered to evoke grief or awe, is a suggestion delivered to a system whose capacity to mark it as external is impaired.

4.2 Misattribution of arousal

Where does the content of a provoked insight come from? The two-factor theory of emotion holds that physiological arousal is ambiguous, and that the mind reaches for available situational cues to label it [29]. The participant in the sustained peak is in a state of marked, unexplained arousal — CO₂ reduced, blood chemistry shifted, perfusion altered — and the nervous system searches for an account of that state. Emotionally directive music supplies one. The resulting “realisation” may be, in part, a misattribution: arousal generated by the breath, given meaning by the soundtrack. We are deliberate about the evidence here. The original two-factor experiments have well-documented replication problems, and emotion is now understood as a loop rather than a simple arousal-plus-label sequence. We therefore lean on the robust core — that ambiguous arousal is shaped by context — and on its better-replicated demonstrations in the misattribution paradigm [30], not on the contested original studies.

4.3 Bottom-up versus top-down

The somatic-processing literature adds a final reason to prefer emergence. Deep processing of emotional material is held to begin in the body and move upward, not to start as cognition or narrative [24, 25]. An external stimulus that triggers an emotional or cognitive reaction before somatic processing has occurred routes the experience top-down and tends to leave it shallowly integrated and poorly anchored in the body. Insight that rises from breath and sensation follows the bottom-up path; provoked insight more often arrives top-down, which is one reason it translates poorly into lasting change.

4.4 The two kinds of insight, compared

The distinction is invisible during the session and becomes legible only afterward; the table sets out where it shows.

This is the scientific reason VELA's acoustic field is neutral, its facilitation non-verbal in the peak, and its framing minimal. A neutral field is not an aesthetic preference. It is the condition under which the nervous system produces something that is genuinely its own.

The best session music is the kind no one mentions afterward — because no one noticed it was there.VELA acoustic-architecture rationale

4.5 Integration over catharsis

The same logic informs a second design choice: VELA treats dramatic emotional discharge as a by-product to be contained, not as the goal. This runs against a strong intuition — the “hydraulic” picture in which emotion builds up like pressure and must be released. That picture is one of the more thoroughly discredited ideas in experimental psychology. Venting does not discharge an emotion; it rehearses it, and reliably increases rather than reduces the state it expresses, to the point that the field called decades ago for a moratorium on catharsis as a therapeutic technique [31]. Activities that lower arousal outperform cathartic discharge. VELA's design follows that evidence: the work of the method is not the peak but the quiet, parasympathetic integration phase, where an unforced reorganisation can consolidate. We note the limit of this argument honestly — the strongest catharsis-fails evidence concerns anger and aggression, and its extension to breathwork-related emotional release is by analogy; the more specific clinical concern is the well-recognised risk of abreaction without integration, which the integration phase and the bottom-up debrief are designed to prevent.

4.6 Limits of the argument

Like any proposed framework, this one has limitations and unproven claims — and taking it seriously means naming them first.

• It is not yet measurable at the individual level. We cannot currently determine, for a given person and a given insight, whether it was endogenous or suggested. Both feel identical in the moment; the claim that they diverge later is itself a hypothesis that requires longitudinal, ideally behavioural, data to test.
• The strongest framework was built for a different intervention. REBUS is a model of psychedelic pharmacology. Its application to breathwork is plausible and, we think, fruitful, but it is an extrapolation, not an established finding.
• One pillar rests on contested classics. The two-factor theory’s original studies replicate poorly. We use the surviving principle, not the original claims, but a critic is entitled to discount this strand accordingly.
• Two pillars are clinical, not experimental. The somatic-processing accounts are influential syntheses, not controlled mechanism studies.
• A neutral field is not a blank one. Removing music and language does not remove context. The breathing instruction, the supine posture, the setting, the facilitator’s presence and the participant’s own expectations all shape what arises; under a predictive-coding account no experience is free of priors. Scent, likewise, is a deliberate input rather than a neutral absence: certain aromatics project directly to limbic structures and can be grounding or mildly anxiolytic, or can invite a deeper process — a controlled, non-directive regulatory cue, not a way of steering content. What VELA claims is narrower and more defensible — that it minimises exogenous, directive shaping — not that it achieves a suggestion-free state.
• Neutrality may forgo real benefit. It is entirely possible that, for some people, music-evoked emotion or skilful framing facilitates processing. A strict neutral-field stance could leave benefit on the table. We do not assume our position is correct; we regard it as the more conservative default, and — importantly — as testable. The conservative choice is also the safer one: removing provoking music protects a participant whose nervous system, in a high-arousal state with lowered defences, is especially open to overstimulation, and minimising external shaping protects the participants as well as the integrity of what emerges — participant safety is the first priority. For participants who find it genuinely hard to reach their emotional material, we are developing VELA (E), an explicitly emotional variant: a more evocative playlist that uses external cues to help them open up. It deliberately invites the very provoked responses the standard protocol otherwise avoids — and once a participant has opened, the work returns to the standard, neutral protocol, with its stronger emphasis on bodily intelligence and emergent insight. The neutral field is the default, in other words, not a dogma.

This is not an untested preference. While developing the protocol we compared an emotionally directive acoustic field against a neutral one, and active verbal guidance against non-verbal support; the neutral, non-verbal configuration produced the steadier, more self-authored responses the method now builds around. That comparison was developmental and observational rather than a controlled trial — but it is the reason the neutral field is a derived feature of VELA, not an assumption. The durability contrast it points to is shown below: provoked and emergent insight may feel comparable on the day — provoked may even feel stronger — yet tend to diverge over the following weeks. The trajectories in Figure 3 are illustrative, not measured data.

Safety begins before the first breath

Screening is mandatory and is delivered as element 1; facilitators are trained and certified to administer and document it. The contraindications below follow from the known hazards of voluntary hyperventilation [11] — hypocapnia, alkalosis, transient cerebral hypoperfusion and the possibility of intense affective release — and are intended to be conservative.

Absolute contraindications

• Cardiovascular disease, including uncontrolled hypertension, arrhythmia or a recent cardiac event
• Epilepsy or other seizure disorder
• Active psychosis, or a psychotic episode within the previous 12 months
• Pregnancy (all trimesters)
• Severe asthma or COPD requiring daily medication
• Acute or uncontrolled glaucoma (precautionary, given pressure changes associated with intensive breathing)
• Major surgery within the previous three months

Relative contraindications

These require individual facilitator assessment and, where indicated, clearance from a physician or relevant specialist:

• Anxiety or panic disorder (low CO₂ tolerance may be present; a modified protocol is available)
• History of dissociation or depersonalisation
• A personal history of a psychotic episode more than 12 months ago, or a first-degree family history of psychosis or schizophrenia (precautionary)
• Acute grief, or severe or recent psychological trauma (within the previous six months)
• Medication affecting respiratory or cardiovascular function
• Diabetes (type 1, or insulin-dependent type 2)

Adverse event management

Certified facilitators are trained to recognise and manage carpopedal (tetanic) spasm, vasovagal responses, acute emotional crisis and dissociative episodes. A written adverse event protocol specifies, for each of these presentations, the steps a facilitator follows — immediate physical and verbal management, criteria for pausing or ending a session, and escalation to external medical care where indicated. Response procedures are standardised and assessed at certification, and adverse event recording is built into the evaluation framework so that incidence can be reported against a denominator — controlled breathwork trials to date report no adverse events directly attributable to the practice, yet only a minority record them at all [4], which is precisely why systematic recording matters.

Studying a method that cannot be blinded

This is the hardest problem the method faces, and we address it directly. Evaluating any intense experiential intervention is methodologically difficult; stating those difficulties plainly, rather than working around them, is a precondition for credible science and for the kind of collaboration this work will require.

6.1 Outcome measures

Proposed measures span patient-reported outcome measures (PROMs) and objective markers. The self-report instruments are all validated, widely used PROMs, which keeps VELA data comparable with the surrounding literature; wherever equipment allows they are paired with objective markers that are far less open to expectancy. Outcomes are analysed with mixed-effects models for repeated measures against a single primary endpoint specified in advance to limit selective reporting, and all results — including null and contrary findings — are reported in full:

• Stress and autonomic state — Perceived Stress Scale (PSS-10 [32]) and, where feasible, heart-rate variability (HRV), end-tidal CO₂ (capnography), respiratory rate and salivary cortisol — measures already used in breathwork and adjacent research [3, 5, 12, 18].
• Affect — Positive and Negative Affect Schedule (PANAS [33]), with qualitative coding of session narratives.
• Depression, anxiety and stress symptoms — Depression Anxiety Stress Scales (DASS-21 [34]).
• Resilience — Resilience Scale (RS-14 [35]).
• Interoceptive awareness — Body Awareness Questionnaire (BAQ [36]), pre and post training, paired where possible with an objective heartbeat-detection task, since self-report captures believed rather than measured interoceptive accuracy [37].
• Sleep quality — Pittsburgh Sleep Quality Index (PSQI [38]), tracked across time points; sleep is a meaningful, comparatively objective outcome and aligns VELA data with the wider breathwork and retreat literature.
• Facilitator competence — structured assessment against a defined rubric: a fixed checklist, scored by a trained assessor, that rates protocol adherence — rate cueing and transition criteria, playlist and tempo adherence, screening and contraindication handling, non-verbal conduct through the peak, safety response, and the sensation-first debrief.
• Adverse events — mandatory reporting of all events of moderate severity or above.

6.2 The control problem

An ideal trial randomises participants to the intervention or to an inert placebo and keeps everyone blinded to which treatment or intervention they receive. That is not an option in a technique that expects engagement from the participants.

The most informative comparison is psychedelic medicine, which has spent a decade confronting the same wall. The acute effects of an active dose are unmistakable, so participants and clinicians routinely work out who received what; blinding fails in practice, and this functional unblinding lets expectancy inflate apparent benefit, particularly in self-reported outcomes [39]. Placebos rarely solve it. You cannot blind a person to whether they have just spent forty-five minutes breathing themselves into an altered state any more than you can blind them to a full psychedelic dose.

The problem runs deeper than blinding. The strongest single piece of evidence comes from a placebo-controlled study run inside real ayahuasca retreats: depression, anxiety and stress decreased after the ceremony in both the active and the placebo groups, with the change largely independent of whether anyone had received the drug [40]. The retreat itself — rest, nature, removal from ordinary life, group belonging, expectation — did much of the work. Any residential breathwork programme inherits exactly this confound. Improvement after a VELA retreat cannot, on its own, be credited to the breathing.

6.3 Making progress without a placebo

The conclusion is not that nothing can be learned, but that no single study will settle anything and that the convenient designs are the weakest. What the wider field actually does — and what we intend to draw on — is a layered approach:

• Within-subject (N-of-1) baselines. A common approach in clinical research is to make each participant their own control across pre-, during- and post-intervention measurement — an N-of-1 logic. A pre-enrolment measurement well before the intervention (a T−14 baseline) estimates each participant’s own trajectory and provides a reference for regression to the mean, the tendency of people who enrol while distressed to improve regardless of treatment.
• Active comparators that hold the setting constant. The decisive question is not breathwork-versus-nothing but breathwork-versus-the-same-retreat-without-the-breathing. Comparing VELA to a credible low-intensity breathing or relaxation condition delivered in the identical environment is the only way to separate the protocol from its context — and breathwork has already served as exactly this kind of comparator in psychedelic retreat research [41].
• Measured expectancy and blinded raters. Expectancy can be measured at entry and modelled as a covariate, and outcomes can be scored by assessors blind to condition, both borrowed from current psychedelic methodology.
• Naturalistic benchmarking. Where a controlled arm is impossible, outcomes can be compared against external datasets and natural history trajectories rather than against an internal placebo.

Cross-over designs: they seem to be attractive options in this context — each participant serves as their own control in two consecutive phases — but they are largely unusable for this intervention. A cross-over assumes the effect of the first condition washes out before the second begins. A breathwork session that produces a genuine insight, or shifts how someone relates to a problem, does not wash out; the carry-over is the point of the intervention. Cross-over may retain a narrow role for acute physiological outcomes that do reset between sessions — CO₂, HRV, respiratory rate — but not for the durable psychological outcomes that matter most.

6.4 The current study

Concretely, and within these constraints, data collection begins with a single-arm, repeated-measures observation of a training cohort across five time points (T−14, day 1, day 14, day 28, day 90), with the instruments above and a pre-specified primary outcome to limit fishing. This design is explicitly exploratory. It can establish feasibility, generate effect-size estimates and surface safety signals; it cannot, on its own, attribute change to the protocol rather than to expectancy, setting or time. We say so in advance, and we report results — including null and contrary ones — accordingly.

6.5 Observations to date

The Academy has delivered certified training across several residential cohorts, most recently in Corfu (2026); further programmes follow in Mallorca (September 2026) and Marbella (spring 2027). Participants in these cohorts complete the instruments listed in 6.1 at the defined time points (T−14 through T90), so that structured data accrue alongside the training. The data gathered so far are uncontrolled observations rather than outcome evidence: participant reports describe recurring themes of autonomic regulation, emotional release and increased professional confidence, but such reports are subject to selection and expectancy bias and are used to refine the measurement protocol and generate hypotheses, not to support claims of efficacy. The first structured results — day 1, day 14, day 28 and day 90 measures from the Corfu (2026) cohort — are currently being analysed and prepared for publication as an exploratory, hypothesis-generating report.

6.6 Collaboration

The protocol is documented to a standard intended to support independent evaluation — standardised procedures, predefined outcomes, and systematic adverse event recording. Teacher training has so far been delivered in German, with English-language training in preparation; the Academy welcomes collaboration with academic and clinical researchers in both German and English, and will report formal partnerships as they are established.

What we do not claim

Drawing the threads together: the mechanisms of Section 3 are imported from established physiology and from studies of adjacent paradigms, and none has been measured in VELA itself; the psychedelic and meditation literature supports analogy, not identity. The protocol’s specific parameters — rates, tempo ceiling, durations, the two-segment structure — derive from theory and experience and have not been empirically optimised. The outcome evidence in hand is uncontrolled and limited to a self-selected, non-clinical population, so no efficacy or clinical claim is warranted. This selection point deserves emphasis in its own right: people who enrol in a multi-day residential breathwork training are a highly self-selected group, likely high in openness to experience, intrinsic motivation for change, and the time and means to attend, so any improvement observed may partly reflect these baseline characteristics rather than the protocol — and generalisability to broader or clinical populations cannot be assumed. The central emergent–provoked thesis, while well-motivated, is not yet measurable at the individual level and leans in part on a contested classic and on clinical syntheses. The authors also have a commercial interest in the method, stated below — a bias that blinded assessment, independent replication and objective measurement can reduce in future work but cannot remove. Taken together, these are the limitations of a protocol at an early, pre-confirmatory stage of evaluation. We state them so that they can be addressed directly in the work that follows.

A method built to be tested

VELA offers a fully specified protocol for CCB — defined at the level of breath, sound, touch, scent, integration and screening, and grounded in physiology that is real even where its application here remains hypothetical. (VELA itself is a method and a training programme; what this paper specifies is its protocol.) Its distinctive commitment is to a neutral field — to arranging conditions under which insight can emerge from the participant rather than be installed by the room — and that commitment is not a slogan but a falsifiable claim, with a clean experiment attached. The contribution we mean to make is methodological. By specifying the procedure and naming the confounds, the method becomes something the field has largely lacked: a standardised breathwork practice that can be measured, replicated, and, where the evidence does not support a claim, abandoned. This document is best read not as a finished protocol paper but as a pre-confirmatory study concept — the protocol, the reasoning behind it, and the studies that would test it, set out together so that others can test them too.

If this is a study concept, it should say what the studies are. The questions below are, in our view, the ones the field most needs to answer, each paired with the kind of design that could answer it.