Category Definition and Method Family
Breathing guidance is content or a service that uses voice, pacing sounds, music, vibration, or visual cues to help a user observe or intentionally alter breathing rate, inhale-to-exhale ratio, depth, route, and pauses. It belongs to WSIC 3.4.1. It may support a bedtime transition, general relaxation, mindfulness practice, or skill instruction. The category name does not identify one protocol and does not mean the content has been shown to treat insomnia, anxiety, lung disease, or cardiovascular disease.
The category includes natural-breath awareness, longer exhalation, equal pacing, diaphragmatic breathing, and other timed patterns, but 4-7-8, box breathing, resonance-frequency breathing, and HRV biofeedback should be identified as specific submethods. Mindful breathing primarily trains a relationship to sensations and thoughts and need not alter rate. Pursed-lip breathing, inspiratory-muscle training, and pulmonary rehabilitation serve specific clinical goals. Methods do not share efficacy, contraindications, or recommended parameters merely because all involve breathing.
Protocol Disclosure and Cue Design
Every guide should disclose rate or count, inhale-to-exhale ratio, breath retention, depth, nasal or oral route, posture, cycles, and exit method, and state whether parameters are fixed, progressively changed, or adjusted from individual signals. “Slow,” “deep,” “belly,” and “natural” are not interchangeable specifications. Even a slow rate may create overbreathing if tidal volume is excessive, while a pattern with retention has a different experience and risk from one without it. A vague title should not hide fast breathing, prolonged holds, or forceful exhalation.
Voice and pacing signals must align accurately with inhalation, exhalation, and pauses, provide a natural-recovery phase, and release any hold explicitly before fading. Cues should invite gentle, comfortable participation, not maximal inhalation, forced emptying, tolerance of air hunger, or perfect synchronization. Audio should not change speed without notice, fill the exhale with speech, insert advertising into a pause, or use challenges and streaks to push excess practice. Parameter and intended-use differences between versions should be comparable in titles and metadata.
Selection, Trial, and Bedtime Workflow
First distinguish an aim such as simple awareness, slower rhythm, longer exhalation, or coping with tension from management of a diagnosed respiratory condition. The first trial belongs in a fully awake daytime setting, seated or supported where a fall cannot occur. Begin with natural breathing and monitor for dry mouth, air hunger, dizziness, tingling, chest tightness, palpitations, cough, wheeze, or anxiety. Resume ordinary breathing or stop when discomfort occurs. Symptoms are not detoxification, oxygen-deprivation training, or a mandatory adaptation period.
At bedtime, lock the display, disable notifications, advertising, and autoplay, and set a stop and gentle ending. The task is not exact obedience to every beat, attaining a device score, or testing lung capacity. Straining to synchronize, repeated restarts, repeated heart-rate checks, and delayed bedtime while choosing a protocol can increase sleep-performance pressure. Test one gentle version across several nights and record perceived sleep onset, awakenings, next-day function, and discomfort. Greater drowsiness does not establish a change in sleep architecture, and a single heart-rate or HRV change does not establish lasting benefit.
Risk Signals and Clinical Boundary
Breathing too deeply or rapidly can lower carbon dioxide and may cause dizziness, tingling, rapid heartbeat, chest pain, or paradoxical breathlessness. Retention may create marked air hunger or a near-faint. Chest pain, fainting, severe or persistent breathing difficulty, blue lips, altered awareness, a new irregular heartbeat, or neurologic symptoms calls for stopping and emergency help according to severity. Drowsiness-inducing, rapid-breathing, and breath-hold practices are inappropriate while driving, cycling, swimming, bathing, at height, operating machinery, or alone with someone requiring continuous vigilance.
A person with asthma, chronic lung disease, sleep-disordered breathing, cardiovascular disease, low blood pressure, recurrent fainting, or recent chest or abdominal surgery should select a method according to clinical advice. Generic audio must not change oxygen, inhalers, positive airway pressure, or rehabilitation. Breath focus may magnify panic, traumatic recall, or compulsive counting, in which case an external sound or stopping may fit better. A caregiver should preview and observe children's content. Holding, suppressing, or accelerating breathing should not become a contest, and no one should override a bodily stop signal.
Equipment, Accessibility, and Situational Safety
Use the lowest comfortable playback level and preserve audibility of smoke, carbon-monoxide, infant-monitor, access, and medical alarms. Headphones can clarify pacing, but overnight use adds pressure, moisture, ear-canal irritation, and cleaning concerns; cables can entangle. Speakers and charging equipment should remain stable and uncovered by bedding. Calls, disconnection, buffering, advertising, and track transitions require testing so that an error signal cannot be mistaken for an inhale, exhale, or hold cue.
Accessible editions can combine clear speech, tones distinguishable in mono, optional vibration, an accurate transcript, adjustable pace, and separate music level, with an immediate pause and return to natural breathing. Color, animation, and left-right channel location cannot be the only phase marker. A user with hearing, vision, attention, language, cognitive, nasal-breathing, or movement limitations should understand the current phase and exit route. A version with changed parameters should be renamed or labeled and should not imply an identical physiological dose.
Sleep Evidence and Clinical Position
A 2026 systematic review included nine studies of slow breathing before bedtime with 457 participants. Self-reported sleep duration and quality signals were more positive, while actigraphy and polysomnography were inconclusive. Objective studies often used only a one-day intervention, and protocols, samples, and training periods varied substantially. An earlier experiment compared 14 self-reported insomniacs with 14 good sleepers and found improvement in some polysomnographic outcomes after presleep slow paced breathing. Its very small sample cannot establish general effectiveness or optimal parameters.
A 2025 randomized trial of 99 older adults with insomnia reported improvement over time in self-rated sleep quality in diaphragmatic-breathing and CBT-I groups. Fewer participants reached final analysis, the control did not receive matched attention, and the breathing intervention included in-person instruction rather than validating generic audio. A broader breathwork meta-analysis found small-to-medium effects on self-reported stress, anxiety, and depression, not sleep, and most studies raised bias concerns. Breathing guidance may be a preference-sensitive, reversible adjunct. Multicomponent CBT-I remains a first-line treatment for chronic insomnia, and persistent respiratory or sleep symptoms require cause-specific assessment.
Claims, Quality, Rights, and Privacy
Claims such as “sleep in one minute,” “raise blood oxygen,” “activate the vagus nerve,” “lower cortisol,” “reset the autonomic nervous system,” “increase deep sleep,” or treat insomnia, anxiety, hypertension, or asthma require reliable evidence matched to the exact protocol, delivery, population, and clinical outcome. Immediate changes in HRV, heart rate, or respiratory rate, subjective calm, completion, plays, and testimonials do not establish disease or sleep efficacy. Reports should separate between-group from within-group, subjective from objective, and short- from long-term results and disclose adverse events, attrition, cointerventions, and commercial relationships.
Quality assurance should verify script version, parameters, phase timing, cumulative drift, cue latency, pronunciation, loudness, true peak, channels, vibration synchronization, fade, pause, screen lock, offline use, disconnection, calls, and advertising. Traditional sources, modern arrangement, translation, narration, music, effects, and masters require separate rights review, while synthetic voice and professional review should be disclosed. If an app adapts pacing from a microphone, breath sound, camera, heart rate, oxygen saturation, or sleep log, it should explain sensor error, inference, and sharing, minimize collection, and support disabling, export, and deletion. Related entries include 4-7-8 breathing guidance audio, bedtime meditation guidance audio, body scan guidance audio, HRV biofeedback, pulmonary rehabilitation, sleep monitoring, sleep-disordered breathing, and CBT-I.