Breathing during sleep normally slows and becomes more regular than while awake. Gasping, witnessed pauses, or a low oxygen reading such as 87% are signs worth taking seriously: obstructive sleep apnea affects an estimated 24–33% of U.S. adults, and about 80% of cases go undiagnosed. A sleep study, not a wearable, confirms it.
The clearest way to think through breathing changes at night is the Rate → Pattern → Oxygen chain. First is the rate: how fast or slow breathing is. Second is the pattern: whether it stays steady or is interrupted by pauses, snoring, or gasping. Third is oxygen: whether those interruptions are actually dropping blood oxygen levels, which is the piece that turns a noisy night into a medical concern. Each section below follows that chain, ending with when the pattern is worth a doctor's visit rather than a wait-and-see approach.
Breathing during sleep is controlled automatically by the brainstem, and it typically slows and becomes more regular than breathing while awake. How steady it stays depends on which sleep stage the body is in. During N3, the deep slow-wave stage associated with the most restorative processes of the night, breathing tends to be at its slowest and most rhythmic. During REM sleep, the stage tied to dreaming, the body enters a state of muscle atonia in which muscles throughout the body relax, including muscles that help keep the upper airway open — which is one reason breathing can look and sound less steady, and why breathing-related problems such as sleep apnea often become more noticeable during REM.
There isn't a single sourced number in this article's research base for what a "normal" breaths-per-minute count looks like on a monitor, and any number quoted without a source shouldn't be trusted as medical fact. What matters more practically is the pattern over the course of the night: a breathing pattern that's basically slow and steady, without long pauses, gasping, or choking, is the baseline worth comparing everything else against.
Many people first notice their breathing rate while sleeping through a smartwatch or ring rather than a doctor's office, and that's where expectations need a reality check. Even trained sleep technologists scoring the exact same overnight polysomnography recording only agree with each other about 75–82% of the time. Consumer wearables do worse: independent testing of the Apple Watch found sleep/wake agreement of only about kappa 0.53, with total sleep time overestimated by roughly 19.6 minutes. The Oura Ring Gen 3 scored kappa 0.65 in manufacturer-funded testing but only 0.21–0.40 in independent testing, and consumer wearables generally struggle to correctly identify brief awakenings. A single night's wearable breathing-rate reading is a rough estimate, not a diagnosis.
What can genuinely change a night's breathing rate and pattern includes alcohol, sleep stage, and underlying airway problems. Alcohol is sedating, but it also fragments the second half of the night and can worsen sleep-related breathing disorders rather than improve them. The most common underlying cause of an irregular breathing rate at night is obstructive sleep apnea, which is estimated to affect roughly 24–33% of U.S. adults — as many as about 85.6 million people — with an estimated 80% of cases undiagnosed. If your nightly breathing-rate numbers look erratic, that prevalence is one reason it's worth more than a glance at an app.
If irregular breathing is disrupting your sleep schedule more broadly, it can help to also review your baby sleep schedule routines if you're managing an infant's sleep alongside your own, or look at best noise machines for sleep if snoring and restlessness are part of the disruption.
Blood oxygen is measured with a pulse oximeter, a sensor that's built into most home sleep tests, in-lab polysomnography setups, and many consumer smartwatches and rings. Overnight, it's tracked alongside breathing effort and airflow so that a dip in oxygen can be matched to a specific pause or shallow breath. This article doesn't have a sourced numeric range for what counts as a "normal" overnight blood oxygen level to state as fact, but the diagnostic framework used to interpret those dips is well defined: sleep clinicians score events using the apnea-hypopnea index (AHI), and obstructive sleep apnea is diagnosed at an AHI of 5 or more events per hour with symptoms, or 15 or more per hour even without symptoms.
Getting an actual oxygen reading typically means a sleep study. An attended, in-lab polysomnography (billed under CPT codes 95810 or 95811) is the most complete option, while an unattended home sleep apnea test (CPT 95806) records fewer channels but is more accessible. Medicare-based cost comparisons put the home option at roughly $169 versus about $625 for an in-lab study; a separate cost study found an average of about $419 for at-home testing versus about $746 in-lab.
A reading around 87% overnight — whether it comes from a home sleep test, an in-lab study, or a consumer pulse oximeter — reflects a real drop, not a device glitch to dismiss. It's generally treated as a possible sign of a breathing-related sleep disorder such as obstructive sleep apnea, since drops like that are exactly what the AHI-based scoring described above is designed to catch and quantify. A single low number on a wearable shouldn't be self-diagnosed at home; it's a reason to bring the reading to a doctor and pursue a proper sleep study, especially given that an estimated 80% of obstructive sleep apnea cases are currently undiagnosed.
Gasping or choking awakenings are one of the most recognizable symptoms of obstructive sleep apnea (OSA), a condition where the upper airway repeatedly narrows or closes during sleep, briefly cutting off airflow until the brain triggers a partial awakening to reopen it. OSA is estimated to affect roughly 24–33% of U.S. adults — as many as about 85.6 million people — and an estimated 80% of those cases are undiagnosed. Left untreated, severe OSA is linked to 3.0–3.8 times the overall mortality risk of someone without it, which is part of why gasping episodes shouldn't be brushed off as just a rough night.
"Stopping breathing" during sleep describes the same repeated pauses that define OSA, confirmed with the apnea-hypopnea index: 5 or more pauses per hour alongside symptoms like loud snoring or daytime sleepiness, or 15 or more per hour even without other symptoms, is the diagnostic threshold. A sleep study, not how tired someone feels the next day, is what actually confirms it.
Continuous positive airway pressure (CPAP) is the standard first-line treatment, but long-term adherence is a real challenge — only about 30–60% of users stick with it consistently over time. For people who can't tolerate CPAP, an implanted hypoglossal nerve stimulator called Inspire is an FDA-approved alternative, typically considered after CPAP hasn't worked; it's approved for an AHI of 15–100, a BMI of 40 or less, and central or mixed apneas making up less than 25% of the AHI. Total Inspire costs run roughly $30,000–$40,000, though Medicare-covered patients typically see an out-of-pocket cost closer to $2,000–$6,000.
Talk to a doctor if any of the following apply to you or a bed partner:
Because an estimated 80% of obstructive sleep apnea cases go undiagnosed, these symptoms are worth raising even if they seem minor — a sleep study, not a home device, is what confirms a diagnosis.