The “Sleep Regression” Myth: Why Your Baby’s Sleep Falls Apart Right Before They Crawl, Stand, or Talk (4–12 Months)

Last updated: August 2026. This article is for general information and is not a substitute for personalised advice from your doctor, health visitor, or paediatrician. See our Editorial Policy.

Quick answer: “Sleep regression” is a parenting term, not a medical diagnosis — and treating it like one sends parents down the wrong path. Yale pediatric sleep physician Dr. Craig Canapari puts it bluntly: there is a lot of talk about regressions happening at fixed ages, but “if you search the medical literature, however, you won’t find anything.” What the research does show is that learning a new physical skill temporarily costs your baby sleep. In one study from Sarah Berger’s lab at the City University of New York, infants in the middle of learning to walk had measurably worse sleep than an age-group estimate, with “temporary increases in night waking and movement at night.” So when your 8-month-old suddenly wakes four times and then pulls to stand in the crib a week later, that is not your sleep training failing and it is not a regression. It is your baby’s brain and body rehearsing. The practical response: hold your routine steady, don’t start a brand-new sleep plan mid-milestone, and make the one safety change that is genuinely urgent — stop swaddling the moment your baby looks like they’re trying to roll. Babies 4–12 months need 12–16 hours of sleep per 24 hours including naps, per the American Academy of Sleep Medicine.

The word “regression” is doing real damage

The framing matters more than it sounds. “Regression” tells a tired parent that something has gone backwards — that progress they earned has been lost, and that they need to do something to get it back. That is exactly the moment when parents start swapping methods every three nights, adding a dream feed, dropping a nap, buying a new product, or restarting sleep training from scratch. All of that adds new variables to a system that was about to settle on its own.

Dr. Canapari, who practises at Yale-New Haven Children’s Hospital, is direct about the evidence base: “There’s very little research on sleep regressions — as noted above, the term is poorly defined.” He compares the concept to “wake windows,” another idea with “a lot more traction among sleep consultants and on parent message boards than in the scientific literature.”

He also points to work by pediatric sleep researcher Jodi Mindell, who went looking for the age-specific spikes in night waking that the popular “4-month, 8-month, 12-month regression” calendar predicts — and, in survey data from thousands of mothers, did not find them. In Mindell’s words, as quoted by Canapari: “The data clearly indicate that there’s no specific age at which all of a sudden you see a shift in sleep.” Worth being precise here: Canapari describes that re-analysis as unpublished and not peer-reviewed, so treat it as an expert’s read of the data rather than a settled finding. But it lines up with what parents actually experience — the disruption is real, it just doesn’t arrive on a shared timetable.

What the research actually shows: new skills cost sleep

This is the part that reframes everything, and it comes from a body of peer-reviewed work rather than a single viral study.

Aaron DeMasi, Melissa Horger, Anat Scher and Sarah Berger studied 78 infants who wore an actigraph — a movement sensor — to measure physical activity during sleep. Their finding, published in Infancy: “Infants in the midst of walking skill acquisition had worse sleep than an age-group estimate.” Babies with more walking experience showed “more temporally sporadic movement during sleep and a steeper hourly increase in physical activity over the course of the night.” The authors describe the pattern as acquisition of new motor skills leading to temporary increases in night waking and movement at night.

The same group looked at what that disrupted sleep does the next day. In a study of 58 infants aged 10 to 19 months — all of them brand-new walkers, recruited within eight days of giving up crawling, split between the New York City metro area and the Tel Aviv and Haifa metro areas in Israel — they tracked night sleep against next-day problem solving. The result was more interesting than “bad sleep, bad performance.” Among infants who eventually solved the task, those with more wake episodes and lower sleep efficiency struggled more. But, as the researchers wrote, “contrary to our expectations, infants who never solved the task had the least fragmented sleep, indicating that an optimal level of fragmentation is needed for efficient problem-solving.”

Read that again, because it is the single most reassuring sentence a sleep-deprived parent of a nine-month-old can read this week: some of that night-time thrashing appears to be part of how the skill gets consolidated. The babies with perfectly smooth, undisturbed nights were not the ones doing the learning.

A 2026 review of infant sleep and development in the first year, published in Pediatric Research, sets the broader context: over the first year total sleep time decreases, night sleep consolidates with fewer awakenings, and healthy sleep parameters are linked to improved memory, language, executive function and sensorimotor skills. Sleep and skill-building are not competing for your baby’s night. They are the same project.

The one thing you must change the moment your baby starts rolling

Most of this article is permission to do less. This part is the exception, and it is not optional.

The American Academy of Pediatrics is unambiguous: “When your baby looks like they’re trying to roll over, you should stop swaddling them.” Not when they roll successfully — when they look like they’re trying. The AAP notes the risk of suffocation is higher if a swaddled baby rolls to their stomach, and that rolling usually happens around 3 to 4 months but can happen earlier.

The rest of the AAP’s rolling-stage guidance:

Keep placing your baby on their back for every sleep. But once they can roll comfortably both ways — back to tummy and tummy to back — you don’t need to keep flipping them back over.
The crib gets stricter, not looser, as they get mobile. No blankets, pillows, stuffed toys or bumper pads. A baby who can now roll can roll into those things, which could block their airflow.
Use a fitted sheet only. If you’re worried about cold, use a wearable blanket or an extra layer of clothing — generally one more layer than you’re wearing.
No weighted swaddles, weighted sleepers, or weighted blankets, and no weighted objects like rice bags inside a swaddle.
Room-share, don’t bed-share. The AAP recommends keeping the baby’s sleep space in your room for at least the first 6 months, and says room sharing can decrease SIDS risk by as much as 50%.

Note also that a sleep sack that does not pin the arms is fine to keep using for as long as you like — it’s the arms-restricting swaddle that has to go at the first sign of rolling.

What’s actually normal at 4–12 months

The American Academy of Sleep Medicine’s consensus figure is that infants 4 to 12 months should sleep 12 to 16 hours per 24 hours, including naps. The AAP explicitly backs this, stating on its HealthyChildren.org guidance (refreshed on 7 August 2026) that it “supports the AASM guidelines.” The CDC repeats the same number in its 9-month milestone guidance.

Sources do differ slightly at the edges. The Sleep Foundation, citing National Sleep Foundation figures, lists 12–15 hours for infants aged 4–11 months rather than 12–16. Treat 12–16 as a range to land inside, not a target to hit — and note that for babies under 4 months, both the AAP and AASM have declined to issue a recommendation at all, because sleep varies too much to generalise.

On night waking, be sceptical of anyone who tells you a precise “normal” number of wakings for a 6- or 9-month-old, because the authoritative bodies don’t publish one. What we can say: for infants, “sleeping through the night” is generally defined as at least six uninterrupted hours, and babies may wake briefly but be able to resettle themselves. One study cited by the Sleep Foundation — a commercial site, worth flagging — found about 38% of 6-month-olds weren’t sleeping through the night, with that figure dropping to under 28% by 12 months. In other words, a substantial minority of one-year-olds still wake, and they are not broken. As Nemours KidsHealth puts it, sleep problems are common in the second half of a baby’s first year, and even a baby who has been sleeping through will sometimes wake in the small hours, just as adults do.

The milestone map — with an important caveat

Here is the sequence to watch for, using the CDC’s checklists. One critical caveat first: since February 2022, the CDC has placed milestones at the age by which at least 75% of children can do them. These are not “average” ages — they are later than the averages you’ll see elsewhere, and they’re designed so that a single missed milestone is worth acting on rather than waiting out. The CDC is also clear these checklists are not screening or diagnostic tools.

By 6 months: rolls from tummy to back; leans on hands to support himself when sitting.
By 9 months: sits without support; gets to a sitting position by herself; makes a lot of different sounds like “mamamama” and “bababababa.”
By 12 months: pulls up to stand; walks holding onto furniture (cruising); calls a parent “mama” or “dada” or another special name.
By 15 months: takes a few steps on his own; tries to say one or two words besides “mama” or “dada.”
By 18 months: walks without holding on to anyone or anything; tries to say three or more words besides “mama” or “dada.”

One notable gap: crawling is not on the CDC’s checklists at all — it was removed, and the CDC now mentions it only as an activity to encourage. So if your baby skips crawling and goes straight to cruising, that is not a milestone missed.

Each of the transitions above is a plausible week for sleep to get worse before it gets better. If you can see the new skill emerging in the daytime, you have your explanation for the nights.

What to actually do during a milestone week

Change nothing structural. This is the wrong week to start sleep training, drop a nap, night-wean, or move your baby to their own room. Add a new variable now and you won’t know what caused what.
Give the skill daylight hours. If your baby is up at 2 a.m. practising pulling to stand, give them far more floor time to practise it at 10 a.m. Supervised awake tummy time and floor play is the outlet.
Keep the response boring and consistent. Whatever your usual settling approach is, run it unchanged. Consistency is what shortens the episode.
Lower the crib mattress before you need to. The night your baby first pulls to stand is the night the mattress height becomes a safety issue.
Protect the sleep environment. Dark room, consistent temperature, white noise if you use it. Don’t let a bad week become a bad month by also letting the room drift bright and hot in late-summer evenings.
Expect it to pass. The research language here is temporary. If it has genuinely become permanent, that’s a different conversation — see below.

If your baby is starting daycare this month, resist the urge to blame the milestone week on daycare or vice versa. If you’re also managing an older child’s schedule shift before preschool, our 14-day toddler sleep reset covers the 1–5 age band, which works differently because of naps. And if the question underneath all of this is whether to sleep train at all, our review of what the 2026 research says about cry-it-out is the more useful read.

When it isn’t a milestone

Talk to your pediatrician if the disruption doesn’t resolve after a few weeks of a consistent routine, or if you notice any of the following, which point past normal skill-building:

Loud snoring, gasping, or pauses in breathing during sleep
Waking unrefreshed, or daytime sleepiness that doesn’t shift
Pain signals — pulling at ears, fever, inconsolable crying rather than restless practising
A loss of skills your baby previously had, which is different from sleep changing while a new skill arrives
Your own exhaustion reaching the point where you’re struggling to function, which is a medical issue worth raising in its own right

The bottom line

The nights got worse because your baby is about to do something new. That’s the most likely explanation for a 4-to-12-month-old whose sleep falls apart for a week or two, and it’s better supported by the evidence than the fixed “regression calendar” doing the rounds on social media. Peer-reviewed work from Berger’s lab found that babies in the thick of learning to walk slept worse — and, strikingly, that the babies with the least fragmented sleep were the ones who never solved the problem-solving task at all. Hold your routine steady, give the new skill room during the day, and make the one non-negotiable change: the moment your baby looks like they’re trying to roll, the swaddle comes off for good. Then wait it out. The word you want isn’t regression. It’s rehearsal.

Sources

🌿 Join Our Parenting Newsletter

Get weekly evidence-based parenting tips, EU-specific guides, and product recommendations — straight to your inbox.