The 4 month sleep regression: what's actually happening, and what helps

Last week she napped for an hour and a half, twice a day, and did a six-hour stretch at night. This week she’s sixteen weeks old, every nap ends at thirty-two minutes on the dot, and last night she was up at one, at three, and at half four, each time looking mildly surprised to be awake. A friend says “four month sleep regression” as if that settles it. It explains some of it, but it doesn’t tell you what to do at half four.

In short: the 4 month sleep regression is a change in how your baby sleeps, not a step backwards. By around four months, counted from the due date, infant sleep reorganises into distinct lighter and deeper stages, and we think that’s why many babies start surfacing between cycles more noticeably: more night wakings, shorter naps, a baby who slept well suddenly not. Nobody has measured how long it lasts; it varies and usually eases as they learn to drift back down. What helps most is watching daytime wake windows and settling the same way every time, so 2am looks like bedtime did.

This article is general information about baby sleep patterns. It isn’t medical advice. For any concern about your baby’s sleep, feeding or health, talk to your doctor or health nurse.

The regression is a progression, seen from the outside

Newborn sleep is fairly simple: active sleep, the twitchy kind with eyes moving under the lids, and quiet sleep, the still kind, in short cycles around the clock. The electrical signature of a distinct lighter stage of non-dreaming sleep, the sleep spindle, is typically present by two to three months post-term, and the lighter and deeper stages we recognise in older children and adults are usually recognisable by around four months, counted from the due date.¹ Your baby’s sleep is being rebuilt on a more grown-up plan.

A more grown-up plan comes with a more grown-up feature: brief surfacing between sleep cycles. In lab recordings of healthy young adults, brief arousals show up around ten times an hour, and we almost never remember them.² A baby who has just acquired these lighter stages surfaces too, but hasn’t yet learned to roll straight back down. So they notice they’re awake, and then that the thing that was there when they fell asleep (an arm, or the motion of the pram) isn’t there now. That noticing is, we think, most of what parents experience as the regression.

The maturation of sleep stages is well described. That this is why some babies start waking more is a reasonable reading, widely repeated and hard to prove from the outside. The reviews of infant sleep across the first year don’t describe a dip at four months, and none was looking at the month-by-month resolution that would show one; one found the biggest gains in sustained night sleep happen across the first four months, with little change from four to nine.³ So the honest version is that the machinery is changing, some babies find that disruptive, and “regression” is the parent’s-eye view of a progression.

Four month sleep regression signs that are still ordinary at this age

These are the things parents describe most often.

More night wakings, sometimes on a rhythm. The classic complaint is a 4 month old waking every hour or so for part of the night, or more precisely at the end of every sleep cycle. Some wake once or twice more than before; some seem to wake at every join.

Naps that end at the same short length. A baby who used to nap for an hour or more suddenly wakes at thirty or forty minutes, often crying, often still tired. It’s the daytime version of the same surfacing. An app-log study of hundreds of families found average nap length dipping between one and five months before rising again, which fits what parents describe, though the paper never names a regression.⁴

Fighting sleep. Arching, crying at the cot, needing far more rocking than a fortnight ago. Often it’s simply overtiredness, the short naps stacking up.

Wanting to feed more at night. Sometimes hunger, sometimes the easiest way back down; more below.

A baby who was easy and suddenly isn’t. This is the one that unsettles people, and it’s why the word “regression” stuck. A 4 month old suddenly not sleeping after weeks of sleeping well feels like something has gone wrong, when usually something has gone on.

Night waking at this age sits inside the normal range: a systematic review of infant sleep norms pooled 34 diary and questionnaire studies and found night wakings in every age band, from none to three or so a night in early infancy, still up to two or more a night well into the second year.⁵ The age is a rough centre too; if your baby was born early, count from their due date.

How long does the 4 month sleep regression last?

The candid answer is that nobody has measured it well. The pages that say “two to six weeks” don’t cite anything, and we couldn’t find a study that does, so we won’t give you a number we can’t stand behind.

The sleep-stage change is permanent. What passes is the disruption, as your baby learns to drift back down without a full production each time. For some babies that’s a rough fortnight; for others a slow improvement over a couple of months, tangled up with a cold and the first tooth. Real-world logs suggest patterns look more settled around five to six months, when longer night stretches begin,⁴ and the first-year reviews describe most babies managing an eight-hour stretch by six months.³ Neither is a promise about your baby, but both say the direction of travel is good.

So if you searched “4 month sleep regression how long”, the answer is: it varies, usually weeks, and it resolves toward better sleep rather than back to the old pattern.

Wake windows and one repeatable settle help more than any method

These make the new machinery easier to run; drop any that don’t suit you.

Watch the wake windows, especially the first one. The stretch a baby can comfortably be awake lengthens as they grow, and at this age it’s short. For babies of twelve to seventeen weeks we model it at roughly 75 to 120 minutes, shorter for the first window of the day and a little longer before bedtime. It matters more now because short naps make a baby tired sooner than the clock suggests. The next sleep depends on when they last woke and on how much they’ve slept today and last night; weigh both to stay out of the overtired zone where settling gets hard.

Protect the first nap. If the morning nap is the one that comes most easily for your baby, that one decent nap puts a floor under the rest of the day.

Pick one way of settling that you’re happy to repeat at 2am. It doesn’t much matter what: a feed, a cuddle, a phrase you say. What matters is that it’s the same thing, done the same way, by whoever is on. Our reading is that a baby who surfaces at the end of a cycle is checking whether the world is as they left it, and a consistent settle is what makes the answer yes. At half four, or whenever yours surfaces: do the same settle you did at bedtime, note the time, and decide nothing else until morning. The pattern is easier to read at breakfast than at 4am.

Offer the same conditions for falling asleep and for resettling. If your baby falls asleep in a dark room with white noise and wakes at 1am in the same dark room with the same white noise, the surfacing is less of an event. If they fell asleep on your shoulder in a bright living room and wake alone in a cot, the noticing is louder. Falling asleep on you is fine; choose, when you can, the version you can reproduce at 1am.

Consider an earlier bedtime on bad-nap days. The last wake window before bed is, we think, the one most often stretched too far. From about fourteen weeks our model puts typical bedtime between half six and eight (before that it sits later, nearer eight to ten), and on a day of thirty-minute naps the earlier end is worth a try.

Accept that daytime sleep may be messy for a while. Our model expects three or four naps at this age, and short ones are common. You don’t have to fix the nap length, only stop the short naps turning into a long, overtired afternoon, which mostly means offering the next sleep sooner than feels intuitive.

Why a quick fix tonight can cost you next month

A few things make the next month harder without making this week better.

Adding a new prop you’ll have to remove later. When a baby is waking every hour, the temptation is to reach for whatever works tonight: twenty minutes of rocking, a particular song on repeat, a walk in the pram. Some of those you’ll happily keep for months. But if you’d rather not be doing it three months from now, this is a poor time to start it, because, as we read it, your baby is currently learning what falling asleep involves.

Swinging between three methods in three nights. Every approach, gentle or otherwise, relies on consistency. Something on Monday, something else on Tuesday and a third thing on Wednesday guarantees none of them worked, and hands your baby three different worlds to wake up into. If you change something, give it a real run before you judge it.

Assuming it’s hunger every time. At four months a baby may genuinely need a night feed or two, and if they feed well and settle after, that’s a feed, not necessarily a sign that bedtime was too early or the naps ran too long. But a baby who takes a few sucks and drops off, then does it again ninety minutes later, may be using the feed to settle rather than to eat. Neither is wrong, and whether your baby is getting enough is a question for your health nurse. A log of feeds against wakings at least shows the pattern faster than a guess.

When the 4 month sleep regression isn’t the explanation

A story that fits everything explains nothing.

If your baby was never a good sleeper, there’s nothing to regress from; the same maturation is happening, but the label doesn’t apply.

If your baby barely registers it, that’s common. Plenty pass through this window with a couple of worse nights, and there’s no reason to go looking for a regression because the internet said one was due.

If something else changed in the same week, a cold, a new room, a return to work, the first tooth, a new skill like rolling, that’s at least as likely the cause. Four months is busy, and the regression takes the blame for a lot of it.

If you’re anxious about the numbers, log less. Three days is usually plenty to see a pattern; a record you’re refreshing at 3am to see how bad it is has stopped helping.

You may not need a tracker at all. If one of you does every night and a note on the fridge already answers “when was the last feed”, you can skip the sharing without guilt, and a notebook can show you the pattern behind the wakings too; an app just does the sums for you.

When to talk to your doctor

These aren’t sleep stories, and they’re worth a call even if a sleep explanation is on offer:

• A fever, or a baby who seems unwell rather than just tired
• Refusing feeds, feeding much less than usual, or fewer wet nappies
• Not gaining weight as expected
• Anything about their breathing during sleep that worries you
• A change in sleep that comes with vomiting, a rash, or unusual floppiness or irritability

“Something feels wrong” is reason enough to call.

Know where the day is, even on the messy ones

This is what Tiny Sleeps is for. One tap when your baby goes down, one when they wake, and the day’s schedule shifts around what actually happened: the next nap is worked out from when your baby last woke and how much they’ve slept today and last night, so it moves when a nap runs short, bedtime moves with it, and the wake windows are sized for your baby’s age. Feeds and nappies are two or three taps, so at 3am you can see whether the last feed was at one or at two. Both parents, and the grandparents or the nanny, see the same log when they open it, and it works offline and catches up later.

A notebook works. If you want the sums done for you and the same log on both phones without retyping at 3am, Tiny Sleeps is built for that; “one, three and half four” looks random until you can see that all three came the same stretch after she’d gone back down.

Download Tiny Sleeps on the App Store for iPhone.

Related

• The 8 month sleep regression, and why it’s a different animal
• How to split the night with a newborn
• Baby sleep tracker for two parents: how to hand over without the group chat

Sources