The 4-month sleep regression, calmly explained

20 August 2026 · 6 min read · by Aleksandr Zhadko, developer of Rostik and dad of two

You had a baby who slept. Not perfectly, but predictably: down at seven, one or two wakes, back down in ten minutes. Then, somewhere around sixteen weeks, that baby was replaced by one who wakes every forty minutes all night and naps for exactly one cycle. Nothing changed. No teeth, no illness, no new routine.

This is the best-documented sleep shift of the first year, and the name it goes by is misleading in a way that matters.

It isn't a regression. It's the sleep system growing up

Newborns have two sleep states and drop almost straight into deep sleep. Between roughly three and five months, that reorganises into the adult architecture: cycles of about 40–50 minutes, beginning in light sleep, with a brief surfacing at the end of each one.

Adults do exactly the same thing — we surface between cycles several times a night, roll over, and go back down without ever remembering it. A four-month-old surfaces, finds that the conditions have changed since they fell asleep (the rocking stopped, the breast is gone, they're in a different room), and calls for help.

The important consequence: this change is permanent. It is not a phase that reverses. What passes is the difficulty of adjusting to it — usually in 2–6 weeks.

What it looks like

Sign	What's happening
Waking every 45–90 min	Surfacing at the end of each cycle and not resettling alone
Catnaps of 30–45 min	One cycle, then awake — daytime sleep is hardest to bridge
Fighting the same routine	Longer settling with methods that worked a fortnight ago
Feeding more at night	Partly hunger, largely the fastest available route back to sleep
More alert, more distracted	Rolling, grabbing, and noticing the room — the same leap that breaks daytime feeds

Timing: most often 3–5 months, occasionally as early as 8 weeks. It tracks development, not the calendar — babies born prematurely usually meet it closer to their corrected age.

What actually helps

Because the cause is a permanent change, the aim is not to restore the old sleep — it's to make the gap between cycles crossable.

Shrink the difference between falling asleep and waking up. If a baby goes down being fed or rocked and surfaces alone in the dark, that gap is what they protest. Putting them down drowsy but aware of where they are — even occasionally — narrows it.
Give a moment before you go in. A minute or two of noise at a cycle boundary often resolves itself. Going in instantly at the first sound teaches the surfacing to become a full wake.
Fix the room, not the baby. Proper darkness and steady white noise remove the two most common reasons a light-sleep surfacing turns into an awake baby.
Watch wake windows, not the clock. They stretch fast at this age — typically to about 1h45–2h30 by four months. Overtiredness makes cycle transitions worse; see wake windows by age.
Feed if feeding works. A night feed at four months isn't a habit you'll regret. Nobody arrives at school being fed back to sleep.
Split the night in shifts. Two parents each getting one unbroken block beats both being woken six times.

What doesn't help

Cutting daytime sleep. The single most common instinct, and it reliably makes nights worse.
Changing method every third night. Nothing gets a fair test, and the baby has one more variable to adjust to.
Starting solids early to "fill them up". There's no good evidence it improves sleep, and the guidance is around six months for other reasons.
Comparing. The baby on the forum who sleeps through at four months exists. So does the one who doesn't until two. Both are common.

When it isn't the regression

Sleep can fall apart for reasons that have nothing to do with development. Speak to your health visitor or GP if broken sleep comes with: a fever, or a baby who seems unwell; a stall in weight gain or a drop in wet nappies; pulling at an ear, or crying that sounds like pain rather than protest; noisy or laboured breathing, snoring or pauses in breathing during sleep; or if you're struggling to cope — that last one is a medical reason on its own, and health visitors would far rather hear it early. If a baby seems seriously unwell, call 111, or 999 in an emergency.

How to know it's passing

The reassuring part is that improvement is visible before it's felt. Weeks two and three feel identical from the inside, because you're measuring by how tired you are — and that keeps climbing even as the nights improve.

What actually changes first is the shape of the night: the wakes stay frequent but get shorter, then the first stretch after bedtime starts to lengthen again, then the 3 a.m. cluster of wakes thins out. Seeing that on a single view of the night is worth more than any single statistic, and it's the difference between "still awful" and "moving in the right direction" — which is the difference that decides whether you keep going with what you're doing or change everything again. Our sleep-by-age guide has the totals to compare against, but the shape is the part that tells you it's ending.

See the night as a shape, not a list of wakes

Rostik draws every night on one clock face, so the first long stretch coming back is obvious — even on the mornings when it doesn't feel like anything improved.

Try Rostik free for 7 days

Rostik is a parental logging app, not a medical device. This article is general information, not medical advice. For safer sleep guidance, follow the advice of the Lullaby Trust and the NHS, and speak to your own health visitor or GP about your baby.