How long does the 4-month sleep regression last?
Just as a longer stretch had started to appear at night, the nights feel broken again.
There is no well-sourced fixed number we can give you. The popular pages we checked all offer a range, none shows the measurement behind its figure, and one page gives two different answers on the same site. This article checks that number rather than setting a new deadline for you.
The short answer is not “two to four weeks”
Across three widely ranked pages, we found four different wordings:
- a few days to a few weeks
- 2–6 weeks, in one page’s summary and main body
- a few days to roughly two weeks, in that same page’s FAQ
- a few days or weeks at most
One page uses the broad “few days to few weeks” wording. Pampers UK uses the 2–6 week wording in its summary and body, then uses the shorter “few days to roughly two weeks” wording in its FAQ. A third UK retailer page uses the “few days or weeks at most” wording.
That disagreement matters, but its meaning is narrow. It does not prove that a familiar 2–4 week estimate is false. It shows that the pages checked do not display a source, dataset or measurement method beside their numbers, so a tired reader cannot tell where any of them comes from.
Adding a fifth estimate would repeat the problem. Without a visible study that follows individual babies night by night and defines what counts as the start and the end, any precise countdown is an editorial choice, not a measured finding.
What is meant to end?
One page creates a useful puzzle for a careful reader. In its explanation of causes it calls the underlying change permanent, then in its duration section it says it will pass in just a couple of weeks.
Those two sentences may be reaching for a distinction between a lasting change in how sleep is organised and a shorter spell of unsettled nights. That would be an understandable distinction to want to make. The page itself does not set it out clearly, so the reader is left to guess what exactly is meant to finish.
We raise this as a reading question, not a verdict on the science or on your baby. If the organisation of sleep is described as lasting, while the disrupted nights are described as passing, a duration figure needs to say which one it is timing. None of the passages we checked does that.
Four months is a window, not a switch
The primary record used here does not measure night waking at all. It reviews how sleep looks on brain-wave recordings in early life.
Grigg-Damberger and colleagues describe different features becoming recognisable to a trained scorer at different ages, counted as post-term age. Post-term here means age counted from around the due date, rather than the birth date. Scorable means recognisable to someone trained to read the recording, not visible to a parent watching at night.
In that review, sleep spindles are typically present by 2 to 3 months post-term. The stages now labelled N1, N2 and N3 are usually scorable by 4 to 4.5 months post-term. K complexes are usually present by 6 months post-term.
That staggered timeline is often flattened into a single switch at four months. The review does not support that flattening. It describes measurable recording features, not behaviour, night waking, or how long a named regression lasts. The authors also caution that sleep recordings in infants cannot reliably predict behaviour or outcome.
The pages checked include their own cautions, and they deserve credit for them. One notes that not every baby shows a visible 4-month regression. Another says four months is an average, not a switch that flips at exactly 16 weeks. Variation is part of their account, even where their duration figures sound precise.
What the NHS pages do - and do not - call it
The two national NHS baby-sleep pages we checked use plainer language. Helping your baby to sleep says babies have their own pattern of waking and sleeping, and that at 3–6 months some may sleep for 5 to 8 hours or longer at night, but not all. Your baby’s sleep patterns describes sleep patterns changing as a baby grows, with some 3–6-month-olds sleeping for 8 hours or longer, but again not all.
Neither page uses the phrase “sleep regression”, and neither offers a “4-month regression lasts X” rule. That is a finding about those two pages only. A separate NHS trust search-result excerpt uses “sleep regression” in material for children aged 3 to 5 years, so this is not a claim about every NHS service.
The contrast is practical rather than technical. Official pages describe changing patterns and changing routines. The search term asks for something tidier: a named phase with a start date and an end date. That tidiness is what the evidence does not supply.
The useful answer for tonight
Our reading, not a measured finding, is this: the pages checked have not given you a trustworthy end date. Treat any duration you see as an unverified range, not a deadline you or your baby are failing to meet.
Nothing here supports promising that a particular action will end or shorten a maturational change, and this is an evidence check rather than a sleep-training guide.
Keep reading: for why the word “regression” itself misleads, see the 4-month sleep regression: what the label gets wrong, and for the other ages you will see on charts, sleep regression ages: why the charts don’t agree. If night feeds are part of the picture, our guide to dream feeds and night checks covers keeping the room dim.
References
- NHS - Helping your baby to sleep - checked national baby-sleep guidance; last reviewed 28 January 2025.
- NHS Best Start in Life - Your baby's sleep patterns - describes changing sleep patterns without using the regression term.
- Grigg-Damberger M et al. (2007) - The visual scoring of sleep and arousal in infants and children - review of when EEG features become scorable by post-term age; behaviour and duration not measured.