The 4-month sleep regression: what the label gets wrong
Your baby slept in longer stretches. Now, suddenly, they are waking again.
Something real is changing in infant sleep around four months, and it is not your baby going backwards. The word "regression" gets the direction wrong, the three top-ranking pages we checked do not show where their duration figures come from, and one of them calls the change both permanent and over in a fortnight. This article does not tell you how to make your baby sleep.
That is our reading of what the evidence actually supports: maturation, not a fault, described by a label that confuses more than it explains.
Something is changing, and it is not going backwards
A sleep cycle is a repeating sequence of lighter and deeper sleep stages. Around the first half-year, that sequence starts to look more adult-like.
The sleep-medicine record puts that change on staggered timelines, not on one night at sixteen weeks. In an evidence-based review by the American Academy of Sleep Medicine Paediatric Task Force, Grigg-Damberger and colleagues reported that the brain-recording features used to score adult-style non-REM sleep become recognisable at different ages. Sleep spindles are typically present by 2 to 3 months post-term, slow wave activity usually by 4 to 4.5 months, and K complexes usually by 5 to 6 months. The stages N1, N2 and N3 are usually scorable by 4 to 4.5 months post-term.
Post-term means counted from the due date rather than the birth date, and "scorable" matters here. This is about when a trained scorer can see a pattern on brain recordings from term-born infants. It is not a statement about night waking, about behaviour, or about anything you can watch for in the cot.
That distinction is the useful reframe. What is changing is the structure of sleep itself, becoming more grown-up in pieces across roughly two to six months. To be fair to the pages that rank highest, the most careful of them say this too: not every baby shows a visible disruption, and four months is an average with no magic switch at exactly sixteen weeks.
Nobody can tell you how long it lasts
The clearest contradiction needs no scientific interpretation. One UK page we checked on 7 September 2026 describes the change as "a permanent change in sleep" and, on the same page, says that "it will pass in just a couple of weeks."
Both cannot mean what they seem to mean. If the underlying change is permanent, what ends is not the change.
The duration figures are no firmer. Across the three top-ranking pages we checked that day, we found four incompatible answers:
- a few days to a few weeks
- a few days or weeks at most
- 2-6 weeks, in the body and summary of one page
- a few days to roughly two weeks, in the same page's own questions section
None of the three pages cited a source for its figure. That does not prove any figure false; absence of shown working is not disproof. It does mean you are being given numbers whose source or measurement is not shown, including two different numbers on one page.
This is why we are not giving you a fifth figure of our own. An article whose point is that the published durations are unsourced should not add another unsourced one.
The word itself is the problem
Durham University's Baby Sleep Info Source, BASIS, puts it plainly: "sleep regression" has "no biological basis - it's simply part of how babies grow and learn," a statement it attributes to Ball (2025).
That sentence is about the term, not about your baby. It does not deny that sleep changes. The brain-recording literature above shows that it does, gradually and in pieces. Hold those two ideas together: the maturation is real, and "regression" is still the wrong name for growing forwards.
The national guidance uses plainer language. The NHS national page Helping your baby to sleep, last reviewed on 28 January 2025, and the NHS Best Start in Life page on your baby's sleep patterns, both checked on 7 September 2026, talk about sleep patterns changing and routines needing to change as babies grow. Neither uses the phrase "sleep regression."
That finding stays narrow on purpose. It covers those national baby-sleep pages only. A search-result excerpt for at least one NHS trust leaflet uses "sleep regression" for children aged 3 to 5 years, which is a different age and a different document, so "the NHS never uses the term" would go further than the checking supports.
What this means for you tonight
Our reading, labelled as such: you have not caused this, there is no method you are missing, and nothing you do tonight ends a maturational change.
That can feel thin at 3am, when you arrived wanting something to try. We are leaving the fix out deliberately, because the record we checked supports an explanation but not a solution.
The next useful step the evidence supports is permission to stop searching for the end date. The numbers have no shown source, the label has no biological basis, and the change underneath is sleep growing up, not breaking down.
Keep reading: if you still want to know what the duration figures are based on, we looked in detail at how long the 4-month sleep regression lasts. If your baby will only nap on you right now, see contact napping: what it is, is it safe, and when does it change. For night feeds in a dark room, our guide to dream feeds and night checks covers keeping the light low.
References
- Grigg-Damberger M, Gozal D, Marcus CL, et al. The visual scoring of sleep and arousal in infants and children. Journal of Clinical Sleep Medicine, 2007;3(2):201-240. PubMed 17557427
- BASIS / Durham Infancy & Sleep Centre, Durham University. Infant Sleep Biology - "no biological basis" statement attributed to Ball, H. (2025). Last reviewed July 2025. Accessed 7 September 2026.
- NHS. Helping your baby to sleep - national baby-sleep guidance; does not use the term "sleep regression." Last reviewed 28 January 2025. Accessed 7 September 2026.
- NHS Best Start in Life. Your baby's sleep patterns - covers 3-6 months as an age band; does not use the term. Accessed 7 September 2026.