Sleep regression ages: why the charts don't agree

You are on your phone in the dark, trying to work out which one this is. 4 months? 8 months? 12 months?

You came for the list of ages. There are many lists, they disagree, and between them they name almost every month of the first two years. We looked at the eighteen pages ranking on the first two pages of this UK search, captured on 7 September 2026, and checked what each one actually publishes.

Between them, those eighteen pages name sleep-regression ages from six weeks old to two years old, and their lists claim nearly every month in between.

What happens when you put the lists side by side

One page gives you two broad bands, 3 to 6 months and 6 to 12 months, and nothing month by month. Three pages give the same fixed six: 4, 6, 8, 12, 18 and 24 months. Another gives a four-band chart of 3 to 4, 7 to 10, 14 to 18, and 24 months, then adds separate sections for 3, 4, 5, 6, 8, 10, 18 and 24 months. Past that, no two lists match.

The gaps fill in from elsewhere. One page lists 6 to 8 weeks. One gives 6 to 9 months and adds that a regression can “strike at any age, be it 4 months, 10 months, 12 months old or beyond”. Another includes 8 to 10 months, another 8 to 11 and 11 to 16 months, another 12, 15, 18 and 24 months, and one puts the “19, 20, and 21 month variations” at 18 months. Another marks toddler leaps spanning 14 to 19 and 17 to 26 months.

Checked month by month, the lists and bands leave almost nothing out between six weeks and two years. Two stretches are thinner than the rest: a few weeks around the two-month mark, which no list claims, and 22 to 23 months, where the only cover is one page’s 17-to-26-month developmental window rather than a stated regression age. The first six weeks are outside the span, and one page rules them out by name. Everything else has somebody’s label on it.

That is what the set does together, not what any single page says. Two of these pages add that a regression can happen at any age.

If nearly every month can be a regression age, the list is not narrowing anything down. It tells you a page exists for the month you are in, not that the month is the reason.

Three pages disagree with themselves

The sharpest problems are not between pages. They are inside them.

One medically reviewed page opens with: “The consensus among sleep specialists is that there are roughly six stages of sleep regression at different ages—when the baby is 4, 6, 8, 12, 18, and 24 months old.” The next paragraph says: “While sleep regression is a common term, there is no medical consensus on its definition. Additionally, little published research exists on infant sleep regression.” Further down it introduces its stage-by-stage section with a five-age list of 4, 6, 8, 12 and 24 months, then gives an eighteen-month section anyway.

One page publishes dedicated articles for the 5-month and the 6-month regression while stating on the same page: “While 5 months is not a common age for a developmental regression, a sleep regression can occur at any age”, with near-identical wording about 6 months.

On a third, the “Typical duration” cell of one chart row begins “Permanent change”.

Then there is the sourcing circle. One UK page carries a six-age list attributed inline to another page on the same results page, dated 2024. That second page cites nobody for its ages. Checked page by page, not one of the eighteen attaches a study citation to the ages it publishes.

The set is honest about this in places, and those passages deserve credit. One page says plainly: “It’s not a scientific or medical term, but you may hear it from parents, sleep consultants, and pediatricians.” Another says the ages are best understood as developmental windows, not fixed events. Both sentences sit on pages organised around fixed ages.

What the one cited study actually says

Only one page in the eighteen attaches a study to the age question, and it uses it to say the ages are not evidenced. Its note reads:

“Despite common belief, there’s no reliable scientific evidence to support the idea that every baby undergoes a regression in sleep at a given age.”

The footnote is to Mindell and colleagues, in the Journal of Sleep Research, 2016: app-recorded data covering 841 children aged birth to 36 months and 156,989 caregiver-logged sleep sessions over 19 months.

That study names no regression ages. It describes sleep patterns as “ever-changing across the first few years with wide individual variability”, with longer night-time sleep beginning to develop more clearly at 5 to 6 months. Being caregiver-recorded through an app, it describes population patterns rather than any individual baby. It was not designed to test the regression ages, and it is not proof that regressions do not exist. Its value here is narrower: the one study cited anywhere in this set on the age question is cited against fixed ages, not for them.

One page in the eighteen goes further and rejects the word. Australia’s national parenting website, the Raising Children Network, says “‘regression’ isn’t a good word to describe what’s happening”, because changes in baby sleep over the first 12 months are developmentally normal, “just like changes in other areas of baby development, like feeding, moving and communicating”.

An older study points the same continuous way. Goodlin-Jones and colleagues filmed 80 infants overnight in four age groups (3, 6, 9 and 12 months) and reported that “most infants woke during the night at all ages observed”, with wide variability between babies and half of the 12-month-olds still typically needing a parent to get back to sleep. It is small, cross-sectional, American and from 2001, and it sampled only four ages, so it says nothing about clustering. It supports the modest point that waking across the first year is common rather than banded.

What to do with that

Our reading, not a medical finding: stop trying to locate your baby on a chart that does not hold still.

If the four-month label is the one you are seeing, our companion audit of the 4-month sleep regression looks more closely at what the phrase gets wrong.

We are not offering a corrected list of our own, and we are not offering a method to make the waking stop. That is not a gap in this article. Given what the eighteen pages show — and fail to show — for their ages, a neater list from us would repeat the problem, and a method would go beyond what this evidence can support.

If your baby seems unwell, is feeding much less than usual, or something feels wrong to you, that is a question for your GP or health visitor rather than another chart.

You are allowed to let “which regression is this?” go unanswered. A run of difficult nights in the first two years is a stretch of time to get through, not a number to identify.

Keep reading: if the bigger change in your house is naps rather than nights, see the one-nap transition: what the sources actually agree on. If the problem is early starts, our checklist for a toddler waking at 5am covers what to check first.

References

1. Mindell et al., Development of infant and toddler sleep patterns: real-world data from a mobile application, Journal of Sleep Research, 2016 — 841 children and 156,989 recorded sleep sessions, naming no regression ages; the study cited by the one ranking page that attaches evidence to the age question.
2. Raising Children Network (Australia), Baby sleep at 2-12 months, updated 17 July 2026 — the Australian national parenting website’s position that “regression” is not a good word for changes that are developmentally normal.
3. Goodlin-Jones et al., Night waking, sleep-wake organization, and self-soothing in the first year of life, Journal of Developmental and Behavioral Pediatrics, 2001 — overnight video of 80 infants at 3, 6, 9 and 12 months, reporting waking at every age observed and wide variability between babies.