Cradle & Care

When it happens · 7 min read

Your baby was sleeping, and suddenly isn't

The specific cruelty of this one is that it usually arrives after a good stretch. Something was working. Then it stopped, for no reason anyone can point at, and every search result tells you it's a regression and offers to sell you a plan.

The word regression is doing a lot of damage here. Around four months, nothing goes backwards — the sleep itself changes shape, permanently, and what worked before stops working because the thing it was working on no longer exists.

Last updated August 3, 2026

The four-month one isn't a regression

Newborn sleep has two states and drops almost straight into the deep one. Somewhere around three to four months that reorganises into cycles closer to an adult's, with lighter phases and brief surfacings between them. Babies don't have regular sleep cycles until around six months, and the change is permanent — this is the new baseline, not a phase to wait out.

What that means practically: they now half-wake between cycles and check their surroundings. If they fell asleep being fed, rocked or held, and surface alone in a dark room, the conditions have changed and they call for them back. Nothing has gone wrong; the wake-up is the maturation, and the crying is about the mismatch.

It's worth knowing that waking is still normal at this stage even when nothing is disrupted. A four-month-old waking one to three times a night for feeding or comfort is inside ordinary. By six months, a bit over half of full-term babies have consolidated night sleep — which also means a large minority haven't, on time, with nothing wrong.

The later ones track skills, not sleep

Around eight to ten months, and again near eighteen months and two years, sleep often breaks again. These tend to sit on top of something new: crawling, pulling to stand, walking, or a leap in language. A brain busy consolidating a new skill rehearses it at 2am, and a baby who can suddenly stand often stands in the cot before they work out how to get back down.

The other engine is separation awareness. Understanding that you continue to exist when you leave the room is a genuine cognitive gain, and its side effect is caring much more about your absence at bedtime than they did a month ago. That one usually looks like protest at the leaving rather than inability to sleep.

These pass on their own in a way the four-month change doesn't, because they're driven by a specific developmental push rather than a permanent restructure. A fortnight to a month is typical. It's genuinely worth writing down when it started, because in week two it always feels like the new normal and it usually isn't.

What actually helps, and what only sells

The single highest-yield change is where they are when they fall asleep, rather than any schedule. Put down drowsy but still awake and the conditions at the start of the night match the ones at 2am. That's the whole mechanism behind most sleep advice, described without the branding.

Protect the front end of the day too. An overtired baby fights sleep harder than a tired one, and light in the morning with dimness in the evening is what sets the circadian clock — which switches on properly somewhere around 12 to 16 weeks and is doing real work by then.

Nothing you can buy fixes this. There is no product that resolves a sleep-cycle restructure, and the sleep-gadget market is largely built on the fortnight when you'd try anything. If you're weighing sleep training, most pediatricians put developmental readiness at around four to six months, and the honest position is that it's a personal call rather than a medical instruction.

When it isn't the sleep

Sleep is often the first thing to wobble when something else is going on, so the useful question is whether it broke alone. Any big transition — a new sibling, a new bed, starting daycare — reliably disturbs sleep for a while, and that's weather rather than a problem to solve.

What's worth a call rather than a plan: sleep disruption alongside fever, tugging at an ear, a change in feeding or nappies, breathing that looks effortful, or the pattern of getting better and then clearly worse again. Snoring most nights, or pauses in breathing, are worth raising specifically rather than filing under bad sleep.

And the part that gets left out: your own sleep debt is a medical fact, not a character test. If you're falling asleep involuntarily, driving while impaired by tiredness, or feeling frightened by your own thoughts at 3am, that's a reason to call your own doctor. It is the most under-reported symptom in this whole period.

Common questions

How long does the four-month sleep change last?
The disruption usually settles over a few weeks, but the underlying change is permanent — sleep cycles don't go back to newborn shape. What settles is everyone's adjustment to the new pattern, not a return to the old one.
Is it a growth spurt?
Sometimes, and a genuine feeding increase for a few days is common. But a sleep change that lasts more than about a week is usually the cycle restructure or a developmental leap rather than hunger alone.
Should we start sleep training?
It's a personal decision, not a medical requirement. Many pediatricians consider four to six months the point of developmental readiness. Plenty of families never do it and end up in the same place a year later.
My baby was sleeping through and now wakes to feed again. Have we gone backwards?
Not necessarily. Night feeds returning during a leap or an illness is common and usually temporary. Worth a call if it comes with weight changes, fewer wet nappies, or feeding that has become difficult rather than just more frequent.

Written from the same sources we cite everywhere else, and it is not a substitute for your own doctor. If something feels wrong to you, that instinct is a good enough reason to call — you know this child better than any page does.