Sleep Regression at 4 Months

The change at four months is the one on the list that is not a regression at all — it is a maturation, and it does not go back. Babies do not have regular sleep cycles until about four months of age, and once those cycles arrive a baby starts surfacing between them: she wakes properly where she used to sail straight through. Nothing has slipped backwards. What you have is a baby whose sleep is now built roughly the way yours is, with none of the practice you have at getting yourself back down.

The later ages on the list are mostly skills being learned, and they ease when the skill lands. This one is a permanent change in how sleep is put together, so what passes is not the change but her difficulty with it.

What is published, and what is not

The honest part first, because it makes the rest easier to read. No publisher this site is allowed to stand on — not the NHS, not the American Academy of Pediatrics, not a paper on PubMed — uses the word regression, prints the famous list of months, or attaches a duration to any of it. There is no measured four-month event, and there is no measured length for one.

What is published is narrower and more useful. The AAP states that babies do not have regular sleep cycles until about four months of age, and that a good sleeper at this stage is a child who wakes frequently and can get herself back to sleep rather than one who sleeps ten hours unbroken — frequent waking being developmentally appropriate. Put those two together and the four-month story stops being a fault to fix and turns into a description of an ordinary month.

So read the ages as a folk map rather than a finding. Parents drew it, nobody measured it, and it is still worth having: it marks the months when nights commonly get harder. The regression timeline lays out all six ages the internet argues about and keeps the same separation on each — what is developing, what tends to happen to the nights, and which of those two anybody has measured.

Why this one sticks when the later ones pass

A young baby's sleep is not organised into the tidy alternation of light and deep stages that adult sleep has. It runs in long, shapeless stretches, which is part of why a small baby can be carried between rooms without surfacing. Around four months the architecture regularises: the cycles get shorter and more distinct, and at the end of each one there is a shallow patch where she comes close to waking.

You do the same thing several times a night and remember none of it, because you turn over and drop back before you are properly awake. A four-month-old comes up, finds the world is not the one she fell asleep in — the arms are gone, the rocking has stopped, the room is dark in a different way — and does the only thing available to her, which is call you. That is not a habit she has acquired. It is a new chance to notice something, several times a night, and noticing improves with age.

Which is why nobody can honestly promise the nights will go back to how they were at three months. What improves is the join between one cycle and the next — over weeks, unevenly, mostly on her own schedule.

What it looks like

Parents describe the same short list, and no publisher measures any of it, so take this as a description rather than a finding: a baby who was easy to settle now waking every couple of hours; naps collapsing to thirty or forty minutes; false starts at bedtime, where she goes down at seven and is awake and furious by half past; and more milk wanted at night.

Naps usually go first and go worst. A thirty-minute nap is one sleep cycle ended at the moment of surfacing — the same event as the night waking, in daylight, at a lower cost to your evening and a much higher one to your afternoon. If naps are the only thing that has fallen apart, nothing unusual is happening: the drive to sleep is weaker in the middle of the day than it is at midnight, so a new cycle boundary shows up there first.

The other half of it is that awake time is growing anyway. A four-month-old manages somewhere around an hour and a half to a bit over two hours between sleeps, on three or four naps, and a day built for a three-month-old starts producing fights at the cot. The wake windows by age guide explains where those bands come from; the four-month wake windows page and the wake window calculator put them on your clock. From four months there is at least a published total to work from: the sleep consensus the AAP endorses gives 12 to 16 hours per 24 hours including naps for infants aged 4 to 12 months.

What else is happening at four months

Two things, and both matter more than the word regression does. The first is that this is the age from which the AAP suggests putting a baby into the cot drowsy rather than already asleep. It is a suggestion rather than a rule, and plenty of babies have not read it — if she wakes fully and furiously the moment her back touches the mattress, you have learned something about your baby rather than failed at something. Try it on the easiest sleep of the day, and drop it for a week when everyone is exhausted.

The second is movement. Rocking on the stomach, kicking and swimming with the arms usually appear at about five months, and they are the machinery for rolling and crawling; by the end of that stretch most babies roll in both directions, though the AAP is careful to say the time frame varies. Rolling has a direct consequence for the night, and it is the one piece of this page that is not negotiable.

Once a baby shows any sign of rolling, the swaddle is finished — arms out, that night, before she manages it in the dark. A wrapped baby who rolls onto her front cannot use her arms to push herself back up, and the arms are the whole safety mechanism. Everything else stays as it was: babies go down on their back for every sleep, day and night, on a firm flat mattress, in a cot clear of blankets, pillows, bumpers and soft toys, in your room for at least the first six months. Nothing weighted goes on or over a sleeping baby, and a baby who has learned to roll still goes down on her back — where she travels after that, in an empty cot, is hers.

What to do, and what to leave alone

Hold the cheap parts of the day: the same short order of events before every sleep, in the same room, in the same light; nights kept boring, dim and quiet, straight back down; and more settling than you were doing a month ago, because she needs more, not because you have made a rod for your own back.

Then leave the structure alone. A bad fortnight is the worst moment to rebuild the day, drop a nap or start a method, because you cannot tell whether anything you changed worked — the phase moves underneath the experiment. If the fourth nap has become a fight most days for a couple of weeks, that is the four-to-three nap transition, a separate decision with its own signs, usually somewhere around three to five months. Two bad nights is not that signal.

If you want to know which it is, write it down for a week on a printed day sheet you can fill in with one hand. What you are looking for is whether the pattern is drifting in one direction or simply being awful at random.

How long it lasts

Nobody knows, and anybody who prints a number is guessing. Parents commonly report two to six weeks of the worst of it. That is unsourced — no publisher recognises the category, let alone times it — and it is here only to put context around the figure you read elsewhere.

What is worth saying is that the two halves come apart: the maturation is permanent and the misery is not. Most families find the nights become gradually less broken rather than fixed on a particular evening, which is easier to see in a week of notes than in any single night. The six-month page is the next stop, and it carries the best evidence in this subject — at six months, night waking is ordinary rather than a fault, and the largest study of it found two lasting patterns rather than one.

When it is not the phase

A hard fortnight at four months has a folk name, which makes it easy to file everything under it. Teeth, a cold, a hot room and a week away from home all cost sleep too. If she is very hard to settle at all, unusually sleepy, feeding poorly, not gaining weight, or her cry does not sound like her usual one — or if something simply frightens you — that is a call to your pediatrician or health visitor rather than a phase to wait out. This page is general information, not medical advice, and it cannot see your baby.