When Will Infant Sleep Through the Night? What Has to Happen First

Five hours, not twelve

The finish line in the research is much closer than the finish line in your head, and almost every disagreement about this question dissolves once you know where it has been drawn.

The five-hour criterion goes back to work published in the 1950s, and it persists because it is easy to measure and because five uninterrupted hours genuinely is a physiological milestone for a young infant. Later researchers have used six hours, eight hours, and fixed blocks such as ten in the evening to six in the morning. Henderson and colleagues, publishing in Pediatrics in 2010, applied several of these criteria to the same infants and produced findings that differ substantially depending on which one you pick.

The practical consequence is that any sentence beginning most babies by such-and-such an age is uninterpretable until you know the criterion behind it, and the criterion is almost never stated in the places parents actually read. When you see a confident age, that is the first question to ask.

It also means something kinder. If your five-month-old regularly gives you a stretch from seven until one, she is, by the standard used across a large share of the literature, already doing the thing. You are simply awake for the other half of the night, which is a genuine and exhausting problem — just not the one you thought you had.

Which infant are we talking about?

The word infant does a lot of quiet work in this question, and the answer at six weeks has almost nothing in common with the answer at nine months.

Under about three months, consolidated night sleep is not really on the table, and not because the baby has failed to learn anything. The machinery is not built yet. Between three and four months it starts changing, which is the origin of what everyone calls the four-month regression and is better described as a maturation. From roughly four months the mechanics permit a long stretch for the first time, though permitting is not producing. Only from six months does when will infant sleep through the night become a partly behavioural question — which is where sleep training enters, covered neutrally, including the option of not doing it, on how to sleep train a baby.

The table sets the four spans against what is actually published for each. Note who publishes which row: the American Academy of Sleep Medicine's consensus statement explicitly declines to make a recommendation below four months, citing insufficient evidence and the enormous normal variation at that age, so the youngest row belongs to the National Sleep Foundation instead. Any page quoting an AASM figure for a newborn has misattributed it.

The body clock is not installed at birth

This is the first of the three systems, and the one that explains most of what happens before three months.

Newborns do not arrive with a functioning circadian rhythm. The internal day-night cycle develops over the early months rather than switching on, and it is shaped by environmental cues — chiefly light, but also feeding times, temperature and the pattern of activity around the baby. A 2022 review in the Journal of Physiological Anthropology on the development of the circadian system in early life sets out both the maturational timeline and the environmental factors that influence it.

What that means at two in the morning: a six-week-old waking at three is not confused, badly trained, or in a bad habit. She has no internal signal telling her that three in the morning differs from three in the afternoon. Sleep spreads roughly evenly around the clock because nothing is yet in place to concentrate it.

This is why the first item on the influence list below is light rather than a technique. Light is the strongest cue the developing system uses, and one of the very few places where ordinary parental behaviour feeds into the biology. It will not accelerate the timeline. It gives that timeline better information to work from — a smaller claim, and a true one.

Short cycles, and what happens at the end of each one

The second system is sleep architecture, and it is the piece that surprises people most.

Infant sleep cycles are considerably shorter than adult ones — commonly published at somewhere around fifty to sixty minutes in the first months, against roughly ninety in adults — and a larger share of that time is spent in the lighter, more active states. Every cycle, in every human being, ends in a brief arousal. Adults have these too and almost never remember them, because we roll over and rejoin the next cycle without ever surfacing properly.

So the target is never the elimination of waking, which is neither possible nor desirable. The target, when there is one, is whether she returns to sleep after an arousal without an adult reinstating whatever she fell asleep with. That capacity develops, and in plenty of infants it develops without anyone doing anything about it.

Galland and colleagues' systematic review of normal sleep patterns in infants and children, published in Sleep Medicine Reviews in 2012, is the standard reference for what typical looks like at each age, and it matters precisely because it reports wide normal ranges rather than a single expected value. The spread between two entirely typical babies of the same age is far larger than parenting content admits.

The feed gap is a medical question wearing a schedule costume

The third system is the one people most often try to solve with logistics, and it is the one that genuinely is not yours to decide alone.

How long an infant can comfortably go between overnight feeds depends on age, weight, growth trajectory, feeding method, whether she was born early, and her medical history. That is a clinical judgement about a specific child. This site does not make it, no app makes it, and no commercial program should either — one that tells you to drop night feeds without referring you to a clinician has removed the only part of the decision with medical content in it.

There is a version of this with a straightforward answer: ask at the next appointment, and ask specifically. Not is she sleeping through yet, which invites a reassuring non-answer, but does she still need calories overnight at this weight and on this growth curve. Bring the log — a week of written clock times is worth more to a pediatrician than anything you can reconstruct from memory.

Until you have that answer, night feeds are not an obstacle to be engineered around. They are a requirement, and building a plan that quietly assumes otherwise is where a lot of well-intentioned advice goes wrong.

What you can actually influence

Having spent four sections on things you cannot change, here is the honest short list of things you can. It is shorter than the internet suggests and every item on it is modest. None of them accelerates circadian development or sleep-cycle maturation, because nothing does. What they do is remove obstacles and supply the developing system with clearer information.

The first three are about contrast and predictability: making day unmistakably different from night, and making the run-up to sleep identical enough to work as a signal. The fourth is about not letting the day's unfinished business arrive at two in the morning. The fifth is the arithmetic point running through this whole site — a bedtime derived from the day beats one assigned to a clock hour, because a baby put down before she is tired will not sleep however well you did everything else.

The safe-sleep item is in a different category from the rest and belongs on every list of this kind. It is not there to improve sleep quality. It is there because it is the thing that matters most and because it is not something to take second-hand from a site like this one — read it at the American Academy of Pediatrics and NICHD, both linked at the foot of this page.

• Light: bright exposure in the morning and through the day, ideally some of it outdoors; dim and uninteresting after bedtime
• A short, identical, predictable pre-sleep sequence, run in the same order every night even when it is going badly
• A sleep space that meets AAP safe-sleep guidance — this one is non-negotiable and is not about sleep quality at all
• Daytime feeding unhurried enough that she is not making up a shortfall at two in the morning
• A bedtime derived from the day's wake time and daytime sleep rather than assigned to a clock hour
• A pause of a minute or so before responding to a noise, once you are comfortable with it — some noises end on their own

The routine evidence, stated at its actual strength

Bedtime routines are the most-recommended intervention in infant sleep and one of the few with a real evidence base, so it is worth being precise about what that base does and does not show.

The strongest study design in this area is the randomised trial. The INSIGHT trial, reported by Paul and colleagues in Pediatrics in 2016, randomised first-time mothers and their newborns to a responsive-parenting intervention with a sleep component, against a control group. Infants in the intervention arm slept longer overnight by 35 minutes at 8 weeks, 25 minutes at 16 weeks and 22 minutes at 40 weeks. That is a genuine finding from a genuinely controlled design, which is rarer in this field than the confident tone of most parenting content would suggest.

Here is the boundary on it, and it matters. Those are differences of twenty to thirty-five minutes in group averages — real, worth having, and not the same thing as a transformed night. The same paper reports that by one year the two groups' sleep duration was similar, which is the sort of detail that rarely survives the journey into a listicle. These trials do not demonstrate that a bedtime routine produces a twelve-hour night, and a group average is not a promise delivered to an individual household.

So run the routine. It is cheap, pleasant, evidenced, and it gives the day a shape for both of you. Just hold the expectation at the size the evidence supports: a routine that has not produced a consolidated night by week three has not failed. It was never the mechanism that produces one.

Making day look like day

The light and contrast lever is the one worth doing properly, because it is nearly free and because most people do half of it. Almost everybody darkens the nights. Far fewer deliberately brighten the days, and the contrast is the active ingredient rather than the darkness on its own.

In practice this means resisting the instinct to turn daytime naps into a nocturnal ceremony. Blackout blinds, silence and tiptoeing for every nap teach a developing system that dark and quiet are simply what sleep looks like, at any hour. Ordinary household light and noise during the day preserve the difference you are building. Older babies whose naps are genuinely disrupted by light are a separate case, and by then the circadian point has largely been made.

The night side is about being boring rather than about being dark. Nothing interesting happens after bedtime: no conversation, no eye contact games, the least light you can manage, and a nappy change only when it is genuinely required rather than as routine.

The last item on the list below is the one that does the most work and is hardest to keep. A fixed morning wake time is the anchor that every other timing on the day hangs from, and it matters most on precisely the mornings you least want to enforce it.

• Morning: open the curtains at roughly the same time daily, and get outdoors into daylight when you can
• Daytime naps: ordinary household light and ordinary household noise, not a blackout ritual
• Feeds before bedtime: lights on, talking, eye contact, the full social event
• Feeds after bedtime: the dimmest light you can manage, no conversation, no nappy change unless you must
• Overnight: nothing interesting happens — no play, no lights, and leave once she is settled
• The same morning wake time every day, including after the worst nights, because it anchors everything else

What it looks like when it does arrive

It is not a switch, and expecting a switch is part of why the arrival so often feels like an anticlimax or a false alarm.

The first long stretch almost always sits at the front of the night, because the deepest sleep of a human night is concentrated in its first hours. So a baby who begins consolidating typically gives you an evening block first, then reverts to shorter cycles in the early hours. If you go to bed at eleven, you may not notice the improvement at all, which is its own small tragedy and an argument for going to bed earlier for a fortnight.

It is also unstable when new. A stretch appears, holds for four nights, vanishes for a fortnight, and returns. Weinraub and colleagues, publishing in Developmental Psychology in 2012, tracked night waking from six months through three years and described distinct developmental trajectories rather than one shared staircase, with waking persisting well past the first year for a meaningful group of children.

The table sets the common expectation against what tends to show up instead. None of the right-hand column is a failure state. It is what the process looks like from inside a house where nobody has slept properly for some time.

Things that will not speed it up

A short list of the interventions parents are most often pushed toward, none of which brings the timeline forward, and some of which make the days meaningfully worse.

The cereal-in-the-bottle suggestion is the most persistent, usually offered by someone a generation older and offered kindly. It is a question for your pediatrician rather than a sleep tactic, and current guidance on when solids are introduced does not support using them for this purpose. Ask rather than assume, and ask before trying it.

The keeping-her-awake-longer family of suggestions fails for a reason worth understanding: an overtired baby settles worse, not better, and the effect compounds across days. The same logic applies to pushing bedtime later. Early waking in particular tends to respond to an earlier bedtime rather than a later one, which is thoroughly counter-intuitive and holds up in practice often enough to try before anything else.

The last item is the one that matters most. Anything sold on the promise of longer sleep needs checking against AAP safe-sleep guidance before it goes anywhere near the crib, and that check comes before any consideration of whether it works. This site does not summarise that guidance, because it is not something to take second-hand — the links are at the foot of the page and they are worth reading in full.

• Cereal in a bottle — a question for your pediatrician, not a sleep tactic, and current guidance does not support it for this
• Keeping her awake longer in the day, which reliably produces a harder bedtime rather than a longer night
• Pushing bedtime later; early waking more often responds to an earlier bedtime, counter-intuitive as that sounds
• Dropping a nap before she is ready, which borrows from the night to pay the day
• Waking her deliberately to reset the night, outside a specific plan agreed with a clinician
• Any product marketed as extending sleep, until you have checked it against AAP safe-sleep guidance

If you do nothing at all

Most of what this page describes happens without intervention. That is not a flourish at the end of an article — it is the developmental picture the whole page has laid out. Circadian rhythm establishes itself. Sleep cycles mature. Feed gaps lengthen as a baby grows. None of it requires a method, a program, or a decision from you, and no window closes if you leave it alone.

The evidence does not stand against that choice either. Pennestri and colleagues, in Pediatrics in 2018, asked whether uninterrupted infant sleep was associated with better infant development and did not find that association on the measures they used. What they did find was an association with maternal mood — which reframes the question. If the waking is genuinely fine for your baby and genuinely not fine for you, then you are the person with the problem, and that is a legitimate problem to solve rather than a selfish one to feel bad about.

Solving it need not mean sleep training. It can mean splitting the night so each adult gets one unbroken stretch, going to bed at nine for a fortnight, accepting help you have been declining, or telling your doctor how you are actually doing rather than the version you give at the school gate.

If you want the schedule side — wake windows, the derived bedtime, the age-by-age charts — the homepage collects it, and how to get a baby to sleep through the night covers the practical levers at length. But nothing here is a requirement. The honest answer to when will infant sleep through the night is: later than you were told, earlier than it feels like tonight, and largely on her own schedule.