Night wakings and sleep regression in children: understanding them and getting through

A waking that isn't a failure

Waking up several times a night is part of normal sleep, at any age. A child's sleep cycle is shorter than an adult's, with more frequent passes through a light-sleep phase where a noise, a change of position, or a sensation is enough to wake them. What changes with age isn't how often they wake. It's their ability to fall back asleep on their own, without calling out.

A child who seemed to never wake up wasn't waking any less than other children: they were falling back asleep before anyone noticed. A regression doesn't create a problem that wasn't there before. It just makes visible a waking that had gone unnoticed until then.

"Sleeping through the night" also arrives at very different ages. In a study that followed 75 infants for a year, sleep consolidated mostly during the first four months, yet at twelve months 11 of those children had reached neither the eight-hour threshold nor the midnight-to-5 a.m. window, two of the three definitions of "sleeping through the night" the authors tested. A review of 34 studies also puts the number of night wakings, depending on the study, between 0 and 2.5 per night in one- to two-year-olds.

This return of night wakings also often gets read as a personal failure: a routine not kept up properly, a mistake made while teaching independent sleep, a step backward in parenting. A regression's timing almost always follows the child's development, not choices made at bedtime in the preceding weeks. A child learning to walk wakes up more than one who isn't moving around yet, no matter how good the evening ritual is.

Why does a child who slept well start waking again?

Most of the time, it's the child's development moving forward. Regressions cluster around fairly predictable windows: major developmental milestones (rolling over, walking, talking), the surge in separation anxiety around 8-10 months and again around 18-24 months, a teething spell, or a change in environment: moving house, starting a new daycare, the arrival of a younger sibling.

The mechanism is almost always the same: a brain busy consolidating a new skill during the day stays more active at night, which lightens sleep and makes full waking easier at every cycle transition. It isn't a disorder. It's a nervous system processing more information than usual.

These windows return with striking regularity from one child to the next. They track developmental stages rather than habits picked up at home. A child who has always slept well and a child who has always had restless sleep generally go through the same windows, at around the same age. The first one doesn't get a pass just because they're a "good sleeper."

How can you tell a regression from a real sleep problem?

A typical regression lasts two to six weeks and fades on its own, without needing any change to the sleep setup. It usually lines up with an identifiable developmental window: around 4 months, 8-10 months, 18 months, or 2 years old are the ones that come up most often in pediatric guidance.

What warrants a medical opinion is something else: wakings that persist past six to eight weeks with no improvement, marked daytime distress, noisy breathing or pauses during sleep, or a clear change in appetite or behavior. Most regressions only need time; some deserve a closer look.

In practical terms: a nine-month-old who had slept without waking since four months and starts waking two to three times a night for three weeks, with no fever or other notable change, is almost always going through an ordinary regression. The benchmark to watch is duration rather than the number of wakings per night: past six to eight weeks with no improvement, or alongside the signs already mentioned, it's better to see a doctor than to wait any longer.

Regression, parasomnia, habit or insomnia: which guide to read

A night waking isn't always a regression, and three other situations look like one. An episode that consistently happens in the first two to three hours of the night, rather than at a time that varies, points to a night terror or an episode of sleepwalking, two parasomnias with causes and management different from what's described here.

A waking that lasts well beyond the developmental window and only settles with a feed, a bottle, or one particular gesture looks more like a sleep-onset habit. Our guide on night wakings and night weaning explains how to tell hunger from habit, and how to undo the gesture gradually.

Finally, trouble falling asleep or wakings that last for months, with an effect on daytime, fit better under our guide on insomnia in children, which also says when to bring it up with a doctor.

What helps you get through a regression

Keeping the evening ritual exactly the same is the most useful lever: it's the one stable landmark the child has while everything else is shifting. Avoiding the temptation, out of sheer tiredness, to introduce a new habit you won't want to keep afterward (rocking all the way to sleep, sleeping in the parents' bed every night) matters just as much: a child learns it just as fast as a typical regression lasts.

Responding consistently, with brief comfort rather than a drawn-out interaction, and keeping the same bedtime despite accumulated tiredness, helps the child get back their ability to fall back asleep on their own more quickly.

Some regressions call for a more specific adjustment. A child learning to pull themselves up often tests this new skill at night, in their crib: laying them back down once usually isn't enough, they'll pull up again to try it once more. Giving them time during the day to practice standing up and sitting back down reduces the need to test it out at night; they end up lying back down on their own once they've mastered the move.

A teething spell follows a shorter timeline, usually a few days to a week, with pain that comes and goes. The real risk at that point isn't the pain itself but the habit picked up to calm it: rocking all the way to sleep, letting a bottle become a required step to get back to sleep. Treating the pain, with the pharmacist's or pediatrician's advice on a suitable pain reliever, rather than compensating with a new bedtime habit, keeps it from outlasting the teething itself.

A change in environment (moving house, starting a new daycare, the arrival of a younger sibling) often disrupts sleep more, because several familiar landmarks disappear at once. Recreating those landmarks despite the change of scenery, with the same comfort object, the same night-light, the same song, helps the child find their bearings again faster, even if the room itself has changed.

Around 18-24 months, the language boost and separation anxiety often overlap: the child understands better that a parent who leaves the room keeps existing elsewhere, which makes the separation more conscious without making it any easier. Short phrases repeated at every departure, like "I'll be back" or "see you in the morning," generally reassure more than staying longer, which can lead the child to expect the same every night after.

Background noise is sometimes added at this point to cushion whatever's interrupting a sleep that's already lighter than usual. How effective it really is remains debated, and our guide on white noise covers what the studies show on both sides.

Despite these different triggers, the approach stays the same at its core: restore what the child already recognizes as quickly as possible, rather than adding a new gesture or object to help them through. That's often the hardest part to hold to, especially in the middle of the night, when any solution that calms things down right away looks like the one worth trying.

When it resolves

Coming out of a regression is gradual, not a clean switch from one night to the next: wakings first space out, then get shorter, before becoming as unnoticeable as before. Consistency in the response, day and night alike, remains what most reliably shortens this period, more than any one-off method tried for a single night.

It isn't unusual for one regression to be followed closely by another: a child might go through a walking-related phase around age one, then a language-related one a few months later, with neither preparing for nor worsening the next. Each episode keeps its own timeline; having already been through one doesn't shorten or lengthen the one that follows.