Baby Sleep Regression: Ages, Signs and What to Do

The four-month change is permanent, and that's okay

Most sleep guides don't say this plainly: the four-month sleep disruption doesn't fully end. The disruption settles. But the sleep pattern that caused it stays forever.

A newborn's brain cycles between just two sleep states, active and quiet, like a very simple machine. Somewhere around three to four months, the brain matures into proper adult-style sleep stages, with light sleep, deep sleep, and REM cycling through the night. Coons and Guilleminault documented this in 31 healthy infants in Pediatrics in 1982, measuring brain-wave activity overnight, and nothing since has changed the basic finding: once that architecture shifts, it does not go back. A 2025 study by Ventura and colleagues in Pediatric Research went further — brain-wave recordings taken during sleep at four months predicted developmental scores at eighteen months. The shift is not just normal. It turns out to be developmentally meaningful.

What this means at 3 a.m.: when your little one wakes briefly between sleep cycles, that brain is looking for whatever was there at falling-asleep time. If that was your arms, the signal goes out. If falling asleep alone is already the routine, resettling mostly happens on its own. This is why how sleep starts begins to matter much more at four months than it did in the newborn weeks. Our sleep schedule by age article covers how total sleep and nap patterns shift across the first year.

Why rough nights track skills, not birthdays

Every sleep regression chart lists the same ages: four months, eight months, twelve months, eighteen months, two years. The ages are useful as rough landmarks. But a 2015 study by Scher and Cohen, published in the Monographs of the Society for Research in Child Development, found something the charts don't show: the disruption tracks the skill, not the date.

Scher and Cohen followed 28 infants from five to eleven months using actigraphy (wrist sensors that record movement overnight), checking in every two to three weeks. When a baby started crawling, night waking went up — not at a fixed age, but at the moment crawling began. And here's the part that surprised the researchers: babies who started crawling earlier than average woke more at night than babies who crawled later. Getting to the milestone faster came with a harder adjustment period.

A 2016 study by Atun-Einy and Scher found the same pattern for pulling to stand. Babies who got there earlier had more disrupted sleep than later achievers. A 2021 time-series analysis by Berger and Moore tracked three families through more than 19,000 daily diary entries and confirmed the same link across multiple milestones. The 2026 review by O'Connor and colleagues in Pediatric Research synthesizes this: crawling infants have more night awakenings than age-matched infants who can't yet crawl, and the onset of sitting, cruising, and walking each brings its own wave of disruption.

Why does this happen? A 2023 study by DeMasi and colleagues in Infancy followed 78 infants with actigraphy through walking onset and found that increased movement during sleep appears to be part of how the brain consolidates the motor memory of a new skill. In the most literal sense, your baby is practicing crawling during sleep — and the disruption is what that process looks like from the outside.

So if your baby is working on something new and the nights get rough, the skill is the reason. When the skill becomes automatic, the nights settle. How long that takes is individual — no peer-reviewed study has validated a specific duration, despite the "two to six weeks" figure that circulates everywhere. What the research does say is that it's temporary, and it resolves without any special intervention.

What eight to ten months adds to the picture

The disruption that clusters around eight to ten months has a second driver on top of the motor milestone piece. Separation anxiety peaks in this window, and it has its own effect on sleep.

A study by Kelmanson, published in Early Child Development and Care in 2012, followed 114 healthy eight-month-olds and found that greater infant separation anxiety was associated with greater bedtime resistance. When a baby has developed enough cognitive understanding to know that you exist somewhere beyond the door (which object permanence makes possible, and which kicks in around this age), that same cognitive leap also makes it possible to actively protest the gap.

So at eight to ten months, two things are often happening at once: new mobility is disrupting sleep from one direction, and a very normal and healthy awareness of your absence is disrupting it from another. Both of these are signs of development moving forward, not backward. If naps are shortening during this same stretch, our baby won't nap article covers the overlapping reasons that tend to cluster at this age.

What eighteen months and two years look like

By eighteen months, walking is well underway, language is exploding, and your toddler has opinions about basically everything, including whether it's time for bed. Sleep disruption at this age tends to come from a mix of ongoing skill consolidation and genuine limit-testing, which is developmentally right on schedule. Separation anxiety can re-emerge in a different form here too, as toddlers develop a clearer sense of what they want and who they want it from.

At two years, another nap transition is often in progress, and verbal resistance at bedtime becomes a new variable. Neither of these is a regression in any meaningful sense. Both are temporary.

Our article on when babies sleep through the night covers Finnish normative data on how often healthy babies wake across the first two years, which is useful for calibrating expectations when the nights feel endless.

How to tell a regression from something else

Most of the time this is pretty straightforward: a baby going through milestone-driven disruption is the same baby during the day — alert, interested in everything, working on the new skill at every available opportunity. Rough nights, fine days.

Illness looks different. Quieter during the day, less interested in eating, sometimes warmer. Teething adds drool and swollen gums and tends to produce daytime fussiness alongside the night waking. A baby who is mid-regression and also teething is going to have a harder time than either alone, which is extremely common in the eight-to-ten-month window when both things often land at once.

Growth spurts are another confound worth naming. An increase in hunger — especially noticeable in breastfed babies — can produce more night waking that looks like a sleep disruption but resolves as soon as feeding amounts adjust. If a baby who was sleeping reasonably well suddenly seems ravenous and waking at unusual intervals, checking whether milk supply or solid intake needs a bump is worth doing before attributing the waking entirely to a developmental phase.

The most useful question is: has something new shown up this week? A new physical skill, a new environment, a new family situation? If the rough nights map onto something your baby is actively working on, the cause is almost certainly developmental. If something else changed, start there.

If disruption is accompanied by changes in weight gain, unusual breathing sounds during sleep, or anything that gives you a nagging feeling, that's worth a call to your pediatrician. It's also worth flagging if the disruption extends well past the period of active skill acquisition with no sign of settling.

The one intervention with actual evidence behind it

A consistent bedtime routine is the only baby sleep intervention with randomized controlled trial evidence behind it. A 2009 trial by Mindell and colleagues in Sleep followed 405 mothers and children aged seven to thirty-six months, randomized to a routine or a control group. The routine group ran a consistent three-step sequence each night (a bath, gentle physical contact, and a quiet activity). Three weeks in, babies in the routine group fell asleep faster, woke less, and their mothers reported measurably better mood.

A predictable sequence signals to a tired baby that the day is ending. It doesn't need to be elaborate — it needs to be the same, in the same order, every night. A bath doesn't have to mean a full tub; a warm washcloth wipe counts. Gentle physical contact can be a brief massage or just being held close for a few minutes. A quiet activity might be a board book, a familiar song, or sitting together in dim light. What matters is the sequence, not the specifics.

For toddlers past eighteen months, giving the routine a visual form helps — when a child can see what comes next, there's less negotiating at each step. Our My First Routine Cards are built around exactly that: a simple picture sequence for the bedtime steps that toddlers can follow and eventually point to themselves.

Worth knowing about new sleep associations during this period: if a baby who was settling independently suddenly needs feeding or rocking during a disruption stretch, and that becomes the new normal, the disruption tends to last longer than the milestone itself. Holding your existing routine steady while offering extra comfort during the day tends to be more sustainable than introducing new props that will need to be present at every wake-up through the night. A baby who learns to fall asleep with a breast or a rocking motion will signal for the same thing at every brief arousal.

Extra daytime connection helps more than most people expect. Babies going through a big developmental stretch often need more physical closeness during the day — more carrying, more floor time with you nearby, more face-to-face interaction. Meeting that need during waking hours tends to reduce the urgency of nighttime signaling, even if it doesn't eliminate it entirely.

There is no evidence that any specific product, white noise setting, or schedule adjustment changes how long milestone-driven disruption lasts. What changes it is the skill becoming automatic.