It isn't really a "regression"
The name is a little misleading. Nothing is going backward, and there's no old version of sleep to return to once it passes. What's actually happening around 4 months is a permanent maturation of sleep architecture, the shift from the simple, newborn-style sleep a baby is born with into the multi-stage, adult-like sleep cycles they'll have for the rest of their life.
Newborn sleep is fairly basic: deep sleep and light sleep, with relatively few transitions in between. Around 4 months, that structure reorganizes into distinct sleep stages, including the lighter stages that adults also pass through several times a night. Once this shift happens, it doesn't reverse. It's the new normal, which is exactly why babies don't simply "snap back" to how they slept before.
What changes exactly
A few concrete things shift during this stage. Sleep cycles shorten to roughly 45 minutes, each one ending in a brief wake or a very light sleep transition before the next cycle begins. A newborn could often sleep right through that transition point without anyone noticing. A 4 month old, with a more mature and more easily disrupted sleep structure, often can't link those cycles together without some help, at least not yet.
The practical result is more night wakings, since each 45 minute cycle boundary is now a potential wake-up point instead of an unnoticed transition. Naps are affected too. A nap that used to run a full sleep cycle or more can shrink down to just one cycle, often landing somewhere around 30 to 45 minutes before a baby surfaces and struggles to continue.
The biology behind this is worth understanding because it explains why the old tricks stop working overnight. Newborn sleep is built almost entirely from two states: active sleep (the newborn version of REM, where twitching, grunting, and irregular breathing are normal) and quiet sleep (deep, still, and harder to disturb). A newborn drops into quiet sleep fast and stays there for a big chunk of each stretch, which is part of why a 6-week-old can sometimes be moved from arms to bassinet without stirring. Around 2 to 4 months, the brain begins laying down the adult-style architecture: distinct light sleep, deeper slow-wave sleep, and REM sleep, cycling through stages roughly every 45 minutes to an hour instead of drifting through one or two long blocks. Each lap through that cycle includes a brief rise toward the surface of consciousness, close enough to waking that any disruption, a noise, a missing pacifier, the absence of the motion or contact they fell asleep with, can tip a baby the rest of the way into a full wake-up.
This is also the point where melatonin production starts to come online in a more organized way and the circadian rhythm begins to take real shape. Before around 3 months, babies don't yet produce melatonin on a day-night cycle at all, which is a big part of why early newborn sleep is scattered evenly across 24 hours with no real concentration at night. As the suprachiasmatic nucleus (the brain's circadian clock) matures and starts responding consistently to light and dark, melatonin output begins to rise in the evening and fall in the morning, nudging sleep to consolidate into a longer nighttime block. That's genuinely good news buried inside a rough few weeks: the same maturation that's causing the choppy nights is also the maturation that eventually produces real day-night sleep.
Short, choppy naps and a sudden wave of night wakings can be hard to track in your head when you're exhausted. Dreamer logs every nap and wake automatically as you go, so you can see in the weekly insights whether the pattern is actually easing, even on nights it doesn't feel like it.
How it's different from a growth spurt or illness
Three things can make a previously good sleeper suddenly fall apart around 3 to 5 months, and they're easy to mix up because they can all show up as more night waking and shorter naps. Telling them apart matters, because the right response is different for each one.
A growth spurt is mainly about appetite, not sleep architecture. It tends to last just a few days, not weeks, and the dominant sign is a baby who suddenly wants to feed far more often, day and night, alongside the usual disrupted sleep. Once the spurt passes, typically within 2 to 5 days, feeding and sleep both settle back to where they were without any lasting change to how sleep itself is structured.
An illness comes with company. Congestion, a cough, a fever, diarrhea, a rash, or simply a baby who seems genuinely unwell rather than just fussy are the giveaways. Illness-related sleep disruption tends to track the illness itself: it gets worse as symptoms peak and improves as your baby recovers, usually within a week or two, rather than persisting on its own separate timeline.
The 4-month regression looks different from both. It isn't appetite-driven the way a growth spurt is, and there's no fever, congestion, or other illness marker attached to it. Instead, it's a multi-week pattern shift in the mechanics of sleep itself: shorter cycles, more wake-ups exactly at cycle boundaries, and naps that compress to a single 30 to 45 minute cycle, all starting fairly suddenly and persisting for 2 to 6 weeks rather than resolving in a few days. If your baby seems otherwise healthy and is feeding normally but sleep has changed shape rather than just gotten harder for a few nights, the regression is the more likely explanation.
Signs you're in it
A few patterns tend to show up together, usually somewhere between 3.5 and 5 months. If multiple things below are happening at once, the 4-month sleep regression is a reasonable explanation.
Night sleep changes
The clearest sign is usually at night: a sudden jump in wake-ups in a baby who was previously sleeping in longer stretches. A baby who was down to one wake-up, or even sleeping through, can start surfacing every 45 minutes to an hour, which lines up almost exactly with the new, shorter sleep cycle length. Unlike a gradual change in wake windows as a baby grows, this shift can show up within a matter of days rather than easing in slowly, which is part of what makes it feel so alarming in the moment.
Nap changes
Naps that used to run 1.5 to 2 hours can drop down to short 30 to 45 minute catnaps almost overnight. This is the single-cycle nap in action: your baby falls asleep, completes one full 45-minute cycle, and then surfaces without the skill yet to link into a second cycle on their own. Some babies will take that short nap and simply be done; others fuss and resist being resettled, which can make a single nap stretch into a longer, more effortful ordeal even though the actual sleep time barely changed.
Bedtime changes
New resistance at bedtime is the third common piece, and it's often the most confusing one because it doesn't fit the "tired baby falls asleep easily" assumption. A baby who is plainly exhausted, rubbing their eyes and yawning, can still fight going down, arching away from the chest or bottle, or popping back awake the moment they're placed in the crib. This isn't defiance. It's a body whose sleep-onset mechanics are mid-reorganization, paired in many cases with a baby who is now alert enough at the edge of sleep to notice, and protest, that something about how they're falling asleep has changed.
See the pattern, not just the bad nights
Dreamer's weekly insights show whether naps and night stretches are actually trending longer, so you're not relying on memory alone during the hardest weeks.
What actually helps
A few adjustments tend to make the biggest difference during this stretch.
- Keep wake windows age-appropriate and consistent. An overtired baby has an even harder time linking sleep cycles together, so consistent timing matters more here than usual. See our wake windows by age guide for current ranges, or our sample 4-month-old sleep schedule for a full day mapped out.
- Practice some self-settling at sleep onset. Rather than feeding or rocking all the way to fully asleep every single time, try putting your baby down drowsy but still slightly awake when you can. This gives them practice with the exact skill, linking sleep cycles independently, that the regression is testing.
- Move bedtime earlier if short naps are causing overtiredness. When daytime sleep drops because of choppy naps, an earlier bedtime helps make up the difference without forcing extra naps that don't want to happen.
- Give it time. This stage typically eases over 2 to 6 weeks as babies adapt to their new sleep architecture, even without any major changes on your end.
What doesn't help
Waiting it out with zero adjustment can actually prolong the rough patch. Since the missing piece is usually the skill of linking sleep cycles independently, not simply time passing, a baby who's never given the chance to practice that skill may keep waking fully at every cycle boundary well past the point most babies start smoothing out.
Overcorrecting in the other direction doesn't help either. Pushing bedtime unusually late in an attempt to "tire them out more" tends to backfire, since an overtired baby's body produces stress hormones that make settling harder, not easier. The better fix during this stage is almost always an earlier bedtime, not a later one.
It's also worth being patient with yourself, not just the schedule. This stage tends to land at the same time as a return to work for many parents, a string of short nights, and a baby who suddenly seems harder to read. None of that means anything is being done wrong. The combination of a developmental shift and ordinary exhaustion is simply a hard few weeks, and it passes faster with small, consistent adjustments than with a complete overhaul of how your family does bedtime.
Key takeaways
The 4-month sleep regression is a permanent, one-way shift from simple newborn sleep into adult-style sleep cycles, not a phase that reverses on its own. Sleep cycles shorten to roughly 45 minutes, each boundary becomes a potential wake-up, and naps often compress to a single cycle until your baby learns to link cycles independently. It typically starts between 3.5 and 5 months, shows up suddenly rather than gradually, and resolves over 2 to 6 weeks. Unlike a growth spurt (appetite-driven, lasts days) or an illness (comes with fever, congestion, or other symptoms), the regression is a multi-week change in sleep mechanics with no other red flags attached. Age-appropriate wake windows, some practice with independent settling, and an earlier bedtime on short-nap days are the levers that actually help; a later bedtime to "tire them out more" tends to backfire because it adds overtiredness on top of an already disrupted system.
Reviewed for accuracy. This guide reflects general pediatric sleep guidance and is reviewed by Dreamer's certified pediatric sleep consultants (CPSCs). It's informational and doesn't replace advice from your child's pediatrician.
Frequently asked questions
How long does the 4-month regression last?
Commonly 2 to 6 weeks, sometimes longer if sleep habits don't adjust alongside it.
Does every baby go through it?
The underlying neurological shift in sleep cycles happens to everyone, but how disruptive it looks varies a lot. Some babies barely show signs of it.
Can I prevent it?
Not entirely, since it's a normal developmental shift. Some practice with independent settling beforehand can soften the disruption, though.
Is sleep training necessary at 4 months?
No, full sleep training isn't required at this stage. If you want a structured, low-pressure approach later on, see our gentle sleep training guide.
Can the 4-month regression happen earlier or later than 4 months?
Yes. The underlying sleep-cycle maturation typically lands somewhere between 3.5 and 5 months, and "4-month" is shorthand for that window rather than an exact birthday. A baby who hits it at 3.5 months or closer to 5 months is still well within the normal range.
Should I feed back to sleep during this stage?
It's fine to feed a baby who's genuinely hungry, but feeding at every single wake-up during this stage can make it harder to tell hunger from a normal cycle-boundary wake, and can build a strong association between feeding and falling back asleep. Many families find it helpful to offer a feed for clearly hungry wake-ups while giving a few minutes of settling time for ones that look more like a cycle transition.
Sources
- American Academy of Pediatrics — infant sleep cycles and developmental sleep changes: healthychildren.org
- Sleep Foundation — infant sleep architecture and circadian rhythm development: sleepfoundation.org
- American Academy of Sleep Medicine — pediatric sleep consensus statements: aasm.org
- Eunice Kennedy Shriver National Institute of Child Health and Human Development — infant sleep and development research: nichd.nih.gov