What causes the 4-month sleep regression?

It's caused by a permanent maturation of sleep architecture. Around 3.5–4 months, the baby's brain transitions from simple 2-stage sleep (active and quiet) to adult-style 4-stage sleep with distinct N1, N2, N3, and REM cycles lasting about 45 minutes each. At every cycle transition, the baby briefly wakes and checks their environment.

How long does the 4-month sleep regression last?

The acute phase typically lasts 2–6 weeks. However, the underlying change is permanent — the baby's sleep architecture has matured and won't revert. What improves is the baby's ability to navigate cycle transitions without fully waking.

Is the 4-month sleep regression a real thing?

Yes, but it's misnamed. It's not a regression (temporary backward step) — it's a permanent developmental upgrade. Sleep researchers call it the 4-month sleep cycle maturation. The baby's brain is now capable of cycling through sleep stages like an adult, which is actually a sign of healthy neurological development.

How do I fix the 4-month sleep regression?

The key principle is consistency of sleep environment: help your baby fall asleep in the same conditions they'll experience when they briefly wake between sleep cycles. If they fall asleep being rocked but wake up in a still crib, they'll fully wake at every 45-minute transition. Put them down drowsy but awake in the environment they'll sleep in.

Can you prevent the 4-month sleep regression?

You can't prevent it — it's a hardwired neurological development. But you can prepare for it by gradually building independent sleep associations before it hits: putting baby down drowsy rather than fully asleep, keeping the sleep environment consistent, and using white noise as a constant cue.

The 4-Month Sleep Regression Explained | Sleep Through The Night

Your baby was sleeping in 3–4 hour stretches. Then, around 3.5–4 months, everything collapsed. Now they’re waking every 45 minutes, and nothing you used to do works anymore.

Welcome to the 4-month sleep regression. Except it’s not a regression. That’s the first thing to understand.

It’s Not a Regression. It’s a Permanent Change.

A regression implies going backward — temporarily returning to an earlier state before recovering. That’s not what’s happening here.

What’s actually happening is a permanent reorganization of your baby’s sleep architecture. Their brain has matured to the point where they sleep more like an adult. That sounds like good news. It temporarily isn’t.

Newborn Sleep vs. 4-Month Sleep: What Changes

Newborns cycle through two sleep states: active sleep (similar to REM) and quiet sleep (similar to deep sleep). Simple, relatively robust to partial arousals, forgiving of imperfect conditions.

Around 3.5–4 months, the brain matures into a four-stage sleep architecture:

N1 — light sleep, the transition into sleep
N2 — true light sleep, where most adult nighttime hours are spent
N3 — deep slow-wave sleep, the most restorative stage
REM — dreaming sleep, memory consolidation

More stages means more transitions. More transitions means more potential arousal points. A newborn might rouse once or twice per sleep cycle. A 4-month-old with adult-like sleep architecture has multiple more opportunities to surface into partial wakefulness within that same window.

And here’s the problem: your baby’s sleep cycle length stayed roughly the same — about 45 minutes. So now they’re completing a full multi-stage cycle every 45 minutes and briefly waking at the end of each one, just as adults do. Adults can roll over and fall back asleep unconsciously. Your baby can’t do that yet — especially if they’ve learned to rely on something external (nursing, rocking, a pacifier being replaced) to fall asleep in the first place.

Two Systems Coming Online at Once

The 4-month transition isn’t just about sleep stages. Two other biological systems are maturing simultaneously:

Melatonin production. The pineal gland’s melatonin output becomes more robust and better-timed around 3–4 months. The evening melatonin surge (the “sleep window”) becomes more defined. Your baby now has a clearer biological bedtime — and if you miss that window, they get overtired, cortisol spikes, and settling becomes significantly harder.

Sleep pressure (adenosine buildup). The homeostatic sleep drive — the pressure to sleep that builds the longer you’re awake — matures around the same time. This means wake windows (the amount of time your baby can comfortably stay awake between naps) become more meaningful. At 4 months, most babies need to sleep every 1.5–2 hours. Letting them get overtired past that window creates a cortisol spike that fights the very sleep pressure you were trying to build.

These two systems need to align. That takes a few weeks of calibration.

Why It Feels So Sudden

Before this transition, your baby could fall asleep anywhere — in the car, in your arms, mid-feed, in the carrier at the grocery store. They had simple sleep needs and a forgiving system.

After this transition, their newly mature sleep architecture makes them far more aware of their environment at each wake-up. If they fell asleep nursing, and they surface into light sleep at the 45-minute mark and nursing isn’t happening — they notice. They cry. You come. They need nursing to get back to sleep. Repeat all night.

This is the core mechanism of the 4-month regression: not that your baby suddenly needs more comfort, but that their newly sophisticated brain now registers the difference between the conditions they fell asleep in and the conditions that currently exist.

What Helps

Consistent sleep environment. Whatever your baby falls asleep in needs to still be present when they surface at the 45-minute mark. If that’s your arms, expect to be needed every 45 minutes all night. The practical path forward is helping them learn to fall asleep in their crib in the same conditions that will exist at 2am — room temperature, white noise, darkness.

Early, consistent bedtime. The melatonin window is real and it’s earlier than most parents expect — typically 7–8pm for a 4-month-old. Putting a baby to bed later in hopes of a later wake-up doesn’t work; it usually results in an overtired baby who sleeps worse. “Sleep begets sleep” isn’t a platitude — the cortisol-adenosine relationship makes it mechanistically true.

Sleep associations that don’t require you. A white noise machine, a consistent room, a swaddle transition to sleep sack — these are associations your baby can rely on without your physical presence at 2am. This is also the window (4–6 months) where evidence-based sleep training approaches begin to be developmentally appropriate, if that’s the route you choose.

What Doesn’t Help

More feeding. At 4 months, most healthy babies are not waking due to hunger at every 45-minute interval. The regressions feels hunger-driven because feeding works to get them back to sleep — but that’s because sucking is soothing, not because they needed calories. Increasing feeds may provide short-term settling but reinforces feed-to-sleep association, making the underlying problem worse.

Co-sleeping as a solution to wake-ups. This is nuanced, so let’s be precise. Bedsharing does not reduce how often a baby wakes — studies by Helen Ball and James McKenna show that bedsharing infants actually arouse more frequently than crib-sleeping infants (arousals are protective and associated with the mother’s sleep cycles). What bedsharing changes is parental awareness of those wake-ups. When the baby is in your bed, you can nurse or resettle without fully waking, and you may not remember doing so. The wake-ups didn’t decrease; your consciousness of them did.

That’s a legitimate choice — reduced parental sleep disruption is real and valuable. But if your goal is to reduce the wake-ups themselves, bedsharing won’t achieve it. And if you start bedsharing during the regression, expect to need to actively transition back later.

How Long Does This Last?

Typically 2–6 weeks of significant disruption while the new sleep architecture settles. After that, a new normal establishes — what that looks like depends heavily on sleep associations and environment.

Babies who had independent sleep skills before the regression tend to find their footing faster. Babies who relied heavily on feeding or parental contact to fall asleep tend to have a harder transition, because the thing they need to fall back asleep is now needed every 45 minutes.

This is also why the 4-month mark is commonly cited as the ideal window to begin thinking about independent sleep skills — the neurodevelopment that makes it possible is just arriving, and you’re already dealing with disruption anyway.

The Upside

Their brain is more capable than it was a month ago. The sleep that emerges on the other side of this transition is more organized, more predictable, and builds toward the consolidated nights you’re waiting for. The architecture change that’s breaking your sleep right now is the same architecture that will, eventually, give you a baby who sleeps through the night.

You’re watching their brain grow. It’s loud at 3am, but it’s real.