Few parenting phenomena generate as much desperation as the four-month sleep regression. A baby who was managing reasonable stretches of sleep suddenly begins waking every hour or two through the night, seemingly for no reason. Parents who thought they were through the worst of it find themselves back to near-newborn levels of sleep deprivation. The reasonable question is: what happened?
The answer is that the baby's sleep architecture changed. This is not a temporary regression that will pass and leave things as they were. It is a permanent developmental shift. Understanding this is both disheartening in one way (things will not simply improve on their own as they did with earlier sleep disruptions) and useful in another (there are things that actually help).
Healthbooq (healthbooq.com) covers infant sleep through the early months, including the common disruptions and what evidence-based approaches look like.
What Changes at Four Months
Newborn sleep is simpler than adult sleep. Babies cycle primarily between active sleep (roughly equivalent to REM) and quiet sleep (non-REM), with relatively brief transitions between them. These transitions are easier to sleep through.
Around three to four months, the sleep cycles mature and begin to include the lighter stages of non-REM sleep that adult sleep contains. The cycle length, previously around 50 minutes, shifts. And between each cycle, the brain has a partial arousal: a brief rising to the surface of consciousness before dipping back down.
This partial arousal is not the problem. Adults have these all night long and typically do not wake fully or remember them. The problem is what happens at the arousal. If the adult falls back to sleep without noticing the transition, they are demonstrating sleep-onset independence in the middle of the night. They can re-settle themselves because the conditions at the start of sleep are still present.
A baby who was fed, rocked, or held to sleep initially falls asleep under those conditions. When they surface at a partial arousal and the conditions are different (they are in a cot, no longer being held, not feeding), the mismatch is alarming. They fully wake and signal for the same conditions that were present when sleep began.
This is what sleep-onset association means in practice. It is why babies who fall asleep independently at bedtime are more likely to re-settle between cycles without fully waking.
Why Some Babies Are More Affected Than Others
A baby who had reasonable stretches of sleep before three to four months but was always being fed or rocked to sleep will often find the regression dramatic. The previously brief, simple cycle transitions are now longer and involve more light sleep, and each one requires re-establishment of the sleep conditions.
A baby who had already been settling to sleep somewhat independently, even partially, may have a smoother transition, because they have more practice at re-settling without adult input.
Neither situation reflects anything wrong with the parent or the baby. The regression affects the vast majority of babies to some degree; it is just that babies with established sleep-onset associations tend to show it more acutely.
What Helps
The regression is not reversible, but it is manageable. The most durable approach is working on sleep-onset associations: supporting the baby to fall asleep with less parental input so that they are better equipped to re-settle between cycles.
This does not require sleep training of any particular kind, and families vary enormously in their approach, their circumstances (bed-sharing, breastfeeding, maternal working schedule), and their priorities. But even gradual, gentle changes to the falling-asleep process move in a helpful direction.
Practical steps include putting the baby down slightly more awake than usual at sleep onset, so they have the experience of drifting off without being fully asleep when they leave contact. Or responding at night with progressively less stimulation: a voice before picking up, a hand on the chest before full re-settling, as a baby who is capable of more will sometimes re-settle with less.
Age-appropriate wake windows matter. A baby who is overtired by bedtime has a harder time achieving settled sleep and is more reactive to partial arousals. At four months, wake windows of roughly 90 minutes to two hours between sleep periods help prevent overtiredness while allowing enough sleep pressure to build for a proper sleep.
Feeds at night remain genuinely needed at this age: the expectation should not be that a four-month-old will sleep through. The target is not necessarily a full night but more consolidated sleep with fewer wakings.
How Long It Lasts
For babies who are managing sleep-onset independently, the disruption often settles within a few weeks as the brain consolidates the new sleep architecture.
For babies with strong sleep-onset associations, the frequent waking tends to persist until something changes about those associations. This is why parents who wait for the regression to pass sometimes find that it does not.
Key Takeaways
The four-month sleep regression is a permanent change in sleep architecture rather than a temporary disruption: around three to four months of age, infant sleep cycles become more similar to adult sleep, with lighter stages of sleep and more frequent partial arousals between cycles. A baby who previously slept through cycle transitions now wakes at each one. Because this is a developmental change rather than a reversible blip, there is no going back to the pre-regression sleep pattern. A baby who had learned to fall asleep independently will link sleep cycles more effectively than one who relied on feeding, rocking, or holding to fall asleep. Settling strategies that build independent sleep onset can help.
