Sleep Regressions, Naps, and Wake Windows: What's Actually Happening
That sudden wave of night waking around 4 months isn't a setback — it's your baby's brain growing up. Here's the science behind sleep regressions, and what realistic naps and wake windows look like through the first year.

If your once-sleepy baby suddenly seems to wake up to the very concept of nighttime somewhere around 3 to 4 months, you haven't broken anything — and you're not imagining the shift. What looks like a step backward is actually your baby's sleep system growing up: cycling through real sleep stages for the first time, the same way it will keep doing for the rest of their life. There's a fairly predictable shape to how this unfolds across the first year, plus one more bumpy stretch around 8 to 10 months. Here's what's happening under the hood, and what a realistic night can look like at each stage.
What's actually changing at 3 to 4 months
Newborns don't really have sleep stages in the adult sense — their sleep is fairly uniform and cycles quickly, roughly every 45 minutes. Research on infant sleep and development describes a shift starting around 3 to 4 months: the brain begins producing adult-like sleep architecture, cycling through the same lighter and deeper stages, including REM, that adults do, with each cycle stretching out toward 90 minutes. A related systematic review found that this circadian and sleep-stage maturation tends to line up with other developmental progress — it's not an isolated sleep event, it's part of the same 3-to-4-month wave of growth that shows up in smiling, tracking, and cooing too.
The practical effect: your baby now surfaces briefly between sleep cycles the way you do, and has to either fall back asleep on their own or signal for help. That's the 4-month sleep regression — not lost ground, but a permanent upgrade that comes with a bumpy adjustment period, typically 2 to 4 weeks, of shorter naps (around 45 minutes is common) and more frequent night waking while your baby learns to link cycles together.
What the research says
A few studies help make sense of what's normal here and what actually helps.
A study using video monitoring of infants in the first three months found that babies who were able to self-settle back to sleep during night wakings, without parental intervention, were more likely to develop long, consolidated stretches of nighttime sleep by 3 months. It's an observational finding rather than a test of any particular method, but it's often cited in support of predictable bedtime routines that give a baby the chance to practice settling.
Longer-term, a longitudinal look at infant sleep behavior found that by 12 months, roughly 70 to 80% of babies are sleeping predominantly at night, with daytime naps progressively shrinking in number across the year, which matches the general nap-count pattern below.
There's also reassuring evidence about room-sharing, which the AAP and NICHD both continue to recommend for the first 6 to 12 months as a SIDS-reduction strategy. A 2023 AAP paper acknowledges the real tension between this guidance and parents' own sleep quality, while noting the safety benefit is substantial. A separate longitudinal study following families well past infancy found that room-sharing without bed-sharing in the early months was associated with fewer behavioral problems in middle childhood — one more reason the separate-surface, same-room setup earns its place in safe-sleep guidance beyond the newborn stage.
Wake windows and naps, roughly by age
These are averages, not deadlines, but they're a useful starting point if you're trying to figure out why a nap refused to happen.
- 3 to 6 months: 12 to 16 hours of sleep total; 10 to 12 hours overnight; 3 to 4 naps of 30 minutes to 2 hours; wake windows of about 60 to 90 minutes.
- 6 to 9 months: still 12 to 16 hours total; 10 to 12 hours overnight; naps consolidating to 2 to 3; wake windows stretching to 2 to 3 hours.
- 9 to 12 months: 12 to 15 hours total; 11 to 12 hours overnight; naps down to 1 to 2; wake windows of 3 to 4 hours.
Nap transitions, typically 3 naps down to 2 around 6 to 8 months, tend to cluster right alongside the developmental leaps below — which is part of why sleep gets bumpy twice in one year, not once.
The second bump: 8 to 10 months
A second, distinct rough patch shows up for a lot of babies around 8 to 10 months, driven less by sleep architecture and more by everything else going on developmentally: separation anxiety, new mobility like crawling and pulling to stand, and the cognitive leap of object permanence — realizing you still exist even when you leave the room, which is exactly why they now protest it. This one tends to run 2 to 6 weeks. Extra reassurance and a steady, predictable response to night waking are the common threads in the guidance here, not a different approach from the 4-month stretch, just patience through a second wave.
Call emergency services now: blue or grey lips, face, or skin during or after sleep; not breathing, gasping, or choking; or real difficulty waking your baby. These are outside normal sleep development at any age and need immediate emergency care, not a call to the office.
Call your pediatrician: a fever paired with unusual, excessive sleepiness; any pause in breathing that worries you, even briefly; or, if it's persistent rather than a one-off — extreme night waking (think 20-plus times a night for weeks) or naps still under 30 minutes well past 6 months — worth a conversation, not necessarily alarming, but worth ruling out reflux or another underlying cause.
Myths worth retiring
Weighted swaddles and weighted sleep sacks are marketed as calming, but the AAP has flagged that they haven't been safety-tested and has advised against them, since restricted movement and overheating are the concerns. Swaddling in general should stop once your baby starts trying to roll, usually 4 to 6 months, since a swaddled baby who rolls to their stomach can't free their arms. Sleep training doesn't have to mean cry-it-out either: guidance broadly favors consistency over any one specific method, so responsive, gradual approaches count too. And adding cereal to a bottle to help a baby sleep longer isn't supported by the research — solids don't reliably change how much a baby wakes at night.
One more myth worth naming directly, since it's a common source of guilt: the idea that babies should be sleeping through the night by 3 months. Newborn and young infant sleep cycles run about 2 to 3 hours, and consolidated overnight sleep for most babies doesn't really show up until 6 months or later — so a 3-month-old still waking two or three times a night isn't behind, that's the expected pattern for that age, not a sign anything needs fixing.
Sources
- American Academy of Pediatrics (2023). The tension between AAP safe sleep guidelines and infant sleep. Pediatrics. https://publications.aap.org/pediatrics/article/153/4/e2023064675/196919/The-Tension-Between-AAP-Safe-Sleep-Guidelines-and
- NICHD Safe to Sleep campaign. Safe sleep environment. https://safetosleep.nichd.nih.gov/reduce-risk/safe-sleep-environment
- Sleep and infant development in the first year (2026). Pediatric Research. https://www.nature.com/articles/s41390-026-04780-4
- Sleep consolidation and growth in early childhood, systematic review (2024). https://pmc.ncbi.nlm.nih.gov/articles/PMC11873771/
- Longitudinal sleep behavior study, 0–12 months. https://pmc.ncbi.nlm.nih.gov/articles/PMC4173090/
- Video evidence that parenting methods predict which infants develop long night-time sleep periods by three months of age (2018). https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5966725/
- Parent–infant room sharing during the first months of life: longitudinal links with behavior during middle childhood (2020). https://pmc.ncbi.nlm.nih.gov/articles/PMC7379577/
Want gentle reminders and tracking built around this stage?
Start free