Your baby was sleeping a solid four-hour stretch. Maybe even five. You started to feel human again. Then, suddenly, they're up every 90 minutes and nothing you did differently explains it.
Welcome to sleep regressions. The term gets thrown around so loosely in parenting forums that it's lost most of its meaning. Bad nap? Regression. Up at 3am? Regression. Baby looked at you weird? Probably a regression.
Some of these are real developmental events with actual science behind them. Others are just bad weeks that got a label. Here's the difference.
The 4-month regression is the only permanent one
This is the big one, and it's the only regression that changes your baby's sleep architecture for good.
Before 4 months, infant sleep is simple: they cycle between active sleep (similar to REM) and quiet sleep. Two stages, back and forth. This is why newborns can fall asleep anywhere, in any conditions, with noise and lights on. Their sleep is biologically uncomplicated.
Around 3-4 months, a baby's brain reorganizes its sleep into the adult-like four-stage pattern that includes light sleep, deep sleep, and REM. This is a one-way change. They don't go back to the simpler pattern.
The problem isn't the new sleep stages themselves. It's the transitions between them. Adults cycle through sleep stages every 90-110 minutes and briefly surface to near-wakefulness between cycles. We don't remember these micro-wakings because we've spent decades learning to roll over and fall back asleep. A 4-month-old has been doing this for zero days. They surface between cycles, realize they're alone in a dark room, and cry.
This is why the 4-month regression hits so hard and why it doesn't "pass" the way other regressions do. Your baby isn't going back to newborn-style sleep. They're learning a new skill, and that skill takes a while to develop.
What actually helps: whatever sleep association your baby has at bedtime is what they'll need at every cycle transition overnight. If they fall asleep nursing, they'll need to nurse back to sleep at 11pm, 1am, 3am, and 5am. That's not a behavioral problem at 4 months. It's just biology meeting logistics. Parents who want to change this eventually start working on independent sleep skills, but there's no rush. The AAP's safe sleep guidelines focus on safe sleep environment, not sleep training methods, and the AAP doesn't endorse any particular training approach.
8-10 months: the one that's actually about milestones
Somewhere between 8 and 10 months, your baby learns to pull up, cruise along furniture, maybe crawl fast. Their brain is processing a massive amount of new motor information, and it keeps processing at night.
You'll find your baby standing in the crib at 2am, crying, sometimes because they pulled up and don't know how to get back down. This is the most frustrating version of a sleep regression because the solution (learning to sit back down from standing) is a physical skill you can't really teach. You can practice during the day and lower them back down at night, but mostly you wait.
This regression overlaps with separation anxiety, which typically peaks between 8 and 10 months according to the AAP. So you've got a baby who's physically wired from new motor skills AND emotionally distressed when you leave the room. It's a rough combination. Most parents describe this as worse than the 4-month regression because at least at 4 months the baby stayed where you put them.
This one typically passes in a few weeks once the novelty of the new motor skill wears off and separation anxiety settles into a less acute phase.
12 months is barely a regression
The 12-month "regression" is the one that gets the most debate. Some sleep consultants insist it's real. Others think it's just a nap transition getting mislabeled.
Here's what usually happens: around 12 months, a baby starts refusing one of their two naps. Parents assume it's time to drop to one nap. They do. The baby becomes overtired by 4pm and then sleeps terribly at night. Night sleep falls apart, and it looks like a regression.
The fix for most 12-month sleep problems is counterintuitive: keep offering two naps even if they refuse one. Most kids aren't actually ready for one nap until closer to 14-18 months. The refusal at 12 months is usually a developmental blip, not a permanent schedule change.
If your baby is 12 months old and sleeping badly, before you google "12-month sleep regression" check whether anything else changed: new daycare, a dropped nap, illness, teething molars (the first molars typically erupt between 13-19 months and they're miserable). The cause is usually more specific than "regression."
18 months: welcome to opinions
The 18-month regression is the first one where your child's personality gets involved.
At this age, toddlers have strong preferences and the word "no" in their vocabulary. They don't just wake up confused or needing comfort. They wake up and actively resist going back to sleep because they want to be somewhere else doing something else. "No bed." "Want down." "Mama play."
Separation anxiety also has a second peak around 18 months, often more intense than the first because the toddler now understands enough to know that you're in the next room having a life without them. This knowledge is personally offensive to them.
The 18-month regression is also the one where toddlers start testing boundaries around bedtime in ways that feel strategic. One more book. Water. Different water. Need socks. Socks are wrong. This isn't manipulation. They genuinely don't want the day to end. But it does require you to set a routine and stick to it more firmly than you did when they were younger.
Most families report this one lasting a few weeks, though some stubborn toddlers stretch it longer.
The 2-year regression nobody warns you about
By 2, most parents assume sleep regressions are over. They're not.
Two-year-olds develop fears, vivid dreams, and enough imagination to be scared of things that aren't there. The dark. Monsters. That one shadow on the wall that looks like a face. Their sleep disruptions aren't about developmental milestones or sleep cycles anymore. They're emotional.
This is also the age when many families transition from crib to bed, which removes the one physical boundary that was keeping the toddler in place. Now they can get out, walk to your room, and stand silently by your bed at 3am staring at you until you wake up with a cardiac event.
There's no "fix" for this stage the way there is for the 4-month regression. It's managed with consistency: a predictable bedtime routine, a dim nightlight if darkness is the issue, and returning them to their bed calmly and without fanfare every time they appear.
What doesn't help during any regression
Some advice that gets repeated constantly but doesn't hold up:
"Sleep when the baby sleeps." This only works if you have exactly zero other responsibilities, no older children, and can fall asleep on command. For most people it's useless advice.
Changing everything at once. When sleep falls apart, the instinct is to overhaul the entire routine. New schedule, new sleep sack, white noise machine, blackout curtains, moved the crib, started sleep training, stopped sleep training, co-sleeping, not co-sleeping. Changing multiple variables at once means you can't tell what helped and what made it worse. Change one thing at a time and give it 3-5 days before deciding it didn't work.
Keeping them up later so they'll be more tired. This backfires almost every time. Overtired babies and toddlers sleep worse, not better. As Cleveland Clinic pediatric sleep specialists explain, fatigue triggers stress hormones that make it harder to settle down, not easier.
When it's not a regression
Sleep regressions are a convenient explanation for bad sleep, but they're not the only one. Before assuming regression, rule out:
- Ear infections. Babies with ear infections sleep worse when lying flat. If they seem fine during the day and miserable at night, ask your pediatrician to check their ears.
- Teething pain. Inconsistent, but real. Molars are worse than front teeth.
- Illness. Even a mild cold disrupts sleep more in babies than adults because they can't blow their nose or mouth-breathe well.
- Room temperature. Babies sleep poorly when they're too warm. The AAP's safe sleep guidelines warn against overheating and recommend dressing infants in no more than one layer more than an adult would need. Most pediatricians suggest keeping the room around 68-72°F (20-22°C).
- Schedule mismatch. Sometimes the wake windows just need adjusting. A baby who's been on the same schedule for three months might need longer wake windows as they grow.
If bad sleep lasts more than six weeks and nothing you've tried has helped, it's worth talking to your pediatrician. Persistent sleep issues can occasionally signal sleep apnea, reflux, or other medical problems that won't resolve on their own.
The pattern underneath all of it
Every regression follows roughly the same arc: sleep falls apart, you survive it, and then it gets better. Not necessarily back to where it was before, but better. The 4-month regression permanently changes sleep structure, but everything after that is temporary disruption while the brain or body catches up.
If you're tracking your baby's sleep and want to see patterns over time, the Sleep Calculator can help you figure out whether the current wake windows still make sense for their age. Aanvi can help on the milestone side. When you log "pulled to stand" and "sleep fell apart" in the same week, the pattern becomes obvious in a way it never does when you're living through it at 3am.
Already past the regression and recalibrating naps? The Sleep Calculator figures out wake windows by age so you're not guessing.