Around 18 months, a previously good sleeper starts waking at 3 AM, fighting bedtime, and crying when you leave the room. It is hormonal, neurological, and temporary. Here is what is happening and what works.
6 min readMay 2026
TL;DR Hits between 17 and 19 months, lasts 2 to 6 weeks. Drivers: separation anxiety surge, language explosion, molar teething, growing independence. Hold the existing schedule. Skip new sleep training; this is not the moment. Add 5 minutes of bedtime cuddle to address the separation piece. Avoid bringing the toddler into your bed — habit-forming and harder to reverse. By week 4-6, baseline sleep usually returns.
It started at month 17. They went down at bedtime like normal, then woke at 2 AM screaming. Twenty minutes of holding, back to sleep. Then awake again at 4. Then refusing the morning crib. By the third night you wondered if the sleep training from a year ago had just expired.
It did not. This is the 18-month regression, and unlike the 4-month one (which was permanent neurological reorganization), this one ends.
What is actually happening
Three things converge around month 18:
Separation anxiety peaks. Toddlers at 18 months suddenly understand object permanence in a new way — they know you exist in another room. The cortisol response to bedtime separation spikes.
Language explosion. The vocabulary jumps from ~50 to ~300 words between 16 and 24 months. Brain development that intense disrupts sleep architecture.
Molar teething. The first molars erupt between 13 and 19 months. Bigger teeth, longer eruption pain.
Growing independence. The "no" phase starts. Bedtime becomes another battleground for the toddler's emerging autonomy.
None of these are problems to fix. They are developmental milestones that show up as sleep disruption.
Not sure if this is actually the regression?
Five questions tells you: the regression you think you're in, an adjacent one, or one of the imposters (teething, illness, schedule problem). Each result comes with a 4-bullet action plan.
Average: 3 to 4 weeks. Range: 2 to 6 weeks. About 10 percent of toddlers have a more severe regression that lasts 6-8 weeks. Anything past 8 weeks is no longer a regression — it is a new sleep pattern that needs different intervention.
The 18-month regression is mostly about new independence. The fix is consistency, not a new sleep system.
What works (evidence-based)
Hold the current schedule. The same wake-up time, the same nap, the same bedtime. The body knows the schedule; the regression disrupts execution, not the schedule itself.
Add 5 minutes of bedtime presence. Sit by the crib, read one extra book, or pat their back. The separation anxiety eases when they know you are leaving voluntarily, not abandoning.
Address molar pain before bed. If you suspect teething: a cold washcloth to chew, infant acetaminophen (talk to your pediatrician for dose), a topical teething gel that does not contain benzocaine.
Comfort-object continuity. A small lovey, the same blanket, white noise — anything that signals "this is sleep time" reduces the new fear response.
Wait it out at the 2 AM wake. Wait 5-10 minutes before going in. Most regression wakes resolve on their own; intervention can reinforce the pattern.
What does not work (and often makes it worse)
Starting sleep training. A toddler in a regression cannot self-soothe through this. Cry-it-out during separation-anxiety peak can extend the regression by weeks.
Bringing the toddler into your bed. Hard to reverse. The "just this one night" turns into 6 months. If you do this once, immediately restore the prior arrangement the next night.
Cutting the nap. Counterintuitively, dropping the nap to "build up tiredness" backfires — overtired toddlers sleep worse, not better.
Pulling out screen time as comfort. Screens before sleep delay melatonin release in toddlers as in adults. The 10-minute fix becomes a 45-minute sleep onset delay.
Adding a late snack. Fine if they actually wake hungry. Counterproductive if they are just up — reinforces "wake = food" pattern.
Verify the schedule is age-right
One contributor to the 18-month regression is a wake-window that has drifted too short. Re-check your toddler's age-appropriate windows.
If naps stay short for more than 4 weeks despite consistent wake windows, check for a real nap transition rather than fighting the regression.
When it is something else
Persistent night terrors (not just crying — sitting up, eyes open, screaming, unresponsive): a sleep-disorder pattern. Resolves on its own usually by age 5 but worth a pediatrician mention.
Snoring or mouth-breathing. Common during a cold, abnormal if persistent. Worth an ENT referral — could be sleep apnea or enlarged tonsils.
Sleep disruption with daytime behavior change. Sudden regression in language, walking, or social engagement alongside sleep changes warrants a pediatric evaluation.
Regression that lasts longer than 8 weeks. Time to consult a pediatric sleep consultant or your pediatrician.
If the regression sticks past 4 weeks, run a 4-week reset: stricter routine, no schedule changes, and protect the bedtime hour as the anchor.
The 4-week reset (if it has lasted that long)
If the regression hits 4 weeks and is still active, a structured reset can help:
Audit the schedule. Wake window between nap and bedtime should be 5-6 hours at 18 months. Too short = under-pressure at bedtime; too long = overtired.
Tighten the wind-down. 30-minute routine: bath, pajamas, book, lights low, in crib. Same order every night.
Reduce nighttime intervention. If you are going in repeatedly, gradually extend wait times (5 min, 10 min, 15 min) over 4-5 nights.
Hold the morning wake-time. Even if they were up at 3 AM, keep the 7 AM wake. Sleeping in shifts the schedule and prolongs the disruption.