"Sleep regression" is one of the most-googled phrases in the first two years of parenting, and one of the least precisely used. It can mean a baby who used to sleep five-hour stretches now waking every two hours; a toddler resisting bedtime for the first time; or anything in between. Knowing which regression you are actually in — and which one is genuinely different from the others — changes what you should and should not do.
Healthbooq gives age-appropriate, evidence-based sleep guidance and helps you log the pattern so you can tell whether what you are seeing fits the typical regression shape or needs a different look.
What a Sleep Regression Actually Is
A regression is a stretch of disrupted sleep in a child who was previously sleeping reasonably well, occurring during or just after a burst of developmental change — motor, cognitive, or both. Despite the name, nothing has gone backwards: the brain is doing more, and sleep is the system that pays the cost in the short term.
Three things make this period temporary:
- The new skill (rolling, crawling, walking, vocabulary) consolidates within a few weeks
- Sleep architecture stabilises around the new neural pattern
- The behavioural changes (bedtime resistance, calling out, more wakes) ease as the underlying driver settles
If a sleep disturbance is not preceded by a stretch of decent sleep, or has no developmental correlate, it is more likely a sleep issue than a regression — see the section at the end on telling them apart.
The Four-Month Regression — the Odd One Out
Around 3–4 months (sometimes as early as 12 weeks, sometimes as late as 5 months), infant sleep architecture matures permanently. Newborn sleep has only two states (active and quiet); from around four months, sleep cycles become adult-like — clearly distinct light (N1, N2), deep (N3), and REM phases, with brief surfacings between cycles roughly every 45–60 minutes.
The clinical implication: those brief surfacings, which previously passed unnoticed because the newborn just rolled into the next cycle, now expose how the baby gets to sleep. If the baby falls asleep at bedtime in arms, on the breast, in the buggy, or being rocked, the same conditions are typically required at every cycle surface — which is why feeds that were every 4 hours suddenly become every 1–2.
This is why the four-month regression often does not "pass" on its own. The maturation is permanent. What needs to change — usually — is the baby's ability to fall asleep at the start of the night without the same external help being needed for the rest of it. See the dedicated four-month regression article for the actual approaches.
The 6–10-Month Regression
Drivers cluster:
- Object permanence consolidating (around 6–8 months)
- Separation anxiety onset (typically 6–10 months, peaking at 10–18 months)
- Major motor milestones — sitting, crawling, pulling to stand
- Often the start of solids around 6 months (NHS / WHO)
- The three-to-two nap transition for many
Unlike the four-month regression, this one is genuinely temporary — usually 2–4 weeks once any one driver settles, sometimes 4–6 if several stack. Hold the routine, keep the cot safe for the new mobility, do not drop naps reactively.
The 12-Month Regression
Coincides with first independent steps (NHS milestone range 9–15 months), language take-off, and the run-up to the 2-to-1 nap transition (which usually completes at 14–18 months, not at 12). The most common error here is dropping the second nap on regression evidence; this almost always produces an overtired bedtime that prolongs the disruption.
The 18–24-Month Regression
Drivers: vocabulary explosion (50 → 200+ words is typical between 18 and 24 months), assertion of autonomy (the "no" phase), eruption of second molars (anywhere from 23–33 months but often earlier), and the first appearance of nighttime fears as symbolic thinking comes online. Bedtime resistance is the headline; calling out at night and stalling tactics are common.
What to Do During a Regression — and What Not To Do
For all regressions except the four-month:
Do- Keep the routine identical — predictability is the most useful tool you have
- Respond to need without escalating involvement (brief, calm, in-cot where possible)
- Keep environment stable: dark room, 16–20°C (Lullaby Trust), white noise if already used
- Wait 2–3 weeks before considering schedule changes
- Use weight-appropriate paracetamol or ibuprofen for clear teething pain on a bad night, per BNF for Children — not as a routine sleep aid
- Re-introduce feeding to sleep, rocking to sleep, or parental bed unless you are happy to keep doing it for months
- Drop naps in response to nap refusal — that is not a readiness sign on its own
- Move to a toddler bed in response to bedtime fights
- Start sleep training in the middle of an active regression — wait for the underlying developmental burst to settle
For the four-month regression, the answer is usually different: gentle work on how the baby falls asleep at the start of the night (placing into the cot drowsy but awake, gradually reducing the parental help) is what makes the change last.
Regression vs. Sleep Problem
Calling something a regression that has actually been a chronic issue can mean weeks of waiting for it to "pass" when it never will. Some quick distinctions:
| | Regression | Sleep problem |
|—|—|—|
| Preceded by good sleep? | Yes | No / never well-established |
| Visible developmental driver? | Usually | No |
| Resolves with consistency in 2–6 weeks? | Yes | No |
| Associated symptoms? | None | Possible: snoring, breathing pauses, weight faltering, daytime distress, ear-pulling, persistent congestion |
If you are 6–8 weeks in and nothing is changing, look for an underlying cause (see GP for snoring/mouth-breathing, weight, reflux, congestion) or address it as a sleep issue rather than continuing to wait.
When to Get Help Sooner
- Persistent fever, vomiting, off feeds, or poor weight gain — see GP / NHS 111
- Snoring, mouth-breathing, or pauses in breathing during sleep — flag to GP (possible adenotonsillar issue)
- Parental mental health affected by sleep loss — this is a legitimate reason to seek support; mention to health visitor or GP
- A regression in a child who has never slept well — this is more likely a sleep issue from the start; consider a specialist sleep consult or a sleep clinic referral
Key Takeaways
Sleep regressions are temporary stretches — typically 2–6 weeks — when a previously settled baby starts waking more, fighting bedtime, or napping less, triggered by developmental change rather than illness or bad habits. The four-month regression is the odd one out: it reflects a permanent shift in sleep architecture, so it does not 'pass' on its own the way the others do — it usually requires looking at how the baby falls asleep at the start of the night. The most common ages are around 4 months, 6–10 months, 12 months, and 18–24 months. The single biggest mistake during any regression is introducing a new sleep crutch (feeding to sleep, parental bed, rocking) that long outlasts the regression itself.
