What "through the night" actually means in the research
The phrase has become so saturated with marketing that most parents do not realize it has a specific clinical meaning. When the pediatric sleep literature reports that 70% of six-month-olds "sleep through the night," it is measuring a five-hour uninterrupted stretch — usually from around midnight to 5 a.m. — not a continuous block from 7 p.m. to 7 a.m. The original 1957 Moore and Ucko study, which produced the first widely cited "through the night" age data, used a midnight-to-five-am window because that was the typical adult sleep span at the time and produced a measurable clinical milestone.
Seven decades later, the definition has held. The 2010 Henderson, France, Owens, and Blampied paper in Pediatrics, still the most-cited modern reference, used a five-hour stretch as the primary outcome. Sadeh's longitudinal work uses six hours. Almost no peer-reviewed study uses the eight-to-twelve-hour adult definition that the parenting industrial complex has trained mothers to expect — because biologically, that pattern is uncommon before the second year.
The practical implication: when a sleep consultant says your baby "should be sleeping through the night by twelve weeks," what they usually mean — and what the literature supports — is a single five- to six-hour stretch. The other six to eight hours are still spread across two to four feeds and as many brief wakings. The Wermom team's editorial review at wermom.com/research covers the original methodological choices behind the five-hour standard and why most consumer sleep guides quietly redefine the term upward.
The realistic age chart, by month
The chart below uses the five-hour stretch as the consolidation benchmark, drawn from the AAP, the National Sleep Foundation's 2024 consensus statement, and the most recent meta-analyses of infant sleep. Every percentage is the share of healthy term babies meeting that stretch on a typical night — not every night, and not every baby.
| Age | Longest typical stretch | % reaching 5-hour stretch nightly | Typical night-feeds remaining |
|---|---|---|---|
| 0–6 weeks | 2–3 hours | < 5% | 3–4 |
| 6–12 weeks | 3–5 hours | ~20% | 2–3 |
| 3–4 months | 4–6 hours | ~50% | 1–2 |
| 4–6 months | 5–7 hours | ~60% | 1–2 (regression common) |
| 6–9 months | 6–8 hours | ~70% | 0–1 |
| 9–12 months | 7–9 hours | ~72% | 0–1 |
| 12–18 months | 8–10 hours | ~75% | 0 |
| 18–24 months | 9–11 hours | ~80% | 0 |
Read the chart with two qualifiers. First, the five-hour stretch is a population average; the individual range is enormous. A six-month-old doing four hours on a typical night is within normal variation. Second, "sleeping through" is not the same as "needing no parental input." Brief partial arousals — what sleep researchers call signaled awakenings — happen four to six times per night across the lifespan, in adults too. The variable is whether the baby self-resettles or calls out.
Why the four-month inflection point matters
Between roughly twelve and sixteen weeks, the infant brain undergoes the single most consequential sleep reorganization of the first two years. Before this window, babies sleep in two simple states (active and quiet sleep) and cycle every 45–50 minutes. After this window, they consolidate to the adult architecture: four sleep stages (light, deeper, deepest, REM), longer cycles, and a clear delineation between night sleep and day sleep. The circadian rhythm — the body's internal clock — also reaches near-adult maturity, and melatonin production stabilizes in evening hours.
This is why the four-month sleep regression exists. The brain is upgrading the operating system, which means previous strategies that worked (rocking to sleep, feeding to sleep, the same nap rhythm) often stop working overnight. It is also why the 50% "through the night" milestone clusters at this age — the architectural change makes longer stretches biologically possible for the first time. Wermom's deep guide to the four-month regression walks through what's actually happening at the neurological level.
The catch: the same architecture that produces longer stretches in some babies produces dramatically worse sleep in others for two to six weeks before stabilizing. About a third of babies regress around four months and then consolidate; about a third stay roughly the same; about a third make a clean step forward. None of these is a sign of how your baby will sleep at twelve months.
The six-to-nine-month window and the night-feed taper
Between six and nine months, two things converge: most babies become biologically capable of going eight or more hours without a feed, and most have established the day-night circadian distinction firmly enough that the longest stretch lands in the night, not the day. This is the window when the population percentage at the five-hour mark crosses from ~60% to ~70%, and where most pediatricians introduce the conversation about night weaning if a family wants it.
"Biologically capable" does not mean "always doing it." About one in three nine-month-olds still wake at least once for a feed, particularly if exclusively breastfed. The CDC's infant feeding guidance recognizes that night-feed frequency varies widely and is not a reliable indicator of caloric adequacy at this age — daytime intake and growth are. Some babies use a night feed for comfort, not hunger, and the distinction matters for whether and how a family chooses to taper.
The Wermom Sleep Desk position, reviewed by our pediatric sleep advisors, is that night weaning is a family choice, not a developmental milestone, and that there is no medical urgency to eliminate night feeds in a healthy thriving baby under twelve months. The pressure to do so is cultural, not biological.
Why the first-year wakings are not your fault
The most over-cited statistic in the sleep training literature is that 28–35% of one-year-olds wake at least once nightly. It is presented in product marketing as a problem to solve. In the pediatric sleep medicine community, it is presented as a baseline — the normal variation of human infant sleep. The 2024 NIH-funded longitudinal infant sleep cohort confirmed this distribution and noted that night waking at twelve months has no measurable association with later cognitive, behavioral, or attachment outcomes.
What does have measurable association: parental sleep loss, postpartum mood disorders, and the daytime functioning of the caregiver. Which means the relevant question is rarely "what is wrong with the baby" — it is "what does the family need." The Wermom team's analysis of the National Institute of Child Health and Human Development's caregiver sleep guidance walks through the harm-reduction strategies that do not require changing the baby — shared night responsibilities, daytime nap discipline for the parent, and the rotation patterns that actually preserve cognitive function in the postpartum year.
What actually predicts earlier consolidation
Three factors show up consistently in longitudinal data as predictive of earlier night consolidation, and they are not the ones most product marketing emphasizes.
Consistent bedtime within a 30-minute window. Babies on a stable bedtime — same hour, same routine, same lighting cues — consolidate roughly two to four weeks earlier on average than babies on variable bedtimes. The mechanism is circadian: predictable evening light and feed cues entrain melatonin release. This effect is measurable from about eight weeks onward.
Daytime exposure to bright light, particularly morning sunlight. Infants exposed to outdoor or window-bright morning light produce more cohesive circadian rhythms by twelve weeks, per Harrison's foundational 2004 study and replicated in subsequent cohorts. The mechanism is the same as in adults: morning light sets the internal clock for the evening melatonin window.
Allowing brief unassisted resettling at sleep transitions. Not cry-it-out — just a 60-to-120-second pause when the baby fusses between sleep cycles before intervening. About 60% of those fusses resolve on their own; intervening immediately teaches the baby that the cycle transition requires parental help. This is the single largest mechanistic finding in the modern self-soothing literature.
None of these is a sleep training method. They are environmental and behavioral inputs that nudge the biology toward consolidation a few weeks earlier on average. Wermom's evidence-based settling guide covers each one with the specific protocols.
Here's how Wermom App makes this 10x simpler
The hardest part of tracking sleep consolidation is the math. A parent in a postpartum fog cannot reliably remember whether last night's longest stretch was four hours or five and a half. Wermom App handles the tracking so the pattern is legible without the cognitive load:
- Longest-stretch tracker — automatically calculates the longest uninterrupted block each night against the age-appropriate benchmark, so the five-hour milestone is a moment, not a guess.
- Regression overlay — flags the 4-month, 8-month, 12-month, and 18-month windows on the timeline so a sudden change is contextualized as developmental, not pathological.
- Bedtime consistency score — rolling 14-day variance on your baby's bedtime, with one tap to tighten the routine — the single largest evidence-based predictor of earlier consolidation.
The shorter answer, for the parent who needs a number tonight
If your baby is under three months and doing two-to-three-hour stretches, that is biology, not a problem to solve. If they are four to six months and consolidating to a single five-hour stretch, that is right on time. If they are nine months and still waking once, that is within the normal third of the population. The chart above is a curve, not a deadline. The Wermom editorial position, reviewed by our advisors, is that the eight-to-twelve-hour expectation parents inherit is a cultural artifact, not a developmental milestone — and that the five-hour clinical definition is the honest number to anchor on. For the broader editorial mission behind the Wermom team's sleep coverage and the advisors who shaped it, see our editorial principles.