Night four was the one that almost broke us.

Nights one through three of gentle sleep training had gone better than I expected — some crying, some pacing the hallway, but she was settling within twenty minutes each time. Then night four happened. My daughter cried for over an hour, harder than she had on night one, and at minute forty I was standing outside her door with my phone in my hand, one thumb over Mei’s contact, ready to text “we’re stopping this.”

Mei found me in the hallway first. She’d been reading while I did the check-ins, and she said something I didn’t want to hear: “This might actually mean it’s working.” I didn’t believe her. Four nights of BloomPath research later, it turned out she was right, and I want to walk through exactly why, because almost nobody warns you about the night that gets worse before it gets better.

TL;DR

  • An “extinction burst” is a real, well-documented behavioral phenomenon: when a behavior (crying to summon a parent) stops working, it often gets more intense for a night or two before it fades — not because the method failed, but because it’s starting to work.
  • Research on graduated extinction methods shows the burst typically shows up somewhere between night two and night seven, often right after a night or two of apparent progress, which is exactly the moment parents are most tempted to quit.
  • A five-year follow-up study in Pediatrics (Price et al., 2012, N=326) found no long-term differences in attachment, emotional or behavioral outcomes, or parent-child closeness between families who used behavioral sleep methods and those who didn’t.
  • The burst is not the method breaking down — it’s the old strategy trying harder one last time. Understanding that distinction is most of what gets a family through it.
  • Sleep training under 6 months isn’t recommended by pediatric sleep researchers, and if a worsening night comes with fever, teething pain, or illness, that’s a different problem — stop and address the illness first.

Why Does Gentle Sleep Training Get Worse Before It Gets Better?

Because crying at bedtime has been working for your child’s entire life, and a few nights of it not working isn’t enough to convince their brain to give up on it yet.

Behavioral scientists call this an “extinction burst” — it’s the same principle that explains why a vending machine that eats your money makes you hit the button harder a few more times before you walk away, not less. For months, your child has learned a clear pattern: cry, and someone comes, picks you up, or stays until you’re asleep. When you shift to a gentle method — Pick Up/Put Down, the Chair Method, gradual fading, whatever you’re using — you’re not removing comfort, but you are changing the terms. Your child’s brain doesn’t process that shift by immediately giving up on the old strategy. It processes it by trying the old strategy harder, for a night or two, to see if it still works.

I found this genuinely useful the way an engineer finds a root cause useful: it’s not a mystery, it’s not a sign we broke our kid, it’s a documented pattern with a name and a mechanism. The Journal of Clinical Sleep Medicine has published discussion specifically on why extinction-based sleep interventions are hard for parents, and the burst phenomenon is central to that difficulty — not because the science is shaky, but because living through it at 2am feels nothing like reading about it in a journal.

How Long Does an Extinction Burst Actually Last?

Usually a night or two, though it can show up anywhere from night two through night seven, and it often arrives right after a stretch that looked like progress.

That timing is the cruel part. One documented pattern in graduated extinction research: baseline crying before the intervention runs 35-40 minutes, and then during the first couple of intervention nights, it can climb past 70 minutes — nearly double — before dropping off. Parents read “gentle sleep training” and picture a smooth downward curve. The actual shape is closer to a curve that dips, spikes hard for a night or two, and then drops. If you don’t know the spike is coming, night four looks exactly like proof the method failed. If you do know it’s coming, night four looks like exactly what the research said would happen.

We kept a rough log on our phone — bedtime, first cry, total crying minutes, when she settled — mostly because I’m the kind of person who needs a spreadsheet to feel calm about anything. Looking back at it after the fact, our curve matched the pattern almost exactly: 22 minutes, 18 minutes, 31 minutes, then 68 minutes on the worst night, then 14 minutes the night after that. If I hadn’t been tracking it, I would have remembered night four as “the night it stopped working” instead of what it actually was, which was the tail end of the process.

Is It Actually Safe to Push Through the Burst? What Does the Long-Term Research Say?

Yes, according to the best long-term data available — a five-year follow-up study found no lasting harm to attachment or parent-child closeness from behavioral sleep methods.

This was the question that mattered most to me, more than whether the method “worked” in the short term. A study published in Pediatrics — Price and colleagues, 2012 — followed 326 infants, 173 of whom were randomly assigned to a behavioral sleep intervention (graduated extinction or “camping out”), and checked back in on the whole group five years later. The researchers found no measurable differences between the intervention group and the control group on emotional and behavioral outcomes, parent-child closeness and conflict, attachment security, or parental mental health at the five-year mark. The short-to-medium-term benefit the researchers did find was a reduction in maternal depression symptoms — which tracks with what almost every exhausted parent already suspects: a family running on two hours of broken sleep a night is not the version of the family that shows up best for anyone, including the kid.

I’m not going to pretend this fully erased my discomfort in the hallway at minute forty. It didn’t. But it gave me something to hold onto besides raw guilt, which is that the discomfort I was feeling was about the sound, not about actual evidence of harm.

What Do You Actually Do on the Worst Night?

Pick one gentle method, commit to it fully for at least five to seven nights before judging it, and use your partner as backup instead of a second opinion mid-crisis.

We used a hybrid of Pick Up/Put Down and the Chair Method — pick her up briefly if she stood at the crib rail distressed, put her back down the second she was calm, no rocking to sleep, and I moved my chair a little farther from the crib every two nights. None of that is a secret method; it’s laid out well in Elizabeth Pantley’s No-Cry Sleep Solution and other gentle-method books. What I want to add here is the part that method guides usually skip: the decision-making script for the actual worst night.

Here’s what got us through night four specifically:

  1. We picked a “we don’t decide anything after 11pm” rule before we started. Any decision to stop, switch methods, or cave gets made in daylight, with both of us awake and not running on adrenaline. This alone stopped at least three “let’s just give up” conversations from becoming permanent decisions.
  2. We tag-teamed instead of both hovering. Mei did the first half of check-ins, I did the second half. Neither of us sat there alone accumulating dread for a full hour — we split the exposure to the crying, which mattered more for our own nervous systems than I expected it to.
  3. We looked for illness signs before assuming it was a burst. Fever, a new tooth coming in, an ear tug — any of those and we’d have stopped and treated the actual problem instead of pushing through pain we misread as protest.
  4. We wrote down the time and duration, even at 2am, half-asleep. This turned a subjective feeling (“this is worse, it’s not working”) into an objective data point I could check against the pattern the next morning.
  5. We agreed in advance on what “not working” would actually look like — no improvement at all after seven full nights, not “one bad night in the middle of an improving trend.” Having that line drawn ahead of time is what kept night four from becoming the night we quit.

What If the Burst Doesn’t Pass — When Should You Actually Stop?

Stop and reassess if there’s no improvement at all after a full week, if your child is under 6 months, or if something other than sleep habits — illness, a developmental leap, a house move — is clearly driving the distress.

Pediatric sleep researchers generally don’t recommend structured sleep training methods before 6 months, since younger infants have real nutritional and developmental needs that can look identical to a sleep-habit issue but aren’t. And if you’re two weeks in with zero trend in the right direction, that’s a different signal than one rough night in an otherwise improving pattern — that’s worth a conversation with your pediatrician, not more persistence.

Eleven years into this, the thing that still surprises me is how much of parenting comes down to knowing which discomfort is information and which discomfort is just discomfort. Night four wasn’t a sign we’d broken something. It was the sound of an old habit trying one last time before it let go.

BloomPath uses illustrated AI characters (Mei and Ethan) to protect our daughter’s privacy. The content is real; the avatars are illustrated. Learn more →

Frequently Asked Questions

What is an extinction burst in sleep training? An extinction burst is a temporary increase in a behavior — in this case, crying at bedtime — right after that behavior stops getting the response it used to get. It’s a well-documented pattern in behavioral science: the old strategy gets tried harder for a night or two before it fades, not because the new approach is failing, but because it’s starting to work.

How many nights does an extinction burst usually last? Most documented patterns show the burst lasting one to two nights, though it can appear anywhere between night two and night seven of a gentle sleep training method. It often shows up right after a stretch of nights that looked like real progress, which is exactly when it’s easiest to mistake it for failure.

Is it harmful to let a baby cry through an extinction burst during gentle sleep training? The best available long-term data doesn’t support that concern. A five-year follow-up study of 326 infants published in Pediatrics found no lasting differences in attachment security, parent-child closeness, or emotional and behavioral outcomes between children who went through behavioral sleep interventions and those who didn’t.

How do I know if it’s an extinction burst or something else, like illness? Check for fever, teething signs, congestion, or a recent developmental leap before assuming it’s a burst. An extinction burst tends to follow a pattern of prior improvement; illness or a genuine developmental change usually comes with other symptoms beyond just harder crying at bedtime.

When should gentle sleep training actually be stopped instead of pushed through? Stop and reassess if there’s no improvement at all after a full week, if your child is under 6 months old, or if something other than a sleep habit — illness, teething, a big life change — is clearly the real driver. One bad night inside an otherwise improving trend is not the same signal as a full week with zero progress.

Products We Recommend

These are the specific things that got us through the week, not a generic sleep-product list:

Want an easy way to track bedtime, wake-ups, and crying minutes without opening a spreadsheet at 2am? The BloomPath app has a sleep log built for exactly this. Download BloomPath →