When sleep regresses: reading the dip in your log.

Somewhere between three and five months, a baby who had started sleeping in longer stretches goes back to waking every couple of hours. It has a name, the four month sleep regression, and it has a shape in the log. Knowing the shape is what stops it feeling like something you broke.

The word regression is the problem. Nothing is going backwards. Around this age a baby's sleep reorganises into cycles with lighter and deeper stages, closer to how adults sleep, and the change is permanent. Between cycles there is a brief surfacing, and a baby who has not yet learned to drop back down on their own wakes up fully and tells you about it. Your baby did not lose a skill. Their sleep got more complicated.

What moves in the log

Open the last two weeks side by side and three things usually stand out.

The longest stretch gets shorter. This is the line most parents watch, and it is the one that dips hardest. A five hour block that had held for a few weeks becomes two and a half, then two. If you read the nap or sleep stretch post, you know the longest stretch is the number that means the most, which is why this dip feels so personal.

The number of night wakes goes up. Two becomes four or five. Many are short. The longer ones tend to land in the second half of the night.

Naps break into pieces. A ninety minute nap becomes forty minutes, because one cycle ends and the baby does not carry on into the next.

What does not move

This is the more useful half. The total sleep over 24 hours often changes far less than the night suggests. Some of it moves into the day, but the sum is usually in the same band it was in three weeks ago. Nights alone look like a collapse. The whole day looks like a redistribution.

Feeds and diapers hold steady. A baby who is waking more but eating and wetting the same as last week is a baby whose sleep changed, not a baby who is unwell. The wakes may come with feeds, and that is fine. What matters is that the daily counts have not jumped.

The weight line keeps its direction. One tired week does not bend a growth curve. If it did, the curve was the thing to look at, not the sleep.

A normal dip or something else

The regression is a change in sleep with nothing else moving. That is the test. Run it against the log rather than against how you feel at 4am.

A dip in the longest stretch plus steady feeds, steady diapers and a baby who is cheerful enough in the daytime is the regression. Give it a few weeks; for most families it settles as the baby gets better at joining cycles together.

A dip in sleep plus a change in something else is a different sentence. Fewer wet diapers than last week. Feeds that are noticeably shorter or that the baby refuses. A temperature. A baby who is hard to settle in the day too, not just at night. Any of those alongside the sleep change is the thing to describe to your midwife, health visitor or doctor, and the log lets you describe it with dates: "since Friday, four wakes a night instead of two, and diapers down from seven to four." That is a sentence a professional can work with.

Reading it without making it worse

Once you have identified the shape, the individual nights stop being data and become weather. Stop reading the log at 4am. Look at it once, in the morning, at the seven day view, and ask one question: is the longest stretch shorter than last week, or is it the same and I am just more tired. Those feel identical at night and look different on the screen.

This is also the week to check the split between night and day is set the way you actually live. In Little Bean the stats separate night sleep from naps using a night window you set yourself, so a 6pm catnap does not get counted as the start of the night and flatten your longest stretch. If your evenings have shifted, move the window. The dip may be smaller than it looks.

Then hand the log over. If both parents are on the same record, the one who did the 2am and 4am wakes does not have to remember them, and the other can read the night in thirty seconds before taking the morning. The plans page covers what each Little Bean plan includes, and what caregiver invites look like.

What the dip is for

If you kept the log through the first weeks, you already know the shape of your own baby's sleep. That was the point of writing down what three weeks of data tells you: it gives you a last week to compare this week against. The regression is the first time that comparison earns its keep. Without it, four wakes a night is a crisis. With it, it is a change, with a start date, a likely cause and a reasonable expectation that it passes.

Log the wakes, skip the analysis, and look once a day. The longest stretch will come back. It usually comes back longer.

Every baby's sleep is different, and a change in sleep is only one signal. If it comes with fewer wet diapers, feeding refusal, a fever, unusual drowsiness or anything that worries you, talk to your midwife, health visitor or doctor. A clear log helps you describe the last week accurately; a professional is the person to act on it. Nothing here is medical advice.