Sleep cycles and the 90-minute alarm rule

What a sleep cycle actually describes

Sleep alternates between non-REM and rapid eye movement, or REM, phases. Non-REM includes lighter stages and deep sleep. NHLBI describes cycles as restarting about every 80–100 minutes, with more deep sleep earlier and more REM later in the night. That describes a changing biological pattern; it isn't a personal timetable you can read from the bedside clock. (NHLBI: sleep phases and stages)

What researchers measured

A retrospective study published online in October 2023 analyzed 6,064 recorded cycles from 369 healthy, good sleepers in controlled laboratory studies. The median cycle lasted 96 minutes, with substantial variation within and between people. Participants were aged 19–35 or 55–80 and had eight-hour sleep opportunities. These recordings weren't a trial of alarm strategies, so 96 minutes isn't a replacement formula. The selected sample and pooled studies limit generalization. The paper reports foundation and research-agency support for several authors, who declared no conflicts of interest. (Cajochen and colleagues: measured sleep-cycle variation; University of Liège: full study text (PDF))

A correct calculation can still answer the wrong question

Imagine a calculator that works backward from a 7:00 a.m. alarm. Five blocks of 90 minutes equal seven and a half hours, giving 11:30 p.m. as the starting time. The arithmetic works. The output is a schedule built from an assumption, not a recording of what your brain will do.

If you add a fixed 15 minutes to fall asleep, the displayed bedtime moves to 11:15 p.m. But the added quarter-hour is still a number someone entered. It has not become a measurement just because it appears beside a precise clock time.

Think of the distinction as calendar time versus observed sleep. A sleep study uses sensors, including brain activity and eye movements, to classify sleep stages. A calculation using bedtime alone doesn't have those observations. The practical inference is to avoid treating its output as a guaranteed stage boundary. (NHLBI: sleep phases and stages)

Those example times are an illustration, not a recommended schedule. You don't need to replace a 90-minute rule with a different magic number. Ask what the number is based on and what decision it is helping you make.

Check the amount of sleep before the number of cycles

Four 90-minute blocks add up to six hours. Calling them four complete cycles doesn't change the total. CDC guidance recommends at least seven hours for adults aged 18–60, seven to nine for ages 61–64, and seven to eight for adults 65 and older. Individual needs and sleep quality also matter. (CDC: recommended sleep duration and sleep quality)

NHLBI recommends making enough time for sleep and keeping bedtimes and wake times consistent. Those are more useful planning anchors than deliberately shortening the night to reach a neat multiple. Allow time for your evening to wind down, too; brushing your teeth and finding tomorrow's keys still happen outside the calculation. (NHLBI: making time for sleep)

One useful planning check

Tonight, look at your actual evening before opening another calculator. Here are three ordinary scheduling questions:

• What time do I realistically need to get up, including the time to get ready?
• What is currently filling the end of my evening: work, chores, entertainment, or preparations I could do earlier?
• Is there one small task I can move, such as packing my bag after dinner, to make bedtime less rushed?

Choose the question that fits your life. This is a planning exercise, not a test of willpower or a promise of easier mornings. If your schedule is constrained, identifying that constraint is useful information; another minute-perfect bedtime may not solve it.

When the problem needs more than arithmetic

If lack of sleep is affecting your daily activities, talk with a healthcare professional. An ongoing sleep problem deserves attention beyond repeatedly adjusting an alarm. (NHLBI: when sleep problems affect daily activities)