The first winter on warehouse nights I added a column to my spreadsheet called "owed". Every morning I typed in how far short of eight hours I had come, and every Sunday I slept until 2 p.m. and cleared the column like a bill I had paid off. It took me months to notice that the column filled back up faster than Sunday could empty it. That running total is sleep debt, and the research on it is a lot less forgiving than my Sunday routine was.
What is sleep debt, counted honestly
Sleep debt is the gap between the sleep you need and the sleep you actually got, added up night after night. If you need eight hours and get six and a half, you are ninety minutes behind. Do that for five weeknights and by Friday you are seven and a half hours short, which is close to a whole night missing from the week.
The subtraction is easy. The first number is the hard one, because nobody hands you your need. The National Sleep Foundation panel (Hirshkowitz et al., Sleep Health, 2015) recommends 7 to 9 hours for adults from 18 to 64 and 7 to 8 hours from 65 on. An expert panel agreed on that range after reading the research. It does not measure you.
Two more things trip people up.
- Time in bed is not sleep. Say the light goes off at 11 p.m., you take twenty minutes to fall asleep (researchers call that sleep latency), and you lie awake for a while at 4 a.m. Your eight hours in bed might hold only seven hours of sleep.
- The debt does not work like a tidy bank balance. Your body does not pay back lost hours one for one, and it seems to write off some of what you lose. The studies below are the best picture I have found of how the ledger really behaves, and a simple spreadsheet gets parts of it wrong.
Six hours a night, and a brain that stopped noticing
The study I think about most is by Van Dongen, Maislin, Mullington and Dinges, published in Sleep in 2003. They kept 48 healthy adults in a lab. Most were given 4, 6 or 8 hours in bed a night for two weeks, and a smaller group was kept awake for three nights in a row, so the researchers could compare slow restriction with no sleep at all.
The main test was dull on purpose. You watch a screen, and when a counter appears you press a button. The researchers counted lapses, their word for the times you respond too slowly because your attention slipped away for a moment.
The eight hour group stayed about where they started. The other two did not.
- The six hour group got a little worse every day. By the end of the two weeks their lapses looked roughly like those of people who had been kept up for one full night.
- The four hour group got there faster and ended up looking like people who had missed about two nights.
The performance scores never leveled off. The debt kept piling up for as long as the study ran.
The part that matters most for anyone doing this at home is what people said about themselves. Their ratings of how sleepy they felt rose in the first few days and then mostly flattened, while their test scores kept sliding. After a week or two on six hours, people rated themselves as only a bit sleepy while performing like someone who had stayed up all night.
That is exactly how my "owed" column felt from the inside. By the third week of short nights I would have told you I had adjusted. I had not. I had just lost the ability to tell.
The authors also built a model to estimate where the damage starts. Their figure was about 15.8 hours awake a day, which works out to a little over eight hours of sleep. I treat that as a lab average from young, healthy volunteers, not a law, and your line may be lower or higher. It is still the reason my starting target is eight and not seven.
How far behind you might be without knowing it
Van Dongen showed how debt builds. Kitamura and colleagues tried to measure how much of it ordinary people already carry, in a paper in Scientific Reports in 2016.
They took fifteen healthy young men whose usual sleep at home averaged a little under seven and a half hours. Then they gave them twelve hours in bed every night for nine nights, with no alarm and nothing to get up for.
On the first nights the men slept far longer than usual. Then the totals slowly drifted down, night after night, and settled at roughly 8.4 hours. That left a gap of about an hour against their normal nights. The authors called it potential sleep debt: a shortfall these men carried every day without feeling sleepy enough to notice.
A few things about that study are worth holding onto.
- The drift took days. The long first nights were paying back debt, and the totals did not level out after one night or even a weekend's worth.
- An hour a night is not dramatic. Nobody in that group would have called themselves sleep deprived.
- The sample was fifteen young men in a quiet lab with nothing else to do, which is small and narrow. Women, older adults and people with different lives might show a smaller gap, a bigger one, or none at all.
I read it as a warning about my own baseline. My "normal" week used to be seven hours, and I suspect I settle nearer eight when given the chance. If so, I was running a small debt all the time, and the weekends I thought were for catching up were only covering the interest.
Why one long Saturday does not catch up on sleep
A lie-in is not useless. After short nights your next sleep gets deeper, and you do feel better when you wake. But whenever researchers test whether a weekend can undo a short week, the answer keeps coming back as "partly".
Belenky and colleagues, in the Journal of Sleep Research in 2003, restricted volunteers to 3, 5, 7 or 9 hours in bed for a week, then gave everyone three recovery nights of eight hours. The groups that had been cut short improved over those three nights but did not get back to where they started. Three normal nights were not enough to clear one short week.
Depner and colleagues, in Current Biology in 2019, tested the pattern most of us live: five short weeknights, a weekend of sleeping as much as you like, then short nights again. The weekend did not undo the drop in how well their bodies handled sugar, and people on that schedule snacked more after dinner. Once the short nights came back, whatever the weekend had fixed was gone.
Then there is the body clock. Sleeping until noon on Saturday tells your internal clock that the day starts late. On Sunday night you are not tired at your usual bedtime, and Monday morning arrives on a clock that thinks it is still early. Wittmann and colleagues named this in Chronobiology International in 2006: social jetlag, the gap between when your body sleeps on free days and when your schedule makes it sleep on work days.
My Sunday habit ran into all three problems. I got some sleep back but not all of it. Then I spent Sunday night staring at the ceiling until 1 a.m., which put the first hour of Monday's debt on the books before the week had started.
How I pay it back now
None of what follows comes from a single study. These are rules of thumb I worked out by putting the research next to my spreadsheet, and I have changed them more than once.
Stop borrowing first
Paying back debt while you add new debt keeps you on a treadmill. Before I add any extra sleep, I make sure the coming nights at least hit the target. If a week is going to be short no matter what, I admit it and plan the payback for the week after.
Add a little each night, at the start of the night
I pay back 30 to 60 minutes a night, and I take it by going to bed earlier. My alarm stays within about an hour of its usual time, weekends included. That keeps the body clock where it is and spares me the Sunday ceiling. If you want a bedtime that does not land in the middle of a cycle, the sleep calculator counts back from your alarm, and I explain the counting in what time should I go to bed.
Use a short nap on bad days, not as the plan
A 20 minute nap in the early afternoon takes the edge off without wrecking the night. A two hour nap at 5 p.m. does wreck it. The nap calculator sets the alarm so the nap ends before it gets heavy. I keep naps for one bad afternoon at a time, because an hour slept at 4 p.m. is an hour of sleepiness you no longer have at bedtime.
Only count the last two weeks
My spreadsheet only looks back fourteen days. I chose that, and no study told me to. Van Dongen ran two weeks, and I have no honest way to put a number on debt from last spring. For short sleep that has gone on for months, the Kitamura pattern suggests you do not repay it hour for hour anyway. You start sleeping enough, keep it up, and let the totals settle over a week or two.
If you are giving yourself the hours and still wake up exhausted day after day, or someone has told you that you snore hard and stop breathing in your sleep, talk to a doctor, because that is beyond what a sleep spreadsheet can see, and I am plain about what I am and am not.
Planning the week with the sleep debt calculator
The arithmetic above is simple enough to do on paper, and I did it on paper for longer than I like to admit. It is also easy to get wrong at 11 p.m., when you are tired and generous with yourself. The sleep debt calculator adds up the nights you enter against a target and shows how many nights of a little extra it takes to get back to even.
Two tips for the inputs.
- Enter sleep, not time in bed. If you are not sure, take twenty to thirty minutes off each night for falling asleep and waking in the night. That is a rough rule of thumb, and yours may be less.
- Pick your target honestly. If you have no idea, start at eight hours. Van Dongen's line sat just above eight and Kitamura's volunteers settled around 8.4, so seven is optimistic for most people. If you ever get a week or more with no alarm, ignore the first few nights, average the rest, and use that as your own number.
If mornings are what hurts most, can't wake up in the morning explains why sleep debt is the first thing I check before I blame the alarm.
Tonight, write down roughly what you slept on each of the last seven nights and put those numbers into the calculator. Then go to bed thirty minutes earlier than last night, and leave the alarm exactly where it is.