Do You Really Need Eight Hours?

It is close to midnight, you are still awake, and somewhere at the back of your mind a small accountant is doing sums. If you fall asleep now you will get six and a half. That is not eight. You have failed the night before it has properly started, and the failing itself will keep you up another twenty minutes.

Almost everyone can quote the number. Very few of us could say where we got it.

The belief, named kindly

The folk belief goes like this: adults need eight hours a night, that requirement is the same for everyone, and every hour you fall short of it is quiet damage being done to you. It is a generous belief in its way. It came from people who wanted us to take rest seriously, at a time when rest was treated as something lazy people did. It is also the reason a great many people lie in the dark feeling like they are behind on a debt.

So it is worth asking a plain question. Who actually says eight?

Nobody says eight

This was the part we did not expect. Go to the two big consensus documents that the field itself defers to, and neither of them recommends eight hours.

The first is the National Sleep Foundation’s duration report, put together by an eighteen-member panel drawn from twelve organisations, working through a structured method for grading recommendations. Their answer for adults aged 18 to 64 is a range: seven to nine hours. For adults over 65, seven to eight. The report is careful in both directions. It says that durations outside the recommended range may still be appropriate for some individuals, and adds in the same breath that deviating far from the normal range is rare, and that people who habitually sleep well outside it may be showing signs of a health problem, or may be wearing themselves down by choice.¹

The second is the joint statement from the American Academy of Sleep Medicine and the Sleep Research Society. Their panel voted that seven to nine hours was appropriate and six or fewer was not, and then made an interesting decision. They published a minimum threshold rather than a range: seven or more hours. They declined to name an upper bound at all, because they could not agree that longer sleep was physiologically harmful, or describe a plausible mechanism by which it would be.² That same document reports twin studies putting the heritability of sleep duration somewhere between 31 and 55 per cent. The field’s own consensus paper concedes, in passing, that need is partly inherited and therefore partly personal.

Eight sits inside both answers. It is simply not either of them.

The two big sleep panels don’t say eight. One says a range of seven to nine. The other just sets a floor: seven or more. Eight was never the rule — it’s the middle of everyone else’s.

What the population actually does

Eight is not the population centre either. In one of the largest datasets on this, 446,118 adults in the UK Biobank, average habitual sleep came out at 7.2 hours with a standard deviation of 1.1.³ Take that at face value and roughly two-thirds of adults are living somewhere between about 6.1 and 8.3 hours a night. Eight is not the middle. It is near the upper edge of ordinary.

Two honest caveats on that figure. Most of it is self-reported, and people tend to overestimate their sleep when you compare their answers against a sleep lab, so the real average is probably a shade lower again. And a volunteer cohort of this kind skews healthier and more European in ancestry than the world at large. It tells you what a very large group of adults report, which is a different thing from what any one body requires.

Where individual need shows up in the biology

Here the evidence gets genuinely interesting, and also more easily misused than any other part of this subject.

In 2009 a research group found a family who slept short and seemed fine on it. The carriers averaged 6.25 hours a night against 8.06 in their non-carrier relatives, and the difference tracked a single point mutation in a gene called DEC2. The team then engineered the same mutation into mice, which showed longer active periods, more wakefulness, and a blunted recovery response after being kept awake. They put it into fruit flies as well, and the flies slept less too.⁴

A decade later, a second and independent short-sleep gene turned up: a rare variant of the β1-adrenergic receptor in a family whose affected members sleep four to six hours their whole lives. Knock-in mice slept about 55 minutes less per day, and the researchers could point at the circuit doing it — a population of wake-active neurons in the dorsal pons that the mutant receptor made easier to switch on.⁵

This is where we have to be careful, because the popular version of these studies is much stronger than the studies are. The human side of the DEC2 work is two affected carriers in one family. The second finding is one family as well, and even inside it one carrier slept seven and a half hours and did not meet the strict definition. What makes both convincing is not the human sample size but the fact that the same edit reproduces the same effect in engineered animals. That establishes a mechanism. It does not hand anybody a diagnosis. True short sleepers are vanishingly rare, and most people who describe themselves that way are simply running short and coping.

The broader genetics point the same way, gently. The Biobank work found 78 genome-wide-significant loci for sleep duration, and together they account for about 0.69 per cent of the variance, with an average effect around one minute per allele.³ Duration is partly biological and thoroughly diffuse. Nothing in that data assigns you a number.

One more piece, from the other direction. Twenty-one healthy adults were kept awake for 36 hours on three separate occasions, with their prior sleep deliberately manipulated beforehand — restricted to six hours a day for a week, or extended to twelve. How badly each person fell apart stayed remarkably consistent from one exposure to the next, and the differences between people were far larger than the effect of all that manipulated prior sleep.⁶ Vulnerability to sleep loss looks like a stable individual trait. It is a small study of young lab volunteers, and it measures response to losing sleep rather than baseline need, which is a related question and not the same one. Still, it means that comparing your own tiredness to someone else’s is a poor instrument.

The part of the folk belief that is true

None of the above is permission to sleep less. The same consensus statement that refuses to name an upper bound is unambiguous about the lower one: habitually sleeping under seven hours is associated with weight gain, higher blood pressure, cardiovascular disease, low mood and increased mortality.² Those are associations from population data rather than proof of cause, and the direction of the arrow is not always obvious, since illness disturbs sleep as readily as poor sleep wears on health. But the floor is where the evidence is strongest, and it is the part of the folk belief that survived.

So the grading falls out cleanly. There is a real kernel — short sleep is worth taking seriously. The claim built on top of it, that eight is the universal requirement and anything less is damage, was never what the evidence said.

The wonder

Sleep need is written in a language older than mammals. Change one letter of one gene, one proline swapped for one arginine, and a person sleeps six hours and wakes genuinely rested. Put that same single-letter change into a mouse, and the mouse sleeps less. Put it into a fruit fly, and the fly sleeps less. Three species separated by hundreds of millions of years, and the same tiny edit turns the same dial in all of them. Whatever sleep is for, it is old enough and fundamental enough that evolution has kept the volume knob in the same place since before there were animals with backbones. And it set your knob, personally, at a slightly different point from the person lying next to you.

What this might mean for your evening

Mostly it means one fewer thing to be graded on. If you are watching a number and doing arithmetic in the dark, you are measuring yourself against a figure that no expert panel actually issued, and that sits above what most adults report getting. The better questions are the dull ones: do you wake without a fight most mornings, and are you upright and reasonably clear by mid-afternoon without leaning hard on caffeine. Those are cruder measures than a number, and they are closer to the thing you actually care about.

It also means protecting the floor rather than chasing the ceiling. Seven or more is where the evidence has weight. Above that, the research is comfortable letting you find your own level, and gives you no reason to think a consistent seven and a quarter is a shortfall.

And it means the last hour of an evening is worth more attention than the tally at the end of it. That hour is the part you can actually shape: the lights, the noise, the pace at which the day is allowed to stop. It is the hour Mythrae is built around, and what we make for it is audio for winding down rather than anything that promises you a number. If you would like to see what we are making for that part of the night, the sleep page is the door, and there is no arithmetic in it.

← Back to The Lab