Sleep Science
How much sleep do you actually need, and who decided on eight hours?
The recommendation is a population range, not a personal prescription. Individual requirement varies substantially, a rare genetic variant permits four hours without penalty, and both ends of the curve carry risk.
The National Sleep Foundation recommends seven to nine hours for adults aged 18 to 64, and seven to eight for those 65 and over. These figures come from expert consensus panels reviewing the epidemiological literature, not from a single experiment establishing an optimum.
They describe a range within which most people function well. They are not a target each individual must hit.
The U-shaped curve, and its complications
Large cohort studies consistently find a U-shaped relationship between self-reported sleep duration and mortality. Both short sleep — usually under six hours — and long sleep — usually over nine — are associated with higher mortality than seven to eight.
The short-sleep side has plausible mechanisms: sleep restriction experiments produce impaired glucose tolerance, increased blood pressure, altered appetite hormones, and impaired immune response.
The long-sleep side is harder to interpret and is probably substantially reverse causation. Illness causes long sleep — depression, cancer, chronic infection, heart failure, sleep apnoea all increase time in bed. Analyses excluding people with existing disease attenuate the association considerably.
Two further caveats deserve weight. Almost all of this is self-reported duration, and people are poor estimators. And time in bed is not sleep — someone reporting nine hours may be spending nine hours in bed obtaining six hours of fragmented sleep.
A small number of people carry mutations — in DEC2, ADRB1 and NPSR1 among others — that allow them to sleep four to six hours with no measurable cognitive or health penalty, and no daytime sleepiness. Ying-Hui Fu's group has identified several families. The variants are rare. The overwhelming majority of people who believe they need only five hours are simply chronically sleep-deprived and have adapted their expectations rather than their physiology.
How to tell whether you are getting enough
The subjective sense is unreliable, because chronic sleep restriction degrades the ability to judge one's own impairment. In Hans Van Dongen and David Dinges's much-cited study, participants restricted to six hours a night for two weeks showed cognitive deficits accumulating steadily to a level comparable to two nights of total sleep deprivation — while their self-rated sleepiness plateaued after a few days. They felt about as bad on day three as day fourteen. They were not.
Better indicators:
- Do you need an alarm to wake? Consistently requiring one suggests you are waking before your sleep is complete.
- Do you sleep substantially longer at weekends? A gap of more than an hour or so indicates a weekday deficit.
- Do you fall asleep within five minutes? That is a sign of significant sleep pressure, not of being a good sleeper. Normal sleep latency is roughly ten to twenty minutes.
- Are you sleepy when inactive — reading, in meetings, as a passenger?
The most reliable personal test is a holiday. Given two weeks with no alarm and no schedule, sleep duration typically drops over the first few days as accumulated debt is repaid, then settles at a stable value. That value is roughly your requirement.
Sleep debt, and what it does and does not do
Recovery sleep is real but incomplete. After restriction, the first recovery night shows increased slow-wave sleep — the deepest sleep is prioritised — followed by REM rebound on subsequent nights. The architecture rebalances.
But repayment is not one-for-one, and some deficits persist. Studies of recovery from chronic restriction find that certain measures, particularly sustained attention, do not fully normalise after a single extended night.
The weekend lie-in is therefore a partial fix at best, and it carries its own cost: shifting sleep timing at weekends produces social jetlag, and a body clock repeatedly asked to move two hours in each direction does not settle.
What changes across a lifetime
Newborns sleep 14 to 17 hours. Requirement declines steeply through childhood. Adolescents need 8 to 10, more than adults, at exactly the age when schools start early and social life starts late.
In adulthood, requirement is relatively stable. In older adults, total requirement changes less than commonly claimed — what changes is the ability to obtain it in one consolidated block. Slow-wave sleep declines markedly, fragmentation increases, and the circadian phase advances. The frequent conclusion that older people need less sleep is not well supported; they get less.
A reasonable position
Seven to nine hours is a sensible default. If you wake without an alarm, feel alert through the day, and do not need to catch up at weekends, you are getting what you need — even if that is six and a half hours.
If you are relying on an alarm, sleeping three extra hours on Saturday, and drinking coffee to function, the number you need is more than the number you are getting, and no amount of adaptation is going to change that.
