How Much Does One Bad Night’s Sleep Actually Cost You?

Unevenly, and the pattern matters more than the total. A meta-analysis of 70 articles covering 147 cognitive tests found the effects of short-term total sleep deprivation ranged from large for lapses in simple attention, at a Hedges’ g of minus 0.776, down to small and non-significant for reasoning accuracy, at minus 0.125.

So sleep deprivation hits sustained attention hardest and leaves careful reasoning comparatively intact, which is not what most people expect and has direct consequences for what you should and should not attempt after a bad night.

What the meta-analysis measured

The study is Lim and Dinges’ 2010 meta-analysis in Psychological Bulletin.

Its motivation was a specific gap. A large amount of research existed on short-term total sleep deprivation, defined as less than 48 hours, but there was disagreement on how to interpret it, arising partly because the relative magnitude of effect sizes across domains was not known.

So they pooled effects for both speed and accuracy across six cognitive categories: simple attention, complex attention, working memory, processing speed, short-term memory and reasoning. Seventy articles containing 147 cognitive tests met their inclusion criteria.

The range spans almost the whole scale of effect sizes: from large, for lapses in simple attention at minus 0.776 with a confidence interval from minus 0.96 to minus 0.60, to small and non-significant, for reasoning accuracy at minus 0.125 with an interval from minus 0.27 to 0.02.

That interval for reasoning accuracy crosses zero. Sleep-deprived people in these studies were not reliably worse at reasoning accurately.

Why the pattern is the finding

A later review summarising this literature puts the shape plainly: acute total sleep deprivation of 24 to 48 hours impairs neurocognitive functioning across almost all cognitive domains, but there is substantial variability in sensitivity across domains, with the largest decrements for sustained attention and working memory, while comparatively more complex tasks are affected to a lesser degree.

Complex tasks holding up better than simple ones is counterintuitive. One proposed explanation in this literature is about engagement rather than difficulty: researchers examining task characteristics noted that although cognitive tasks required different types of processing, they encouraged participants to remain attentive to and engaged in the task, whereas vigilance tasks were less intrinsically interesting and engaging.

That account should be held loosely. In a closely related field, research on noise and cognition, a 2023 review reports that increasing task engagement did not reduce disruption in the way earlier work had suggested, and that a task-difficulty finding failed to replicate. Engagement explanations of this general kind have not fared well recently, and this one has not been tested directly here.

What this does not cover

Two scope limits, significant for anyone reading this after a short night rather than a missed one.

The meta-analysis concerns total sleep deprivation, meaning complete sleep loss over an extended period of less than 48 hours. That is an all-nighter, not six hours instead of eight.

The same paper notes total deprivation is less common than long-term chronic sleep loss, in which sleep time is partially reduced over days or weeks. Chronic partial restriction is the more common real-world pattern and is not what these effect sizes describe. Nor is it the same question as catching up at weekends.

So the figures here are a reasonable guide to the morning after a night with no sleep, and a poor guide to the cumulative effect of a bad fortnight.

What to do with the pattern

Reading across the findings, the practical implication is about task selection rather than effort.

If lapses in simple attention take the largest measured hit and reasoning accuracy takes the smallest, then the things to be most careful about after severe sleep loss are those requiring sustained vigilance over time: monitoring, checking, reviewing, anything where the failure mode is missing something rather than getting it wrong.

The corollary is not that you should do complex work while exhausted. It is that a familiar intuition, that you should save the hard thinking for when you are rested and knock out routine work while tired, may have it backwards on this evidence.

If the bad night followed a drink, alcohol before bed explains why the damage concentrates in the second half of the night. And a daytime nap has its own small measured effect, strongest for alertness, which is the domain this analysis found most vulnerable.

When this stops being about one night

Persistent unrefreshing sleep is different from an occasional bad night, and none of this research addresses it.

If you regularly cannot fall asleep or stay asleep, wake unrefreshed most days, snore heavily or have been told you stop breathing at night, or find daytime sleepiness affecting your driving, see a doctor. Insomnia and sleep apnoea are both common and both treatable, and neither is addressed by managing your task list better.

If you are deliberately going without sleep to meet deadlines on a regular basis, that is worth discussing with someone too, whether a doctor or whoever sets the deadlines.

Common questions

What suffers most when I have not slept?

Sustained attention, by a wide margin. A meta-analysis of 70 articles covering 147 cognitive tests examined short-term total sleep deprivation across six categories: simple attention, complex attention, working memory, processing speed, short-term memory and reasoning. The largest effect was for lapses in simple attention, at a Hedges’ g of minus 0.776 with a confidence interval from minus 0.96 to minus 0.60. The smallest was reasoning accuracy at minus 0.125, with an interval crossing zero and non-significant. A later review describes the largest decrements as being in sustained attention and working memory, with more complex tasks affected to a lesser degree.

Does that mean I can still do complex work?

Not exactly, and the finding is narrower than that. What the evidence shows is that reasoning accuracy was the least affected measure and simple attention lapses the most affected, in studies of total sleep deprivation under 48 hours. One proposed explanation, from researchers examining task characteristics, is that cognitive tasks encouraged participants to remain attentive and engaged, whereas vigilance tasks were less intrinsically interesting. That account deserves caution: in the adjacent literature on noise and cognition, a 2023 review reports that task-engagement and task-difficulty explanations have failed to replicate. It is a reason to be wary of monitoring work when exhausted, not a licence to attempt anything demanding.

Does this apply to sleeping six hours instead of eight?

No. The meta-analysis concerns short-term total sleep deprivation, meaning complete sleep loss over a period of less than 48 hours. The paper itself notes this is less common than long-term chronic sleep loss, in which sleep time is partially reduced over days or weeks. Partial restriction over time is the pattern most people experience and is not what these effect sizes measure. Treat the figures as describing the morning after an all-nighter rather than the cumulative effect of a run of short nights.

How bad is the biggest effect, really?

A Hedges’ g of minus 0.776 for lapses in simple attention is a large effect by conventional standards, and its confidence interval from minus 0.96 to minus 0.60 does not come close to zero. That is one of the more robust findings in cognitive psychology and it concerns the specific thing measured: lapses, meaning moments where a response does not come, on tests of simple sustained attention. It is not a general measure of being worse at your job. If sleep problems are persistent rather than occasional, that is a matter for a doctor, since insomnia and sleep apnoea are common and treatable.

Save the checking, not the thinking

The received advice about working tired is to do the easy things and postpone the hard ones. This evidence suggests that instinct is aimed at the wrong distinction.

What collapses under sleep loss is the ability to stay present through something that does not hold your attention. Lapses in simple attention showed the largest effect in the whole analysis. What holds up comparatively well is accurate reasoning, which showed the smallest and was not statistically significant.

The distinction that matters is therefore not hard versus easy. Whether it is engaging versus dull, as one proposed account has it, is less certain than it looks, since engagement explanations in adjacent fields have recently failed to replicate. What is solid is the pattern itself: vigilance suffers most, reasoning accuracy least.

None of which makes working exhausted a good plan. It does mean that if you have to, the review of someone else’s work and the long checklist are the tasks to be most suspicious of yourself on, and those are exactly the ones people reach for when they feel too tired to think.