Sleep deprivation: the cognitive cost most people underestimate

A classic 2003 study by Hans Van Dongen and David Dinges put 48 healthy adults on one of four schedules for 14 days: 8 hours of sleep per night, 6 hours, 4 hours, or 0 hours. Performance was assessed daily on standardized cognitive tasks. The findings rearranged what the field had assumed about chronic sleep restriction.

The 4-hour and 6-hour groups showed steadily worsening performance across the two weeks. By day 14, their cognitive impairment matched the impairment seen in subjects after a single full night of total sleep deprivation. The subjects themselves reported feeling roughly fine — their self-assessment had recalibrated to the new normal (Van Dongen et al., 2003).

This is the central finding of modern sleep research: chronic moderate sleep restriction produces cognitive deficits that the impaired person cannot reliably detect.

1. The performance cost

The Van Dongen study used the Psychomotor Vigilance Task (PVT) — a simple reaction-time task that's been used in sleep research for decades. The PVT is unusually informative because it's resistant to learning effects and sensitive to sleep state.

PVT performance after 17-19 hours of wakefulness matches the impairment seen at a blood alcohol level of 0.05% — past the legal limit for driving in most countries (Dawson & Reid, 1997). Most people don't conceptualize chronic sleep restriction in terms equivalent to a moderate alcohol level, but the data is consistent.

2. The self-assessment problem

Subjects in sleep restriction studies overestimate their own functioning. The Van Dongen team's most consequential finding wasn't the impairment — it was the unreliable awareness of the impairment. Subjects rated themselves as roughly functional even as objective measures dropped into the territory of severely sleep-deprived comparison subjects.

This has practical implications. The shift worker, medical resident, or parent of a newborn experiences chronic moderate sleep restriction and rates themselves as adapted. They are not adapted. They are recalibrated to a new baseline that they can't accurately compare against their full-rest baseline.

3. What's actually affected

Sleep restriction affects different cognitive functions differently:

Most affected: sustained attention, working memory, reaction time, decision-making under risk, emotional regulation.

Moderately affected: episodic memory, learning, creative problem-solving.

Least affected: crystallized intelligence, well-rehearsed motor skills, simple pattern recognition.

The asymmetry explains why a tired person can drive a familiar route to work without noticeable problems but can't do anything that requires sustained attention or novel decision-making with full competence.

4. The recovery question

The 2003 study and subsequent work suggests recovery from chronic sleep restriction takes substantially longer than the duration of the restriction. Two nights of 10 hours doesn't fully restore baseline after two weeks of 6-hour nights. The full restoration timeline is on the order of weeks, not days (Belenky et al., 2003).

This argues against the popular "sleep debt" model, in which short nights can be repaid on weekends. The data are not consistent with that framing.

5. The practical implication

For most adults, the minimum sleep that doesn't produce measurable cognitive deficit is closer to 7 hours than the popular "5-6 hours is enough" claim. Individual variation exists — some people genuinely function well on 6 hours, identifiable by genetic short-sleeper variants — but they are rare. Most people who report functioning fine on 6 hours are operating in the Van Dongen-style adapted-to-impairment zone.

The cost shows up in things you don't notice you've lost.