Why “debt” is the wrong metaphor
The bank metaphor — borrow some sleep on weekdays, repay it on weekends — fits our intuition but not the physiology. The 2003 Spiegel and 2008 Banks & Dinges studies in Sleep ran adults at 4–6 hours per night for 1–2 weeks, then offered recovery. A single recovery night never restored psychomotor vigilance to baseline. Full recovery took 9 or more days of 9–10 hour sleep. Chronic short sleep (months or years) is associated with elevated cardiovascular, metabolic and mood risk that occasional recovery weekends do not fully reverse.
How much do I actually need?
The NSF 2015 expert panel published age-banded ranges. Adults 18–64 land at 7–9 hours; adults 65+ tier slightly lower at 7–8. The ranges are wider than older guidance because individual variation (5–10% of adults are bona-fide short sleepers, 1–3% long sleepers) is real. The way to test where you fall: track total sleep (including weekend lie-ins) for two weeks of unconstrained schedule and average it. That number is closer to your need than any individual weekday count.
The sleep behaviours with the strongest evidence
- Consistent wake time — even on weekends. The single highest-leverage circadian anchor.
- Bright light within 30 minutes of waking — 10,000 lux for 20 minutes, or daylight on a balcony.
- Caffeine cutoff 8–10 hours before bed — half-life of 5–7 hours; afternoon coffee is still active at midnight.
- Cool, dark, quiet bedroom — 18–20°C is the NSF consensus optimum for most adults.
- Wind-down ritual — 30–60 minutes off screens before bed reduces sleep-onset latency.
When this becomes clinical
Persistent insomnia, loud snoring with witnessed apnoeas, feeling unrefreshed after 8+ hours, involuntary daytime sleep, or insomnia paired with low mood or anxiety all warrant a provider conversation. Untreated obstructive sleep apnoea is implicated in roughly 1 in 4 adults with hypertension and meaningfully raises cardiovascular and stroke risk.