What the Research Says
What the Caffeine Research Actually Shows
Caffeine is the most studied psychoactive substance on earth. The research is unusually clear — here's what it actually shows about performance, timing, and tolerance.
The Most Studied Stimulant in the World
Caffeine has an unusually good evidence base — it's cheap, legal, widely consumed, and easy to study in controlled conditions. The result is hundreds of high-quality RCTs with consistent findings. Most people's caffeine habits are poorly calibrated relative to what the research supports.
What Replicates Strongly
Caffeine reliably improves sustained attention, reaction time, and vigilance. The effect is robust across doses of 100–400mg, study designs, and populations. The dose-response curve is roughly linear up to about 300mg, after which additional gains are smaller and side effects increase. For cognitive work requiring sustained focus, caffeine is one of the few interventions with a genuine, large effect.
The half-life is 5–6 hours on average — longer than most people assume. A 200mg dose (roughly a large coffee) at 2pm leaves approximately 100mg active at 8pm and 50mg at 1am. Polysomnography studies show that late-afternoon caffeine degrades slow-wave sleep even when subjects report sleeping fine. Individual half-life varies 2–10 hours based on CYP1A2 genotype, liver function, and oral contraceptive use.
A "caffeine nap" outperforms caffeine or a nap alone. Drinking coffee then immediately napping for 20 minutes times the caffeine peak to hit just as you wake, producing better alertness scores than either intervention alone. The mechanism: adenosine (which caffeine blocks) clears during the nap, leaving caffeine receptors uncontested when you wake.
Tolerance develops within days of regular use. Daily caffeine use produces near-complete tolerance to subjective alertness effects within 7–12 days. The implication: habitual caffeine users are mostly restoring baseline rather than enhancing above it. Strategic use — with regular breaks — preserves efficacy.
Timing Is More Important Than Dose
The "delay your first coffee 90 minutes after waking" recommendation has a mechanistic basis: adenosine accumulates during sleep and continues clearing for 60–90 minutes after waking. Blocking its clearance with early caffeine may reduce the natural cortisol-driven alertness peak. Some RCT evidence supports this, though effect sizes are modest.
The more impactful timing question is the cutoff. Studies measuring objective sleep architecture (not just subjective quality) consistently support a caffeine cutoff of 12–2pm for most people — though individual variation is large.
What the Research Can't Tell You
Your optimal dose, timing, and cutoff depend on your CYP1A2 genotype, sleep architecture, and baseline adenosine sensitivity — none of which you can read from the literature. The averages are a starting point. Your own data is more actionable.