Caffeine hides the tired signal instead of adding energy
Caffeine does not put energy into your brain. It sits in the docking sites of a molecule called adenosine, one of the brain's signals to slow down, and while it sits there that message is not received. The cup hides a signal. It does not refill anything.
A signal made from spent fuel
Brain cells run on ATP, the molecule cells use to carry energy. When ATP is broken down, one of the leftovers is adenosine. Reichert, Deboer and Landolt's 2022 review describes how, after long stretches of nerve activity while awake, ATP builds up outside the cells and is broken down to adenosine there. Adenosine is thought to rise outside the cells as waking goes on, and it has been proposed as one of the ways the brain keeps a running total of how much it needs sleep.
Adenosine acts through receptors, protein docking sites on the surface of cells. Of the four kinds, two matter most for sleep: A1 and A2A. Drugs that switch either of them on increase deep, slow-wave sleep. At A1 receptors, adenosine turns down the release of glutamate, the brain's main "go" messenger, which may quiet the areas that keep you awake.
Blocking, not fuelling
A drug that fits a receptor without switching it on is called an antagonist. Caffeine is an adenosine antagonist that binds A1 and A2A receptors about equally well. About half an hour after you drink it, it reaches the brain and occupies those receptors, so adenosine arriving there has fewer places to land.
Which receptor carries the wake-up effect has been tested in mice. Caffeine still kept mice awake when their A1 receptors had been genetically removed, but not when their A2A receptors had. Removing A2A receptors from just one small region, the shell of the nucleus accumbens, was enough to stop caffeine's wake-promoting effect. The main substance the body turns caffeine into, paraxanthine, also binds these receptors in animal studies, so it may add to the effect.
How long one cup lasts
The liver clears caffeine, mostly with one enzyme, CYP1A2. The rate is described by a half-life: the time for the amount in the body to fall by half. The clinical review by Evans, Richards and Battisti puts it at about 5 hours in the average adult. Smoking can shorten it by up to half, and late in pregnancy it can stretch to as long as 15 hours.
That makes the arithmetic simple and the answer surprising. A 100 mg cup at 3 pm with a 5-hour half-life still leaves about 33 mg in the body at 11 pm, eight hours later. A 2023 meta-analysis of 24 studies by Gardiner and colleagues found that caffeine cut total sleep by about 45 minutes and sleep efficiency by about 7%, and lengthened the time taken to fall asleep by about 9 minutes. Their estimate: to avoid losing sleep, a coffee of about 107 mg should be drunk at least 8.8 hours before bed, and a 217.5 mg pre-workout drink at least 13.2 hours before.
Interactive Choose a cup: how much caffeine, when you drink it, how fast your body clears it, and when you go to bed. Drag across the chart, or use the time slider, to see how much is left at any hour and how many receptors it is sitting in. Every control works with Tab and the arrow keys.
Why the daily cup feels less like magic
A brain that meets caffeine every day adjusts. Reichert and colleagues summarise animal work in which adenosine levels rose during daily caffeine, and some studies in which cells made more A1 receptors. They report that caffeine lowers sleepiness most clearly after more than a day without it, or when sleep pressure is unusually high. When caffeine was given over several days, studies found no significant change in how sleepy people felt compared with a placebo.
Rogers and colleagues tested 369 people after a night without caffeine. In the regular, heavier drinkers, the night without caffeine showed up as more sleepiness, lower alertness and poorer reaction-time and memory scores, worse by the afternoon, and caffeine improved those measures in them. In people who rarely drank it, caffeine made them less sleepy but also more anxious and jittery, so their alertness and mental performance barely changed. Caffeine sped up simple movements and reaction times in both groups. The authors' conclusion: tolerance to caffeine's effect on sleepiness means that frequent use fails to make people more mentally alert. Much of what the morning cup does for a regular drinker is undo the night's withdrawal.
Adenosine is not the whole story of tiredness
The popular picture has adenosine piling up all day like sand in an hourglass, to be swept away by sleep. The measurements are messier. In rats, adenosine jumps when the active period begins and then stays roughly level, more like a step than a slope, and newer sensors in mice show it changing over minutes, not hours. Reichert and colleagues conclude that adenosine is unlikely to be what drives the build-up of sleepiness over an ordinary day, though it may add to sleep pressure when you stay awake far longer than usual. The body clock, and other sleep-promoting substances, also play a part. Caffeine blocks one real signal; it does not reach the others.
How much is a lot
For most adults, the US Food and Drug Administration cites 400 mg a day as an amount not generally associated with harmful effects, while noting wide differences in how sensitive people are and how fast they clear it. It advises anyone pregnant, trying to become pregnant or breastfeeding to ask a health care provider about limits. It estimates that toxic effects such as seizures can appear after rapidly taking in about 1,200 mg, less than half a teaspoon of pure caffeine powder.
In short
Caffeine works by blocking adenosine receptors, mainly A2A, so one of the brain's "slow down" messages goes unheard. Nothing is added; the signal is hidden for as long as the caffeine lasts, which for an average adult means half of it is still there five hours later. Regular drinkers adapt, so the daily cup largely brings them back to normal rather than above it.
Where this comes from
- Adenosine, caffeine, and sleep–wake regulation: state of the science and perspectives (Journal of Sleep Research 31(4), e13597) linked only, not reproduced
doi.org/10.1111/jsr.13597 - Caffeine (StatPearls, NCBI Bookshelf NBK519490) linked only, not reproduced
www.ncbi.nlm.nih.gov/books/NBK519490/ - The effect of caffeine on subsequent sleep: a systematic review and meta-analysis (Sleep Medicine Reviews 69, 101764) linked only, not reproduced
doi.org/10.1016/j.smrv.2023.101764 - Faster but not smarter: effects of caffeine and caffeine withdrawal on alertness and performance (Psychopharmacology 226, 229–240) linked only, not reproduced
doi.org/10.1007/s00213-012-2889-4 - Spilling the Beans: How Much Caffeine is Too Much? (consumer update) linked only, not reproduced
www.fda.gov/consumers/consumer-updates/spilling-beans-how-much-caffeine-too-much