The coffee ends.
The exposure continues.

An antagonist, not a replacement for sleep.

Adenosine is a signalling molecule involved in sleep–wake regulation. Its effects depend on the receptor and the brain circuit. During extended wakefulness, adenosine signalling in particular regions contributes to the pressure to sleep. This is one component of a much larger system, alongside the circadian clock, light exposure and the networks that maintain wakefulness.

At usual dietary exposures, caffeine blocks adenosine A₁ and A₂A receptors. An antagonist occupies a receptor without producing the same effect as its usual signalling molecule. Blocking these receptors can reduce sleepiness and promote vigilance. It does not remove the need for sleep or simply “clear adenosine out of the brain”. A₁ and A₂A receptors do different things in different circuits; the diagram below deliberately explains competition rather than pretending to reproduce an entire brain network. [1]

Why the time on the cup is not the end of the dose.

Caffeine is mainly metabolised in the liver, particularly through CYP1A2. A commonly reported adult half-life is roughly 3–7 hours, although it can fall outside that range. Half-life is the time required for the amount to decrease by half under the model’s assumptions. Pregnancy, medication, smoking and other individual factors can change clearance. The number of hours since the drink is therefore only part of the explanation. [2]

The interactive model isolates one principle: a larger initial dose can leave a larger residual amount even after several half-lives. It assumes immediate absorption and simple first-order elimination, ignores active metabolites and repeated drinks, and is not a prediction of blood concentration or sleep quality. The sliders are a way to understand the arithmetic, not a personalised caffeine test.

What happened when sleep was measured?

First, combine controlled studies.

Gardiner and colleagues’ 2023 systematic review included 24 controlled crossover studies. The analysis of objectively measured total sleep time used 20 studies and 340 participants. Caffeine was associated with an average reduction of 45.3 minutes compared with control, with a 95% confidence interval of 29.0–61.5 minutes less sleep. This describes the tested doses and timings collectively; it is not the expected effect of every cup. [3]

The review’s dose-and-timing regression produced estimates of 8.8 hours for a 107 mg coffee and 13.2 hours for a 217.5 mg pre-workout serving. Those are model-derived estimates, not biological boundaries. Moving ten minutes across one of them does not turn a drink from harmless into harmful. Nor does a non-significant result for a lower dose establish that it cannot affect a particular person.

Then test dose and time in the same people.

A later double-blind crossover trial studied 23 men, average age about 25, with habitual intake below 300 mg per day. They completed placebo and six caffeine conditions: 100 or 400 mg taken 12, 8 or 4 hours before bedtime. Sleep was assessed at home with partial polysomnography and diaries. [4]

No statistically significant effect was detected for 100 mg at the tested times. At 400 mg, several sleep outcomes worsened, with some effects present even 12 hours before bed. Subjective sleep ratings did not always reflect objective disruption. This small, male-only experiment helps explain why dose matters, but cannot establish a universally safe four-hour rule or predict responses in pregnancy, insomnia or other populations.

Chocolate is not a small dose of every stimulant.

Cocoa contains caffeine and theobromine, but chemical relatedness does not make substances interchangeable. Theobromine is another methylxanthine; its effects and potency differ from caffeine. Describing chocolate as equivalent to cannabis or amphetamine confuses chemical associations with what an ordinary food actually does.

In a placebo-controlled, within-person study of 80 healthy participants, Baggott and colleagues compared 250, 500 and 1,000 mg theobromine with placebo and 200 mg caffeine. Theobromine had limited subjective effects at 250 mg; higher doses produced more negative mood effects and raised heart rate. These findings do not validate a fixed number of chocolate squares as a pre-workout intervention. The study tested isolated compounds rather than a typical chocolate bar. [5]

Flavanols are a separate research question. COSMOS randomised 21,442 older adults to cocoa extract or placebo within a factorial trial. The supplement provided 500 mg flavanols per day. Over a median 3.6 years, the primary cardiovascular endpoint was not significantly reduced: hazard ratio 0.90, 95% CI 0.78–1.02. A favourable secondary result for cardiovascular death is interesting but does not overturn the primary result or establish that eating chocolate prevents cardiovascular disease. An extract is also not nutritionally equivalent to a confectionery product. [6]

Turn the evidence into a better question.

If sleep is the concern, record the amount and timing of caffeine as well as the bedtime. “One coffee” is a poor exposure measure because servings and preparation vary. Consider coffee, tea, energy drinks, pre-workout products and other sources together. A consistent sleep window and a short log make patterns easier to interpret than judging one isolated night.

A practical self-observation is to move the final caffeinated drink earlier while keeping the rest of the routine as stable as feasible. Compare several nights, including time to fall asleep, awakenings and next-day functioning. This is an informal observation, with expectation effects and ordinary night-to-night variation; it is not a diagnostic experiment. Reduce gradually if withdrawal symptoms make an abrupt change difficult.

What remains uncertain is the personal dose–timing combination that affects a specific sleeper, how repeated intake changes that response, and how much can be inferred from subjective sleep alone. Persistent sleep problems deserve assessment beyond simply adjusting coffee.

Read the sources.

1. Reichert et al. Adenosine, caffeine, and sleep–wake regulation: state of the science and perspectives (2022). Mechanistic review; animal and human evidence are distinguished.
2. Temple et al. The Safety of Ingested Caffeine: A Comprehensive Review (2017). Metabolism, half-life and sources of variation.
3. Gardiner et al. The effect of caffeine on subsequent sleep (2023). Systematic review and meta-analysis.
4. Gardiner et al. Dose and timing effects of caffeine on subsequent sleep (2025). Randomised crossover trial; 23 men.
5. Baggott et al. Psychopharmacology of theobromine in healthy volunteers (2013). Placebo-controlled experiment.
6. Sesso et al. COSMOS cocoa-extract trial (2022). Primary clinical endpoint and secondary outcomes.