Reviews/Caffeine and Performance/Caffeine improves alertness
Cognition · CAFF-COG-01 · v2
Borderline: within 4 points of the Established/Likely boundary, so a re-grade could land either side.
Across many independent, blinded, randomised trials, caffeine consistently made people respond faster on reaction-time and vigilance tasks, in e-gamers, athletes, general adults and sleep-restricted soldiers ([1],[4],[5],[6],[7],[8],[9],[11]). Several trials also found people felt more alert after caffeine ([6],[8]). The panel treated the reaction-time half of the claim as strongly and repeatedly supported, and the alertness half as supported but with real caveats. One trial found the alertness lift largely disappeared in people who were not habitual drinkers and may partly reflect undoing overnight withdrawal rather than a true boost ([5]). Other trials qualified the picture: one saw reaction time improve only at a higher dose ([4]) and another found its reaction-time changes were not statistically significant ([10]). The direction of the effect is consistent enough that the panel placed the claim at the top tier while noting these tempering findings.
The corpus is a dense bank of double-blind randomised crossover and parallel trials. Reaction time fell with caffeine in [1], [5], [6], [7], [8] and [9], at 6 mg/kg in [4], and vigilance speed was maintained or improved in [11]. Self-rated alertness rose in [6] and [8], and PVT-indexed sustained attention improved in [3] and [11]. Doses ranged from 80 mg ([8]) to 6 mg/kg ([4],[9]) across e-gamers, taekwondo and CrossFit athletes, general young and older adults, and Special Forces personnel, giving broad convergence. The replication across independent labs and populations is what anchors the direction of the effect. The contested point: The division is over the alertness half and whether 'improves' holds unconditionally. [5] finds caffeine does not raise mental alertness in non-low consumers and attributes the apparent benefit in habitual users to reversal of overnight withdrawal, while [4] makes the reaction-time gain dose-contingent and [10] reports a non-significant reaction-time result.
All 11 papers behind this claim with their DOIs and what each does to it, the panel’s arbitration, the evidence gap map and the open research questions. Free, and sent to your inbox.
We’ll also send you other claims from Caffeine and Performance, one at a time, and tell you when a release or a pivotal trial changes a grade. Unsubscribe by replying.
One reviewer's only job is to attack the finding. The challenges below were weighed before the grade was set.
The alertness claim is undercut by [5], which shows caffeine does not enhance mental alertness in non-low (non-habitual) consumers and that the apparent benefit in habitual users may reflect reversal of overnight withdrawal rather than net cognitive enhancement.
[10] reports no statistically significant difference in reaction time between any caffeine dose and placebo — an in-corpus null that contradicts the unqualified 'improves'.
Where it remains uncertain. The alertness benefit was largely absent in non-habitual consumers and may reflect withdrawal reversal ([5]), the reaction-time gain appeared only at 6 mg/kg in [4], and [10] found its reaction-time changes non-significant.
Aleatory
Several trials have small N (e.g. N=15 in [1] and [9], N=18 in [7], N=22 in [8], N=20 in [11]), so individual response varies around a consistent mean direction.
Epistemic
[5] shows the alertness effect depends on habitual intake and overnight abstinence, and the corpus does not resolve how much of the alertness signal is net enhancement versus withdrawal reversal.
Inferential
Alertness in [6] and [8] rests on subjective ratings, so inferring an objective cognitive gain from the rating is a step the corpus does not fully close.
The load-bearing papers from a corpus of 11, each classified by what it does to the claim. Every DOI is taken from the corpus the claim was graded on.
Primary RCT evidence of reduced simple reaction time after 3 mg/kg caffeine.
“In comparison to the placebo, caffeine decreased simple reaction time (0.20 ± 0.01 vs. 0.19 ± 0.01 s, P < 0.01)”
Panel auditInternal validity: medium · scope of inference: Simple reaction time in young professional e-gamers at a single 3 mg/kg dose.Key limitations: very small sample · specialized population · single dose
Dose-resolved RT improvement, establishing a 6 mg/kg threshold.
“Reaction time only improved following 6 mg/kg of caffeine intake (Physical reaction time: -0.04 s, 95% CI -0.08 to -0.01 s, P = 0.036, d = 0.5; Motor reaction time: -0.04 s, 95% CI -0.07 to -0.01 s, P = 0.008, d = 0.6) compared to the placebo condition.”
Panel auditInternal validity: medium · scope of inference: Reaction time in healthy physically active young adults; effect dose-dependent (6 mg/kg only).Key limitations: effect only at higher dose · small absolute effect · attention not improved
Confirms reaction-time benefit across habitual-use strata while qualifying the alertness claim.
“Caffeine enhanced physical performance (faster tapping speed and faster simple and choice reaction times) in both medium-high and non-low consumers.”
Panel auditInternal validity: high · scope of inference: Reaction time and mental alertness in habitual consumers; the paper frames benefits partly as reversal of overnight withdrawal.Key limitations: benefits may reflect withdrawal reversal · alertness gain offset by anxiety in non-low consumers · no significance values in abstract
Only within-group pre-post change and an energy-drink-vs-caffeine comparison are reported; with no non-caffeine arm the pre-post PVT gain cannot be attributed to caffeine over practice or expectancy.
Panel auditInternal validity: medium · scope of inference: Comparison of caffeine delivery formats (energy drink vs plain caffeine), not caffeine vs no caffeine.Key limitations: no non-caffeine control arm · within-group pre-post only · direction stated in conclusion not isolable from design
4 of 11 papers shown. The full review lists all of them, with the panel’s arbitration →
Claims are graded at each quarterly release, and re-graded between releases only when a pivotal trial lands. Every change is a new version of the claim.
Convergent primary RCT evidence places the claim at the Established floor, held down rather than higher by the habituation confound and a dose-threshold and one non-significant result.
Release 2026-q4 re-graded every claim from its corpus under protocol v0.19. An earlier release's grade is shown for the record; it came from an earlier protocol and is not a movement.
Assumptions that carry weight but are not themselves established.
The panel records the gap that would most move the claim, and the study that would close it.
Habituation-controlled alertness design. The alertness half rests heavily on [5], which finds the benefit largely absent once habitual use and overnight abstinence are controlled. The panel rates this the highest-importance open gap for the claim, and names the study design that would close it: A trial stratifying by habitual intake with a withdrawal-controlled baseline, measuring alertness in both naive and habitual consumers.
Open question: How much of caffeine's alertness effect is net enhancement versus reversal of overnight withdrawal? Why it is still open: [5] is the only trial that stratifies by habitual use and finds no mental-alertness benefit in non-low consumers.
Priors is a sceptical first reader, not peer review, and not medical advice.
Release 2026-q4 · graded 21 Oct 2026 · protocol v0.19