Marius ComperStay-Awake Brain

Public education, not medical advice; for persistent insomnia or daytime sleepiness, see a doctor.

This tool never learns “you”; it only repeats group-trial results: group averages, no individual calibration.

Method and sources

← Back to the hub

Method and sources (D1–D10 + D13)

D1 · Wüst 2024 (S-Wuest-fulltext, PMID 38759474)

Sleep Med Rev 76:101940. Full abstract verified (eutils): k=44; one night at 2–6 h sleep opportunity; subjective sleepiness SMD=0.986; sustained attention (10-min tasks) with reaction times SMD=0.512 and lapses SMD=0.489; no significant effects on choice reaction time, cognitive throughput, working memory, or inhibitory control; duration links non-significant (sleepiness β=-0.214, ns at 0.01; attention ps>0.090). Inherited unsourced figures and MRI-stratification wordings were deleted; the slope card was contradicted by the abstract and reframed as effects. Full-text reading is optional.

D2 · Reichert 2016 summary page (S-Reichert-page, id=1358)

Verified 7.7KB fetch: adenosine accumulation → the sleepiness signal, mechanism framing only. The A1 paper's specific claims are TO-VERIFY, out of the core until verified.

D3 · Pierce review (S-Pierce-record)

Record verified: the sole basis of drowsy-driving warnings, qualitative wording, no numeric anchors.

D4 · Drake 2013 timing trial (S-Drake-abs, PMID 24235903 + S-Drake-COI)

J Clin Sleep Med 9(11):1195-200. Abstract verified: fixed 400 mg dose vs placebo at 0/3/6 h before bed, at home, self-report + portable monitor; all three timings significantly disturb sleep (p<0.05); the 6-hour group with “important disruptive effects”. The dose TO-VERIFY closed at abstract level. “Over an hour” for the 6-hour group comes from secondary records (AASM/medicalxpress/sciencedaily), marked on the card. Single cohort + actigraphy reading; conflict: Zeo Inc investigator-initiated grant + multiple author pharma ties (funding page verified), noted on the card.

D5 · Taiwan labeling rules (S-TFDA-quote)

FDA bulletin 药物食品安全週報 no. 1071 (2026/03/27, page text ?id=1612&tid=5578 verified): liquid drinks ≥20mg/100mL exact label / <20mg “20mg/100mL以下” / coffee-tea-cocoa ≤2mg “低咖啡因”; instant sachets with the per-serving total. <20 shows ≤ ceiling (MAX, not exact).

D6 · Williamson 2000 (S-Williamson-abs, PMID 10984335)

Occup Environ Med 57:649-55. Abstract verified: n=39 (30 transport + 9 army); 28 h awake vs alcohol to ~0.10%; 17–19 h (clock 22:30/01:00, at night) some tests ≥0.05%; longer → the maximum 0.10% dose (no fixed hour anchor). Build-time correction: the brief's PMID proved to be an unrelated paper (a growth-hormone topic); the correct one is 10984335.

D7 · Arnedt 2001 (S-Arnedt-abs, PMID 11235795)

Accid Anal Prev 33:337-44. Abstract verified: n=18 (males 19–35); 30 minutes of simulated driving, BAC 0.00/0.05/0.08%; lane position + speed deviation + off-road events. The wakefulness arm ran a whole night. On the three tracking measures, 18.5/21 h ≈ 0.05/0.08%; on speed deviation and off-road events alcohol beats wakefulness; traffic-free simulation. Hour anchors come from the abstract (gate iv closed at the abstract, full text optional).

D8 · Lamond 1999 (S-Lamond-abs, PMID 10646165)

J Sleep Res 8:255-62. Abstract verified: N=22 (ages 19–26); 28 h awake vs alcohol/placebo to 0.10%; 4 standardized tasks; falls with BAC and with hours (task-dependent). A within-one-trial partial trend only: the abstract gives no test clock times, so no BAC equivalence anchors (P1).

D9 · Dawson & Reid 1997 (S-Dawson-record, PMID 9230429)

Nature 388:235 (single-page letter, no abstract): record verified (Nature, 1997-07-17, vol. 388, p. 235). Tracking/regression details await full text (gate ii open): the card stays a “record”, with no hour anchors.

D10 · FDA caffeine consumer page (S-FDA-record)

Fetched and checked 2026-09-18 (46KB, 200 OK): the 400 mg/day line confirmed (“for most adults… not generally associated with negative effects”, authorized paraphrase). The current page's trying/pregnant/breastfeeding section states no figure, only advises seeing a doctor — this calculator sets no pregnancy number line (the brief's pregnancy line was not set, the current page states no figure; evidence over instruction, as with the Wüst slope precedent).

D13 · Method note

Brief-level comparative reading: TO-VERIFY, unclosed at build; this page does not rely on it.

Deleted (never cited)

D11 Taiwan insomnia page, D12 lower study: no stable sources, deleted in rev.3, the safety strip permanently number-free. The microsleep-duration claim: source blocked, deleted. The rev.1 “600mg night-shift alertness” wording: unsourced, deleted. The “18-study count”: replaced with the named inventory. GWR/amyloid/A1 specifics/glymphatic/insomnia analysis: optional editorial content, gates unpassed, outside the core.

Calculator honesty

The line is an authorized anchor (FDA paraphrase), the product milligrams are user-entered / label values, and their comparison is derived, not measured. The TW <20 tier shows ≤ ceiling.

After a sleepless night don't push it; leave big decisions until you've slept.