Melatonin for Cognitive Function

Verdict: Weak, disputed; not proven to aid cognition

Melatonin is not a reliable way to improve cognition. A few small meta-analyses hint at a modest benefit in Alzheimer's and mild cognitive impairment, but the signal is weak, contested, and not endorsed by any regulator or medical society for this use.

C 🟠 C Weak Evidence Disputed

🔬Why this grade7-layer evidence engine

The grade is Weak (Tier C, disputed) because the evidence points in different directions depending on who is treated. Three meta-analyses in cognitively impaired adults found small MMSE gains -- MD 1.08 overall and 2.63 in mild cognitive impairment (PMID 41185054), MD 1.48 with low-dose treatment in Alzheimer's dementia (PMID 35450525), and MD 1.82 in Alzheimer's patients treated over 12 weeks (PMID 33957167). These improvements are real but clinically marginal (roughly 1-2 MMSE points) and drawn from small trials.

The case weakens further outside dementia. Two earlier meta-analyses suggested melatonin cut ICU delirium (RR 0.72-0.75; PMID 40662882, PMID 41602948), but the larger 2025 double-blind DEMEL trial found no benefit at all (RR 0.986; PMID 40608082), undercutting that claim. In healthy adults, daytime melatonin actually reduced reaction accuracy (SMD -0.74; PMID 33957167) -- a harm signal, not a nootropic effect.

Regulators and clinics reinforce the cautious read. The FDA sees 'no basis to conclude that melatonin is GRAS,' and EFSA's only approved claims concern sleep onset, not cognition; the UK NHS treats it as prescription-only. Mayo Clinic says it 'doesn't seem to improve cognition,' Cleveland Clinic calls it possibly inappropriate for dementia patients, and the Alzheimer's Association states no supplement is proven to help -- so no authority backs melatonin for cognitive benefit.

⚖️

Scoring transparency

All scores computed by a 7-layer evidence engine — fully auditable
Raw score 0.42
D
C
B
A
S
← counter-evidence / ineffectiveeffective / strong evidence →
Final grade
C · Disputed
Confidence
77%
Broadly consistent
Evidence level
E2
Multiple high-quality MAs (≥2 independent, consistent)

How strongly each layer supports this effect

lower = less supportive
L5 Clinical bodiesAuthoritative stance
0.28
L2 PubMedPrimary literature
0.45
L3 MechanismPlausibility
0.45
L1 ExamineGlobal benchmark
0.50
L11 AI re-checkIndependent read
0.50
Against Mixed Supports
View the full decision path (audit trail)
  1. compute_raw_score — 加權公式: L2×0.30 + L3×0.25 + L5×0.25 + L11×0.10 + L1×0.10 = 0.416
  2. tier_from_score — 依分數區間映射至 tier letter
  3. apply_hec_rules — 高品質 SR/MA 顯示 positive (1 篇 > 0 negative)
  4. tier_strict_requirement_check — Tier 條件達標,未降階
  5. detect_disputes — 偵測到 1 個 hard + 0 個 soft dispute
  6. decide_status — 依 tier + dispute 結果決定 status

📄PubMed studies (6)L2 · primary research & systematic reviews

Effect of melatonin on cognitive function in adults with cognitive impairment: a multi-dimensional meta-analysis of randomized trials
PMID: 41185054 2025 統合分析 n = 518
Finding: Melatonin improved MMSE (MD 1.08, p<0.0001); subgroup with mild cognitive impairment MD 2.63 (p<0.000001)
Academic Effect size: [object Object]
View on PubMed
The Dose and Duration-dependent Association between Melatonin Treatment and Overall Cognition in Alzheimer's Dementia: A Network Meta-Analysis of RCTs
PMID: 35450525 2022 統合分析
Finding: Medium-term low-dose melatonin had highest post-treatment MMSE (MD 1.48, 95% CI 0.51-2.46) among all compared treatments
Effect size: [object Object]
View on PubMed
Neurocognitive effects of melatonin treatment in healthy adults and individuals with Alzheimer's disease and insomnia: A SR & MA of RCTs
PMID: 33957167 2021 統合分析
Finding: AD >12 weeks: MMSE MD 1.82 (1.01-2.63, p<0.0001); mild AD MD 1.89 (p<0.0001); healthy daytime use decreased accuracy SMD -0.74 (p<0.00001)
Effect size: [object Object]
View on PubMed
Melatonin supplementation reduces delirium incidence in critically ill patients: a systematic review and meta-analysis
PMID: 41602948 2026 統合分析 n = 3,706
Finding: Melatonin reduced delirium incidence RR 0.75 (95% CI 0.63-0.90, p=0.001); no effect on duration, LOS or mortality
Effect size: [object Object]
View on PubMed
Melatonin Use in the ICU: A Systematic Review and Meta-Analysis
PMID: 40662882 2025 統合分析 n = 3,895
Finding: May reduce delirium RR 0.72 (95% CI 0.58-0.89, low certainty); may slightly reduce ICU LOS; may improve perceived sleep quality
Effect size: [object Object]
View on PubMed
Melatonin for prevention of delirium in patients receiving mechanical ventilation in the ICU: a multiarm multistage adaptive RCT (DEMEL)
PMID: 40608082 2025 RCT (double-blind) n = 334
Finding: No difference: low-dose 54.4% vs placebo 55.2% (RR 0.986, 95% CI 0.803-1.211); 0.3 mg had better PK profile but no clinical benefit
🟢 High quality Mixed funding Effect size: [object Object]
View on PubMed

🏛️Regulatory & authoritative positionsL4/L5 · FDA / EMA / NIH ODS / Cochrane / Mayo …

L4a US FDA
Cautious
no basis to conclude that melatonin is GRAS source↗
L4b EU EFSA
Supportive
melatonin and reduction of sleep onset latency source↗
L4c UK NHS
Cautious
Melatonin is available on prescription only. source↗
L4d TW TFDA / 衛福部
Against
褪黑激素產品屬藥品列管 source↗
L4e WHO
Not addressed
altered melatonin excretion source↗
L5a NIH Office of Dietary Supplements
Cautious
short-term use of melatonin supplements appears to be safe source↗
L5b Mayo Clinic
Against
doesn't seem to improve cognition source↗
L5c Cleveland Clinic
Cautious
may also not be appropriate source↗
L5e Specialty Society (condition-mapped)
Not addressed
Not a single food, beverage, ingredient, vitamin or supplement source↗
PMID 100% verifiedevery citation checked via NCBI Entrez
🔬6 PubMed studiesindependently re-checked by multiple sub-agents
engine_version: v1.0 claim_id: CLM-COND-cognitive-function-INT-melatonin-001 繁體中文版 →