Curcumin for Cognitive Function
Curcumin shows only weak, inconsistent evidence for supporting cognitive function: any benefit appears limited to specific domains like working memory in older adults using high-bioavailability formulations, and major health authorities agree the evidence is too thin to recommend it for memory or brain health.
Why this grade7-layer evidence engine
The grade is held to weak (C) because the strongest evidence directly contradicts itself. Two 2025 meta-analyses split on the core outcome of global cognition: Wang 2025 (PMID 40308636, 9 RCTs, n=501) found a significant benefit (SMD 0.82, 95% CI 0.19-1.45), but a cognitive-aging meta-analysis the same year (PMID 40579315, n=531) found no global effect (SMD 0.14, 95% CI -0.78 to 1.07). With one positive and one null high-tier source, neither dominates.
What signal exists is narrow and conditional. The null meta-analysis still found benefits for working memory (SMD 1.01) and processing speed (SMD 0.37), echoed by small RCTs in older adults using bioavailable formulations (Small 2018, PMID 29246725, Theracurmin; Cox 2020, PMID 32512782, Longvida working memory). Effects appear only at sufficient duration, age 60+, and with specialized formulations, not generic supplements or younger users. Notably, three of five studies were funded by formulation manufacturers, flagging industry-bias risk.
Health authorities reinforce caution rather than endorsement. NIH's NCCIH states plainly, 'We don't know enough to definitively conclude if turmeric or curcumin is beneficial for any health purposes,' and Cleveland Clinic says 'more research is needed,' while Mayo Clinic and Harvard Health do not address cognition at all. EFSA rejected curcumin health claims and set a 3 mg/kg/day intake limit, and NHS Wales warns against use with liver or biliary disease, given documented liver-injury reports.
Scoring transparency
All scores computed by a 7-layer evidence engine — fully auditable▸View the full decision path (audit trail)
- compute_raw_score — 加權公式: L2×0.30 + L3×0.25 + L5×0.25 + L11×0.10 + L1×0.10 = 0.446
- tier_from_score — 依分數區間映射至 tier letter
- apply_hec_rules — 高階證據未達主導 (1 positive vs 1 negative)，由 raw_score 決定
- tier_strict_requirement_check — Tier 條件達標，未降階
- detect_disputes — 偵測到 0 個 hard + 1 個 soft dispute
- decide_status — 依 tier + dispute 結果決定 status