Vitamin B12 for Cognitive Function
In people who already get enough B12, supplementing does not improve memory, thinking, or slow cognitive decline, and it is not an effective way to prevent dementia. B12 only helps the brain when it corrects a genuine deficiency, and even that benefit for cognition is uncertain.
Why this grade7-layer evidence engine
This earns a Counter-Evidence (D) grade because the strongest, most recent trials converge on no effect. Two large meta-analyses anchor the verdict: Markun 2021 (16 RCTs, 6,276 adults; PMID 33809274) found null results across every cognitive domain, and Alzahrani 2024 (9 RCTs; PMID 39655146) found a pooled SMD of -0.03 (p=0.18). The biggest single trial, B-PROOF (2,919 elderly with high homocysteine, 24 months; PMID 25391305), was also null despite lowering homocysteine.
Crucially, supplementation failed even when it corrected a real shortfall. Dangour 2015 (PMID 26135351) gave 1 mg/day for a year to older adults with moderate B12 deficiency and saw no neurologic or cognitive gain, breaking the assumption that fixing the biochemistry fixes the brain. The lone positive landmark, VITACOG (PMID 21780182), helped only a subgroup with elevated homocysteine plus adequate omega-3, and that signal was never replicated. A 2025 meta-analysis (PMID 40966571) detected a very small benefit (Hedges' g=0.11) that vanished in cognitively healthy people.
Health authorities agree. The NIH Office of Dietary Supplements states B12 has no benefit on performance absent a nutritional deficit, and the WHO does not endorse routine B12 supplementation, treating it as a remedy for diagnosed deficiency rather than a prevention strategy. Cleveland Clinic finds no evidence such supplements boost memory or cognition, while Mayo and Harvard confirm benefit is limited to deficiency, noting even 1,000 µg/day did not improve cognition in Alzheimer's patients.
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 = 0.44
- tier_from_score — 依分數區間映射至 tier letter
- apply_hec_rules — 高品質 SR/MA 顯示 negative 主導 (3 negative > 0 positive)，下層 RCT 不能推翻
- apply_hec_override — HEC-1 高階證據 negative — 強制由 C 改為 D
- tier_strict_requirement_check — Tier 條件達標，未降階
- detect_disputes — 偵測到 0 個 hard + 0 個 soft dispute
- decide_status — 依 tier + dispute 結果決定 status