10 Plant Compounds, One Ranking: What a New Network Meta-Analysis Found on Cognition
A January 2026 network meta-analysis pooled 25 RCTs and 1,861 healthy older adults to rank bacopa, ginkgo, guarana, curcumin, raisin, tart cherry, and more head-to-head across five cognitive domains.
This article is for educational and research purposes only. It does not constitute medical advice. Consult a qualified healthcare professional before making any health-related decisions.
Comparing botanicals head-to-head, not one at a time
Most of the evidence for plant-derived cognitive compounds comes compound by compound: one review on bacopa, one on ginkgo, one on curcumin, each drawing on a different and only partly overlapping set of trials. That structure makes it hard to answer a question people actually ask, which is how these options compare against each other rather than against placebo alone. A network meta-analysis published online on 20 January 2026 in Frontiers in Pharmacology set out to close that gap. The authors, Xueyan Feng, Shuangbo Fan, and Fengqin Wei, pooled 25 randomized controlled trials covering 1,861 healthy older adults and 10 plant active substances, then used network meta-analysis to rank them against each other across five cognitive domains (Feng X, Fan S, Wei F. "The effect of plant active substances on cognitive function in healthy older adults: a systematic review and network meta-analysis of randomized controlled trials." Front Pharmacol 2026;16:1672171. DOI: 10.3389/fphar.2025.1672171. PMID: 41640686).
What was actually compared
The review was prospectively registered with PROSPERO (CRD420251032046) and searched Embase, PubMed, the Cochrane Library, and Web of Science from each database's inception through 31 December 2024. Eligible trials were randomized and placebo- or active-controlled, enrolled adults aged 50 and over with no dementia, Alzheimer's disease, mild cognitive impairment, or other neurological or psychiatric diagnosis, and reported at least one standardized cognitive outcome.
The 10 substances were bacopa monnieri, ginkgo biloba, guarana, salvia, caffeine, wild green oat (Avena sativa), curcumin, blueberry, tart cherry, and raisin or grape extract. Several trials tested combination products rather than single ingredients, including a bacopa compound stacked with lycopene, astaxanthin, and vitamin B12, and a ginkgo combination with Gotu kola and DHA. Doses varied widely even within a single substance: trials of ginkgo biloba ranged from 120 mg per day (Nathan 2002; Solomon 2002) to 180 mg per day (Mix 2000; Mix 2002), while bacopa trials spanned 300 mg per day (Calabrese 2008; Morgan 2010) up to 4.32 g per day in one trial (McPhee 2021), a dose that sits well above the 300 to 450 mg per day range typically used in the standalone bacopa clinical trial literature.
Results by cognitive domain
Learning and memory drew the largest evidence base, 23 of the 25 trials. Raisin or grape extract ranked first (SUCRA 95.1%) and tart cherry second (SUCRA 89.5%), with mean differences over placebo of 1.09 (95% CI 0.46 to 1.71) for raisin and 0.96 (95% CI 0.13 to 1.78) for tart cherry. Bacopa also beat placebo on this domain, with a smaller mean difference of 0.49 (95% CI 0.07 to 0.90). The authors note the raisin-versus-tart-cherry difference itself was not statistically significant, so the two polyphenol-rich fruit interventions are better read as a tied top tier than as one clearly superior option.
Complex attention, assessed in 18 trials, showed essentially nothing. The top-ranked intervention, a bacopa compound, had a SUCRA of only 63.9% and did not separate from placebo at a level the authors considered meaningful. Across this entire domain, none of the 10 substances produced a statistically convincing benefit over placebo.
Executive function, covered in 16 trials, was the domain with the clearest multi-substance signal. A bacopa compound ranked first (SUCRA 91.3%, mean difference 1.28, 95% CI 0.78 to 1.78), followed closely by curcumin (SUCRA 89.3%, mean difference 1.12, 95% CI 0.72 to 1.70) and tart cherry (SUCRA 88.9%, mean difference 1.21, 95% CI 0.47 to 1.96). Standalone ginkgo also reached statistical significance here, though with a far smaller effect (mean difference 0.30, 95% CI 0.01 to 0.59).
Language (4 trials) and perceptual-motor function (3 trials) were assessed in too few studies to support firm conclusions. The authors flag these as exploratory: a bacopa compound and raisin ranked highest for language, and guarana ranked highest for perceptual-motor speed, edging out plain caffeine in one direct comparison (mean difference 1.09, 95% CI 0.35 to 1.85). Guarana's natural caffeine content means this result sits alongside, rather than instead of, the better-established evidence for caffeine's acute cognitive effects, and the authors note the comparison was not designed to separate guarana's other constituents from its caffeine content.
No single winner, and real limits on the evidence
The headline finding is not "X is the best nootropic." It is that different plant compounds appear to act on different cognitive domains, and none of the 10 substances tested showed a benefit across all five. Raisin and tart cherry led on memory; a bacopa compound, curcumin, and tart cherry led on executive function; nothing reliably moved complex attention; and the language and perceptual-motor results rest on too few trials to generalize.
The authors are candid about evidence quality. Of the 25 trials, 6 were rated low risk of bias and 4 high risk, with the remaining 15 not clearly reporting how allocation was concealed, which the authors score as an overall moderate level of certainty. Trial durations, doses, extraction methods, and the specific cognitive tests used varied considerably across studies, which complicates direct comparison even within the network meta-analysis framework. Only 8 of the 25 trials reported on adverse events at all, and those that did described mostly mild gastrointestinal complaints with no serious toxicity signals. The authors specifically flag a theoretical bleeding risk when ginkgo is combined with anticoagulant medication, consistent with ginkgo's known effect on platelet aggregation, and recommend consulting a clinician before combining it with blood-thinning drugs. Visual inspection of funnel plots did not suggest obvious publication bias, though the authors caution this is a qualitative assessment with a relatively small number of trials per domain.
What this means in context
This network meta-analysis does not overturn the single-ingredient literature on sites like this one, it supplements it with a head-to-head ranking that most individual trials cannot provide. Bacopa's showing here, a modest but real effect on memory and a notably larger one on executive function through a multi-ingredient compound, is broadly consistent with the picture from dedicated bacopa trials: real but modest effects that take weeks of consistent use to appear, not an acute cognitive boost. The strongest signals in this analysis, for raisin, tart cherry, and curcumin, come from fewer prior systematic reviews on this site and from a smaller evidence base overall than bacopa or ginkgo individually carry, which is exactly why the authors call for dosage-optimization and longer-duration trials before treating any of these rankings as settled.
Key takeaways
A January 2026 network meta-analysis of 25 RCTs in 1,861 healthy older adults ranked 10 plant active substances across five cognitive domains rather than testing each against placebo in isolation. Raisin and tart cherry ranked highest for learning and memory; a bacopa-containing compound, curcumin, and tart cherry ranked highest for executive function; nothing tested reliably improved complex attention; and language and perceptual-motor findings rest on too few trials to generalize. Evidence quality was rated moderate overall, with mixed risk-of-bias ratings and adverse-event reporting in only a third of trials. The authors call for longer, better-dosed, better-designed trials before any of these substances can be considered a validated cognitive intervention for healthy older adults.
FAQ
Does this mean raisins or tart cherry are proven to improve memory?
Not proven in a strong sense. They ranked highest among the 10 substances tested for learning and memory in this specific pool of trials, with statistically significant improvements over placebo. But the overall evidence was rated moderate quality, individual trials varied in design and dosing, and the authors explicitly call for further research before treating the ranking as a settled conclusion.
Does this change what is known about bacopa monnieri?
It adds to it rather than overturning it. A bacopa-containing compound ranked at or near the top for executive function and also beat placebo on memory in this analysis, consistent with the modest, consolidation-focused benefits documented in dedicated bacopa trials. The doses used in the trials feeding this meta-analysis ranged more widely than the 300 to 450 mg per day typically studied for bacopa on its own, so the two bodies of evidence are not directly interchangeable.
Why did ginkgo biloba rank lower than expected?
Standalone ginkgo reached statistical significance only for executive function, and with a small effect size (mean difference 0.30) compared to the top-ranked interventions in that domain. It did not separate clearly from placebo on memory or complex attention in this pooled analysis. The authors also flag a theoretical bleeding risk when ginkgo is combined with anticoagulant medication.
Is this evidence strong enough to act on?
The authors themselves rate it moderate-certainty and call for future trials to address long-term safety, dose optimization, and whether combining substances produces synergistic effects beyond what any single one achieves. Three of the five cognitive domains examined, language, perceptual-motor speed, and complex attention, were covered by too few trials to support confident conclusions either way.
Primary source: Feng X, Fan S, Wei F. "The effect of plant active substances on cognitive function in healthy older adults: a systematic review and network meta-analysis of randomized controlled trials." Front Pharmacol. 2026;16:1672171. DOI: 10.3389/fphar.2025.1672171. PMID: 41640686. PMCID: PMC12864429. PROSPERO registration: CRD420251032046. This article summarizes published, peer-reviewed research and is not a substitute for individualized medical advice.
Disclaimer
This article is for educational and research-context purposes only. It does not constitute medical advice and is not a substitute for consultation with a qualified healthcare practitioner. The network meta-analysis discussed reflects published, peer-reviewed findings in healthy older adults, not personalized clinical guidance, and does not establish efficacy or safety in people with existing cognitive impairment, other medical conditions, or those taking prescription medications. Always consult a doctor before beginning, modifying, or stopping any supplement.