Polyphenols and Memory: Why Anthocyanins Stand Out

Brain-supplement marketing treats polyphenols as one big happy family. The trials do not. The best current evidence points to a small improvement in episodic memory from certain subclasses—especially anthocyanins and possibly resveratrol—while Ginkgo, isoflavones, and cocoa-flavanol supplements have far less convincing records.

That distinction matters because drug prevention trials for cognitive decline have not delivered a reliable answer either. Available drugs such as donepezil have modest effects in established dementia; they are not proven preventive treatments for healthy aging. Polyphenols are therefore worth studying as low-risk adjuncts, not as a substitute for exercise, sleep, blood-pressure control, or medical evaluation of new memory symptoms.

The signal is real, but small and uneven

The largest overview, a 2022 systematic review and meta-analysis by de Vries et al., assessed 66 randomized controlled trials of polyphenol supplements and memory in adults age 40 and older. Forty-nine trials provided calculable effect sizes. Across that subset, supplementation produced a statistically significant standardized memory effect of about 0.24. By conventional standards, that is small—and publication-bias testing found funnel-plot asymmetry, which means the uncorrected estimate likely overstates the average benefit.

Bias correction weakened the pooled result but did not erase it. That leaves a reasonable conclusion: some effect probably exists, but it is likely smaller than the published literature initially makes it appear. The review also excluded 17 of its 66 eligible trials from quantitative pooling because raw data were unavailable, another reason to avoid treating 0.24 as a precise promise.

More useful than the overall number is where it came from. Anthocyanins and resveratrol showed statistically significant effects on episodic memory, the ability to remember events and information after a delay. Working-memory findings were weaker and less consistent. Ginkgo biloba and isoflavones did not show significant subgroup effects. A separate 2022 meta-analysis of dietary flavonoid trials also found positive but heterogeneous cognitive effects, supporting the idea that compound class matters.

For blueberries, the product and dose are part of the intervention

"Blueberry supplement" is not a scientifically meaningful dose. A 2024 randomized, controlled multi-study trial by Cheng et al. tested three wild-blueberry preparations in healthy older adults with self-reported memory complaints. After three months, the bioavailability-enhanced WBE 111 extract, containing 7 mg anthocyanins plus L-cysteine and L-glutathione, significantly improved delayed word-recognition memory.

Two crude powder products did not improve cognition: WBP 500 supplied 1.35 mg anthocyanins and WBP 1000 supplied 2.7 mg. The study cannot separate the effect of the higher anthocyanin dose from the co-formulation's possible effect on absorption. It does establish a practical point: products from the same fruit can produce different trial results, and a label that says "blueberry" tells you very little.

The dose contrast also fits trials summarized by de Vries et al. In the strongest anthocyanin trial in that review, 215 healthy adults aged 60 to 70 took 600 mg/day for 24 weeks and showed significant episodic-memory improvements on CANTAB subscales. A separate 24-week trial in 102 adults aged 40 to 60 found no cognitive effect from 90 to 150 mg/day of Aronia extract. These studies differ in source, age, and design, so they do not prove a single dose threshold. Still, they make low-dose powders a weak bet.

Memory complaints may identify the more responsive group

The benefit appears more consistent in people who already have room to improve. A 2025 blueberry-focused systematic review and meta-analysis by da Silva et al., which included a Bayesian post hoc analysis, found positive cognitive effects in elderly adults with prior cognitive decline, including mild cognitive impairment or subjective cognitive decline. Effects were smaller or absent in cognitively intact older adults.

A ceiling effect is one plausible explanation: a person already performing near the top of a cognitive test has less measurable room to improve. It is not proof that anthocyanins treat mild cognitive impairment or prevent dementia. The trials measured cognitive test outcomes over weeks or months, not rates of dementia diagnosis or long-term independence.

Blood vessels are a plausible mechanism, not a finished explanation

The leading human mechanism is cerebrovascular. In a double-blind randomized trial, Wood et al. reported in 2023 that wild-blueberry polyphenols improved vascular function and cognitive performance in healthy older adults versus placebo. Measuring both endpoints in the same trial strengthens the vascular hypothesis, but parallel changes do not prove that the vascular change caused the cognitive result.

A placebo-controlled crossover study of cocoa flavanols during severe hypoxia found better cognitive function, cerebral oxygenation, and less mental fatigue at an altitude-equivalent challenge of 5,500 meters. That is an informative proof of concept for blood-flow-sensitive cognition, not evidence that cocoa supplements prevent cognitive decline in daily life. Antioxidant pathways may also contribute: a 2025 epicatechin review and meta-analysis supports neuroprotective effects against oxidative-stress-induced cognitive impairment. Claims about BDNF, synaptic plasticity, or the gut-brain axis remain largely preclinical.

Cocoa is the cautionary example

Cocoa-flavanol findings show why a mechanism and a few positive trials do not settle a supplement question. The large COSMOS-Mind randomized trial found no cognitive benefit from 500 mg/day cocoa flavanols in a prevention setting. The CANN randomized trial likewise found no improvement in cognition or brain structure from combined DHA-rich fish oil and cocoa flavanols in older adults with memory complaints. An acute trial also found that cocoa flavanols, with or without caffeine, did not improve working memory or attention.

Those null results do not erase every favorable cocoa study, but they make a reliable cognitive benefit hard to claim. Differences in participant health, formulation, duration, and cognitive endpoints may explain some disagreement. They also fit the possibility that smaller positive trials overestimated an effect. Cocoa may alter cerebral oxygenation under severe hypoxia without improving ordinary cognitive performance months later; those are different questions.

What a rational experiment looks like

If you have subjective memory decline or mild cognitive impairment and want to test a supplement alongside standard care, anthocyanin-rich interventions have the most defensible rationale. Choose a product that discloses anthocyanin content and resembles a studied formulation as closely as possible. Do not infer efficacy from berry powder weight, a colorful capsule, or an antioxidant score. The available evidence does not identify one universally effective dose or prove that a bioavailability enhancer beats an equal anthocyanin dose without one.

• Expect, at most, a small change in episodic-memory test performance over weeks to months.
• Do not expect proven prevention of dementia or reversal of established disease.
• Skip Ginkgo and isoflavones if your goal is memory support; the largest subgroup analysis found no significant effect.
• Treat resveratrol as plausible but less settled: its 2023 meta-analysis found cognitive benefits in older adults, but high heterogeneity limits confidence in any single estimate.
• Do not buy cocoa flavanols for cognition on the basis of the current prevention evidence.

Anthocyanins merit a cautious trial for people with existing memory concerns, particularly when the product's dose and formulation are transparent. They are an adjunct with modest evidence, not a brain-insurance policy.

Finally, new or worsening memory problems deserve clinical attention. Supplements can coexist with evidence-based habits; they cannot diagnose depression, sleep apnea, medication effects, vascular disease, or neurodegenerative illness. The most honest summary is modest: anthocyanins have a credible but small episodic-memory signal, strongest in people already noticing decline. The rest of the polyphenol aisle remains much less certain.