Ginkgo biloba extract produces small, inconsistent effects on cognition and peripheral circulation, and it does not prevent dementia in healthy older adults. The best-designed trials show changes too modest to notice in daily life, and the most rigorous prevention studies were negative. It is a biologically active plant extract with real pharmacology, real drug interactions, and a evidence base that is thinner than its marketing suggests.

How the extract actually acts on the body

Ginkgo leaf contains two pharmacologically distinct families of compounds. Flavone glycosides dominate the extract by mass. Terpene lactones, specifically ginkgolides A, B, C and the bilobalides, are present in small amounts but carry much of the biological punch. Standardized extracts such as EGb 761 are defined by both fractions, not by the herb's total weight. The label ratio "50:1" describes how much dried leaf was concentrated, not how much active compound you swallow.

The most concrete mechanism is platelet-activating factor antagonism. Ginkgolide B binds the platelet-activating factor receptor at nanomolar affinity, blocking a lipid signal that drives platelet aggregation and inflammatory cell recruitment. This is not a vague "blood-thinning" claim. It is a specific receptor interaction, and it explains both the modest circulation effects and the bleeding risk.

A second mechanism involves mitochondrial and vascular effects. In animal and cell models, ginkgo fractions improve endothelial nitric oxide availability and reduce oxidative stress in neurons. Human data are weaker. The blood-brain barrier permeability of ginkgolides is low, and brain concentrations after oral dosing are estimated in the low nanomolar range. That matters: a mechanism that works in a dish may barely reach the tissue that matters.

A third pathway is microcirculation. Ginkgo extracts measurably increase forearm blood flow and reduce blood viscosity in small trials. The effect size, roughly a 10 to 15 percent change in flow-mediated measures at typical doses, is real but small. It is the kind of shift that shows up on a vascular ultrasound, not in how you feel walking up stairs.

Plain language: ginkgo does have a defined biochemical target, but the dose that reaches the brain and the size of the downstream effect are both modest.

What the trials show, in numbers

The most-cited cognitive trial is the Ginkgo Evaluation of Memory study, published in JAMA in 2008. It randomized 3,069 adults aged 75 and older to 240 mg per day of EGb 761 or placebo for a median of 6.1 years. The primary outcome, incident dementia, was not reduced. The hazard ratio was approximately 1.12, with a confidence interval spanning 0.94 to 1.33. In plain terms: no benefit, and possibly a slight excess that was not statistically reliable.

The GuidAge trial, published in 2012, followed 2,854 adults aged 70 and older in France for 5 years using 240 mg per day. It also failed to prevent dementia. A 2012 Cochrane review pooling these and smaller trials found no consistent cognitive benefit in healthy adults, and small, inconsistent effects in people with mild cognitive impairment.

Where ginkgo performs best is in small, short trials of peripheral circulation and tinnitus. Typical studies enroll 40 to 120 participants, run 12 to 24 weeks, and report improvements of a few points on symptom scales. That is a signal worth studying, not proof of clinical benefit. A 2013 systematic review of tinnitus trials concluded the evidence was insufficient to recommend ginkgo, with most studies at high risk of bias.

Dose and standardization matter. The trials that showed any effect used 120 to 240 mg per day of a standardized extract, not raw leaf powder. A 500 mg capsule of unstandardized leaf is not equivalent to 120 mg of EGb 761. This distinction explains a large share of the confusing internet reports.

Where the evidence is genuinely thin

Two populations deserve separate treatment. People with mild cognitive impairment sit between normal aging and dementia, and they are the group most likely to buy ginkgo. The GEM study included a subgroup with mild cognitive impairment and found no reduction in progression to dementia. The 2012 Cochrane analysis reached the same conclusion. Buying ginkgo to slow memory decline is spending money on an unproven strategy.

People taking anticoagulants or antiplatelets are a safety edge case, not a cognitive one. Because ginkgolide B blocks platelet-activating factor, ginkgo can add to the effect of warfarin, aspirin, or clopidogrel. Case reports describe bleeding events, including intracranial hemorrhage in a small number of patients. These are rare and confounded, but the mechanism is plausible and the combination is avoidable.

Counterintuitive detail: ginkgo's most consistent human effect is not on memory at all. It is a modest increase in bleeding time and a small reduction in platelet aggregation. The herb is more reliably a mild antiplatelet agent than a cognitive enhancer. Anyone taking it for "brain health" is, physiologically, taking a weak blood thinner.

A professional caveat: the supplement market is poorly standardized. Independent analyses of ginkgo products have found extracts with terpene lactone content far below label claims, and occasional adulteration with other species. A 2016 analysis of 24 ginkgo products found that only a minority met the labeled flavonoid and lactone specifications. You cannot assume a bottle delivers the trial dose.

Common misconceptions, corrected

The first misconception is that "natural" means "safe with everything." Ginkgo is a pharmacologically active extract with a documented interaction profile. It is not a vitamin.

The second is that ginkgo is proven for dementia prevention. It is not. The two largest trials, at 240 mg per day for 5 to 6 years, were negative. That is the strongest evidence available, and it points the other way.

The third is that ginkgo cures tinnitus. The best reviews say evidence is insufficient. Symptom-scale improvements in small trials do not translate into a reliable treatment.

The fourth is that more is better. Doses above 240 mg per day have not shown added cognitive benefit in trials, and higher exposure increases bleeding risk without a demonstrated upside.

The fifth is that all ginkgo products are equivalent. Standardization matters. A product without verified terpene lactone content is not the same chemical entity the trials used.

Comparison with common alternatives

Option Best evidence Typical dose in trials Key caveat
Ginkgo extract (EGb 761 type) Negative for dementia prevention; small effects on circulation 120–240 mg/day standardized Bleeding risk; product quality varies
Aerobic exercise Moderate evidence for cognition and mood 150 min/week moderate intensity Requires sustained behavior change
Mediterranean-style diet Observational and trial support for cardiovascular risk Food pattern, not a pill Slow, indirect cognitive effects
Omega-3 supplements Mixed; no clear cognitive benefit in healthy adults 1–2 g/day in trials Fishy aftertaste; bleeding at high doses
Cholinesterase inhibitors Modest symptomatic benefit in diagnosed Alzheimer's Prescription only Gastrointestinal side effects; prescription required
Placebo Zero by definition Not applicable Honest baseline for comparison

The table's message is straightforward. Ginkgo is one option among several, and it is not the best-supported one for cognition. Exercise and diet have stronger evidence for brain health, though neither is a quick fix. Prescription drugs have a defined role in diagnosed disease, not in healthy adults.

What to do if you are considering it

Start by defining the goal. If the goal is dementia prevention, ginkgo is the wrong tool, and the GEM and GuidAge results are the reason. If the goal is a mild circulatory or tinnitus symptom, discuss it with a clinician before buying, because the evidence is weak enough that a trial period should have a defined endpoint.

Check the label for standardization. Look for a stated extract name, a defined flavonoid glycoside percentage, and a terpene lactone percentage, typically around 24 percent and 6 percent respectively in EGb 761-type products. If the label does not specify these, treat the product as unverified.

Set a stop rule in advance. A reasonable trial is 8 to 12 weeks at 120 to 240 mg per day of a standardized extract, with a symptom log. If nothing measurable changes, stop. That is a fair test of the evidence, not a permanent subscription.

Review your medication list. If you take warfarin, aspirin, clopidogrel, or any anticoagulant, do not start ginkgo without medical advice. The same applies before surgery, where ginkgo should be stopped at least two weeks beforehand, consistent with general guidance on antiplatelet supplements.

Red-flag warnings:

Stop and seek care for unusual bruising, nosebleeds that will not stop, blood in urine or stool, or severe headache while taking ginkgo. Do not combine ginkgo with anticoagulants or antiplatelets without a clinician's review. Do not use ginkgo as a substitute for evaluation of new memory loss, which needs a diagnostic workup. Treat unstandardized products and multi-herb "brain blends" as unknowns, not as equivalents to trial extracts.

Reconciling the thresholds: the 240 mg per day dose in the large negative trials is the same dose used in many positive circulation trials. So the dose is not the explanation for the disagreement. The explanation is endpoint and population. Prevention trials in healthy older adults measure dementia incidence over years, while circulation trials measure symptom scales over weeks. When you compare the two, the negative prevention data are more relevant to the question most people are actually asking.

Limits of what can be concluded

The evidence base has real holes. Most positive ginkgo trials are small, short, and industry-adjacent, with high risk of bias. The large trials were better designed but tested a narrow prevention question in older adults. There is little long-term data on cognitive outcomes in midlife users, and almost no trial data on the popular combination products sold today.

Product standardization is a second limit. Even a well-designed trial cannot tell you whether the bottle on your shelf matches the extract that was tested. Without independent verification, the dose on the label is a claim, not a measurement.

The honest summary is bounded. Ginkgo has defined pharmacology, small measurable vascular effects, and no demonstrated ability to prevent dementia. It carries a real bleeding interaction and a real product-quality problem. For most people asking "does it really work," the answer is that it works a little, on the wrong outcomes, with more risk than the marketing implies. That is not a reason to ban it. It is a reason to stop treating it as a proven brain supplement and start treating it as what it is: a modestly active plant extract with narrow, unproven uses.