- Cocoa flavanols are plant compounds (including epicatechin) concentrated in raw cocoa. The key trials used a standardized cocoa extract providing 500 mg of flavanols a day — not chocolate.
- In COSMOS-Web (3,562 older adults), the cocoa extract did not significantly improve memory in the group as a whole at one year — the trial's prespecified primary result.
- Memory did improve in participants whose usual diets were lower in quality or lower in flavanols, which the authors frame as a "depletion–repletion" effect. That's a subgroup finding, not proof for everyone.
- Chocolate is a poor delivery system: heat, roasting, and alkalizing ("Dutch processing") reduce flavanols, labels rarely state flavanol content, and you'd add sugar, fat, and calories.
- If your diet is already rich in tea, berries, apples, and cocoa, a supplement may add little. If it isn't, food changes are the first step.
What cocoa flavanols are
Flavanols are a family of plant compounds found in cocoa beans, tea, apples, berries, and grapes. Cocoa is unusually rich in them, particularly in a flavanol called epicatechin. Researchers became interested because flavanols affect the blood vessel lining — in 2015 the European Commission approved a health claim, based on an EFSA scientific opinion, that 200 mg of cocoa flavanols a day may contribute to maintaining vascular elasticity and normal blood flow.[4]
Note what that claim is about: blood vessels, not memory. The memory question needed its own trials.
The early signal: a small hippocampus study
In 2014, Columbia University researchers published a small randomized trial in Nature Neuroscience. Healthy adults aged 50 to 69 were assigned to a high- or low-cocoa-flavanol diet for three months. The high-flavanol group showed enhanced function in the dentate gyrus — a part of the hippocampus linked to age-related memory change — on specialized brain imaging, along with better performance on a related memory task.[2]
It was a short trial with a narrow memory test, and one author was employed by Mars, Inc., which also provided a grant. Promising, not definitive — which is why the larger COSMOS trials mattered.
COSMOS: the large trials
The COSMOS trial (COcoa Supplement and Multivitamin Outcomes Study) randomized 21,442 older adults — women 65 and older and men 60 and older — to a cocoa extract providing 500 mg of cocoa flavanols a day, a multivitamin, both, or placebo, with a median of 3.6 years of treatment and follow-up.[3] Several cognitive sub-studies were built into it.
COSMOS-Web: the memory result
COSMOS-Web enrolled 3,562 older adults who completed online memory testing. The prespecified primary outcome — an improvement in memory across all participants at one year — was not statistically significant.[1]
The more interesting findings came next. Among participants in the lower thirds of habitual diet quality (measured by the Alternative Healthy Eating Index) or habitual flavanol intake, the cocoa extract was associated with improved hippocampal-dependent memory. In a subset of 1,361 people with a urine flavanol biomarker, increases in the biomarker tracked with memory improvement.[1] The authors proposed a "depletion–repletion" model: flavanols may help most in people who aren't getting enough.
NOT SHOWN: Cocoa flavanols for memory in older adults overall — COSMOS-Web's primary endpoint was not met. EMERGING: Cocoa flavanols for memory in people with low-flavanol or lower-quality diets — subgroup finding that needs confirmation.When a trial misses its main target but finds an effect in a subgroup, that's a clue, not a conclusion. The finding here is biologically sensible and was supported by a biomarker, which strengthens it — but it should be confirmed in a trial designed specifically for people with low flavanol intake.
Why dark chocolate isn't a substitute
It's tempting to read "cocoa flavanols help memory" as "eat dark chocolate." Several problems get in the way:
- Processing destroys flavanols. Fermentation, drying, roasting, and especially alkalization ("Dutch processing") reduce cocoa's epicatechin and catechin content, which is why chocolate is usually relatively low in flavanols compared with raw cocoa.[5]
- Labels don't tell you. "70% cacao" describes cocoa solids, not flavanols. Two bars with the same percentage can differ widely.
- The dose comes with baggage. Getting a trial-like flavanol dose from chocolate — if it were even possible with a given bar — would add substantial sugar, saturated fat, and calories every day.
- Caffeine and theobromine. Cocoa products contain stimulants that can affect sleep for sensitive people, especially in the evening.
Enjoy dark chocolate as food. Just don't count it as a brain-health intervention.
| Factor | Standardized cocoa extract | Dark chocolate |
|---|---|---|
| Flavanol content stated | Usually, on products standardized for flavanols | Rarely |
| Matches trial dose | Possible (500 mg/day used in COSMOS) | Unreliable |
| Sugar and calories | Minimal in capsules | Significant at larger amounts |
| Processing losses | Extracts are designed to preserve flavanols | Roasting and alkalizing reduce them |
| Trial evidence for memory | Subgroup benefit in COSMOS-Web | None directly |
Food first: where else flavanols come from
If the depletion–repletion idea is right, the first move for most people is dietary: tea, apples, berries, grapes, and cocoa all contribute flavanols. That approach also improves overall diet quality — itself the factor that defined who benefited in COSMOS-Web. Our guide to foods that fight brain aging covers the broader pattern.
A cocoa-extract supplement is a reasonable consideration for someone whose diet is low in these foods and who wants a standardized dose. It's less likely to add much for someone who already drinks tea daily and eats fruit regularly.
Safety and who should be careful
Cocoa extracts at trial doses were taken by thousands of older adults in COSMOS. Still, consider:
- Caffeine and theobromine sensitivity — take earlier in the day if sleep is an issue.
- Medications and conditions — check with a pharmacist if you're sensitive to stimulants, take several prescriptions, or have heart rhythm concerns.
- Product quality — supplements aren't approved for effectiveness before sale. A third-party seal (USP, NSF) means an independent lab verified contents and screened contaminants; it doesn't prove a product works. Confirm the seal for the exact product.
For an honest framework on single ingredients versus multi-ingredient formulas, see nootropic blends vs single-ingredient supplements.
What to actually buy
If you want to try a standardized extract, compare products by stated cocoa flavanol content per serving (the COSMOS dose was 500 mg/day). Avoid products that list only "cocoa powder" milligrams.
Cocoa flavanol supplement (flavanol content stated)
Look for the milligrams of cocoa flavanols per serving on the label, not just cocoa extract weight.
Check Amazon price →Nordic Naturals Algae DHA
DHA is the other dietary nutrient with cognitive-aging trial data that we cover in depth on the site.
Check Amazon price →Frequently asked questions
Do cocoa flavanols improve memory?
In the COSMOS-Web trial of 3,562 older adults, a 500 mg/day cocoa flavanol extract did not significantly improve memory overall at one year. Memory did improve in people whose habitual diets were lower in quality or flavanols.
Is dark chocolate good for your brain?
Dark chocolate contains some flavanols, but processing reduces them, labels rarely state flavanol content, and the sugar and calories add up. It hasn't been shown to improve memory in trials.
How much cocoa flavanol was used in the COSMOS trial?
500 mg of cocoa flavanols per day, delivered as a cocoa extract supplement.
Who might benefit most from cocoa flavanols?
The COSMOS-Web findings suggest people with lower diet quality or low habitual flavanol intake. Improving diet with tea, berries, apples, and cocoa is a sensible first step.
Are cocoa flavanol supplements safe?
Thousands of older adults took them in COSMOS. They contain some caffeine and theobromine; check with your clinician if you're sensitive to stimulants or take interacting medications.
Citations
- Brickman AM, et al. Dietary flavanols restore hippocampal-dependent memory in older adults with lower diet quality and lower habitual flavanol consumption. PNAS. 2023;120(23). PMID 37252983. https://www.pnas.org/doi/10.1073/pnas.2216932120
- Brickman AM, Khan UA, Provenzano FA, et al. Enhancing dentate gyrus function with dietary flavanols improves cognition in older adults. Nature Neuroscience. 2014;17:1798–1803. https://www.nature.com/articles/nn.3850
- COSMOS Trial. COcoa Supplement and Multivitamin Outcomes Study — study overview. https://cosmostrial.org/
- EFSA Panel on Dietetic Products, Nutrition and Allergies. Scientific Opinion on the modification of the authorisation of a health claim related to cocoa flavanols and maintenance of normal endothelium-dependent vasodilation. EFSA Journal. 2014;12(5):3654. https://www.efsa.europa.eu/en/efsajournal/pub/3654
- Payne MJ, Hurst WJ, Miller KB, et al. Impact of fermentation, drying, roasting, and Dutch processing on epicatechin and catechin content of cacao beans and cocoa ingredients. J Agric Food Chem. 2010;58(19):10518–10527. https://pubs.acs.org/doi/10.1021/jf102391q