Every Study on Green Tea and the Ageing Brain: A Working Register

Somebody drinking green tea in Sendai in 2006 was asked how many cups a day. Somebody in a village in Ishikawa was asked the same in 2007, somebody in Niigata in 2011, somebody in Singapore in 2003. Between them, a few hundred thousand people have now been asked what they drink and then watched for years to see what happens to their memory.

This page keeps track of what those studies found. All of them, including the ones that found nothing.

I built it because the alternative is what you get everywhere else: one striking number, repeated without its context, usually the same number, usually from a study too small to support it. If you want to know what the research actually says, you need to see the whole set at once, including the failures and the arguments.

I sell Japanese tea, so I have an obvious interest in this literature looking good. The best answer I have to that is to show you all of it and let you judge. Where a study was funded or written by a beverage company, I say so in the table.

This page gets updated as studies publish.

Most confusion about this subject comes from treating all studies as equal. They are not, and the difference is not about quality. It is about what kind of question each design can answer.

A snapshot study measures everything at one moment. It might find that tea drinkers today have better memories than non-tea-drinkers today, which is genuinely interesting and proves nothing about direction. People whose memory is already failing often stop cooking, stop shopping and stop making tea. A snapshot cannot tell you whether the tea protected the memory or the memory protected the tea habit. Researchers call these cross-sectional.

A follow-forward study asks people what they drink while they are still well, then waits. Five years, ten, twenty. Because the drinking was recorded before anyone got ill, reverse causation is much less of a worry. These are called prospective cohort studies and they are the backbone of this field. Their weakness is that tea drinkers differ from non-tea-drinkers in many ways, and statistical adjustment can only correct for the differences the researchers thought to measure.

A trial assigns people at random to get the thing or a placebo. This is the only design that can demonstrate cause. It is also expensive, slow, and nearly impossible to run for the decades over which dietary effects on the brain seem to operate. There is exactly one proper trial of tea and cognition and it lasted a year.

A pooled analysis combines many studies to get a number that no single study was large enough to produce. It inherits every flaw of the studies it pools.

One more thing worth knowing. When a study reports a striking result from a small number of cases, be careful. If forty people in a study develop dementia and nine of them are non-tea-drinkers, shuffling three people between groups changes the answer dramatically. Researchers express this as a margin of error around their headline number, and when that margin is wide enough to include "no effect," the result is a lead rather than a finding. This matters below.

Researchers in a town in Ishikawa Prefecture followed 490 people aged over 60 for about five years, and found that daily green tea drinkers were roughly a third as likely to show cognitive decline as non-drinkers.¹ Coffee showed nothing. Black tea showed nothing.

This is the study everyone cites, and it matters because it was the first to follow people forward rather than photograph them once. It is also the source of the single most misused number in the field. For dementia specifically, as opposed to the broader measure of cognitive decline, the reported figure looks like a 74 per cent risk reduction and gets quoted that way constantly. Too few of the 490 developed dementia for that number to mean much, and the researchers' own margin of error included the possibility of no effect. The broader cognitive decline result is the one that stands.

Thirteen thousand six hundred and forty-five people, followed for 5.7 years, with dementia identified through Japan's long-term care insurance records rather than self-report.² Five or more cups a day came with about a quarter lower risk, and the finding held when the analysis was restricted to people with no memory complaints at the start.

This is where the field stopped being a curiosity. The sample was twenty-seven times larger than Nakajima's and the outcome was recorded administratively rather than reported by participants.

In Niigata, researchers followed 13,757 people and published twice.

After eight years, green tea showed no association with dementia at all.³ Coffee looked excellent, with the heaviest drinkers at roughly half the risk of the lightest.

After twelve years, on the same cohort, green tea did show an association: about a quarter lower risk in the top quartile, with risk falling about 4.8 per cent for every extra 150ml (5 fl oz) cup per day.⁴

I keep coming back to this pair. It is the clearest demonstration available that whatever green tea does to the ageing brain, it does slowly, and that a study of the wrong length will find nothing while a longer look at the identical people finds something. It should make anyone cautious about treating a single result as settled, in either direction.

Three studies have looked at this in Singapore, all from NUS, and they are strangely absent from most English writing on the subject even though they cover the tea Singaporeans actually drink.

The most important followed 957 Chinese Singaporeans aged 55 and over.⁵ Regular tea drinkers had around half the risk of developing neurocognitive disorders. Green tea came out slightly ahead of black and oolong, but both were protective, which is not what the Japanese studies found.

The finding that makes this study essential is about genetics. People carrying the APOE ε4 gene variant, the strongest common genetic risk factor for Alzheimer's, appeared to benefit most, with a risk reduction reported at around 86 per cent.

The Japanese brain imaging study below found no association whatsoever in APOE ε4 carriers.

Same gene, opposite result, both from serious groups. Nobody has resolved it. If you carry ε4 and want to know whether tea helps you specifically, the honest answer is that the two best studies to ask disagree.

This one is small but unusually designed. Researchers recorded what 1,155 people aged 44 to 66 were drinking in 1995 and again in 2000, then assessed their cognition around 2014 and 2015.⁶ That is roughly twenty years, and crucially the measurement was taken in middle age rather than old age.

Two to three cups a day came with an odds ratio of 0.56, and the effect was strongest in men. Four or more cups a day showed no significant benefit.

That last sentence is the interesting one. This is now the second human study pointing at moderate rather than maximal intake, and it is a pattern I did not expect to find.

MRI scans of 8,766 dementia-free Japanese aged 65 and over.⁷ Higher green tea intake tracked with smaller volumes of cerebral white matter lesions, the small patches of damage to the brain's wiring that accumulate with age and predict later dementia. Coffee showed no such association. Hippocampal and total brain volume were unchanged.

The effect was modest, roughly 3 per cent less lesion volume at three cups a day compared with one, and about 6 per cent at seven or eight. It is a snapshot, so it cannot show causation, and the authors say so.

What makes it matter is that it points at a location. White matter lesions are a blood vessel problem. If tea is doing something, this suggests it is doing it through the small vessels of the brain rather than through anything specific to Alzheimer's pathology. That fits the UK Biobank finding that the association was strongest for vascular dementia.

Ninety-nine people aged 60 to 85, all with mild cognitive impairment or memory complaints, randomly assigned to 2g of matcha daily or a coloured cornstarch placebo, in capsules, for twelve months, with neither participants nor researchers knowing who had which.⁸

The primary cognitive measures did not move.

A secondary measure did: participants in the matcha group became measurably better at recognising emotion in facial expressions. Sleep quality trended in the right direction without reaching significance.

A year is very short against a signal that took the Niigata researchers twelve years to detect, ninety-nine people is a small sample, and everyone in the trial already had symptoms. So this is not evidence that matcha does nothing. But it is the only randomised trial in existence here, and it did not confirm the headline claim, and any page discussing this subject that leaves it out is not being straight with you.

Every entry below follows the same order: who was studied and for how long, then what the study found, then what limits it. Nineteen studies, including the ones that found nothing.

**Tsurugaya, 2006.**⁹ A snapshot of 1,003 people aged 70 and over in Sendai. Higher green tea intake went with better cognitive test scores. Everything was measured at a single moment, so it cannot show which came first.

**Nakajima, 2014.**¹ Four hundred and ninety people aged over 60 in Ishikawa, followed about five years. Daily green tea drinkers were roughly a third as likely to show cognitive decline, while coffee and black tea showed nothing. Small, and the widely quoted dementia figure rests on too few cases to be reliable.

**Ōsaki, 2016.**² Thirteen thousand six hundred and forty-five people aged 65 and over in Miyagi, followed 5.7 years. Five or more cups a day came with about a quarter lower dementia risk. Diet was recorded only once, at the start.

**NILS-LSA, 2020.**¹⁷ One thousand three hundred and five people aged 60 to 85 in Aichi, followed 5.3 years with cognition tested every two years. Risk of cognitive decline was about 30 per cent lower at every level of intake from one cup a day upwards, and coffee showed nothing. The benefit did not grow with intake: one cup and four or more cups produced almost identical results.

**Murakami, 2021.**³ Thirteen thousand seven hundred and fifty-seven people aged 40 to 74 in Niigata, followed eight years. Green tea showed no association at all, while coffee drinkers at the highest intake had about half the risk. Superseded by the twelve-year analysis of the same people.

**JPHC Saku, 2025.**⁶ One thousand one hundred and fifty-five people aged 44 to 66 when first surveyed in Nagano, assessed about twenty years later. Two to three cups a day came with an odds ratio of 0.56, strongest in men. Four or more cups showed no significant benefit, and the sample is small.

**Murakami, 2025.**⁴ Thirteen thousand six hundred and sixty people in Niigata, followed twelve years. The top quartile of drinkers had about a quarter lower risk, falling roughly 4.8 per cent per extra cup. This is the same cohort that showed nothing at eight years.

**SLAS, 2008.**¹⁰ Two thousand five hundred and one Chinese Singaporeans aged 55 and over assessed at baseline, with 1,438 followed for one to two years. Tea drinkers showed less cognitive impairment and less decline. The follow-up is very short.

**SLAS, 2010.**¹¹ A snapshot of 716 Chinese Singaporeans aged 55 and over, given a full battery of cognitive tests. Tea was associated with better global cognition, memory, executive function and processing speed. Single moment in time.

**SLAS, 2016.**⁵ Nine hundred and fifty-seven Chinese Singaporeans aged 55 and over. Regular tea drinkers had about half the risk of developing a neurocognitive disorder, with green tea slightly ahead of black and oolong, and the largest benefit in APOE ε4 carriers. Only 72 cases in total, and that ε4 finding is directly contradicted by the Japanese imaging study.

**MrOS, 2018.**¹² Three thousand eight hundred and forty-four American men aged 65 and over. No association whatsoever between black tea and cognitive decline. Black tea only, and by the same lead author as the Singapore studies, which is worth knowing.

**UK Biobank, 2022.**¹³ Three hundred and seventy-seven thousand five hundred and ninety-two people, followed nine years. Tea drinkers were 16 per cent less likely to develop dementia, with vascular dementia 25 to 29 per cent lower. Overwhelmingly black tea, so strictly a different question from the Japanese work.

**CABLE, 2020.**¹⁸ A snapshot of 722 cognitively intact older Chinese adults who had their spinal fluid analysed. Frequent green tea drinkers had lower levels of tau, one of the two proteins that accumulate in Alzheimer's disease, though there was no difference in amyloid. A snapshot, and it measures proteins rather than symptoms.

**JPSC-AD, 2025.**⁷ MRI scans of 8,766 Japanese aged 65 and over, all free of dementia. More green tea went with smaller volumes of white matter lesions, while coffee showed nothing and overall brain volume was unchanged. A snapshot, with no association found in APOE ε4 carriers or people with depression, and part-funded by Suntory, which sells green tea.

**Matcha trial, 2024.**⁸ Ninety-nine people aged 60 to 85 with mild cognitive impairment or memory complaints, given 2g of matcha daily or a placebo for twelve months, with nobody knowing who had which. The primary cognitive measures did not move, while social acuity improved and sleep quality trended without reaching significance. Small, short, everyone already had symptoms, and several authors work for Ito En, a tea company.

**Kakutani, 2019.**¹⁴ A systematic review of the observational studies published up to August 2018. It concluded that the evidence appears to support a protective effect, and called for better-designed cohort studies to settle it. All three authors worked at Suntory.

**Li, 2024.**¹⁵ A pooled analysis of 38 cohorts covering 751,824 people and 13,017 dementia cases. Risk falls about 4 per cent per additional cup of tea per day, in a straight line, while coffee behaves non-linearly. It pools very different teas and very different populations into one number.

**Dose-response analysis, 2026.**¹⁶ Ten cohort and nested case-control studies pooled together. Dementia risk falls steadily as tea intake rises, while coffee follows a U-shaped curve with the lowest risk around two to three cups. Very recent, and I have only read the abstract.

Four interesting disagreements.

Does coffee do the same thing? The Niigata cohort found coffee strongly protective. The brain imaging study found nothing for coffee. The Saku study found a coffee benefit only in its older participants. The large pooled analyses find tea straightforwardly linear and coffee U-shaped. Anyone telling you confidently which drink is better is going beyond the evidence.

Does it depend on your genes? Singapore says ε4 carriers benefit most. Japan's imaging study says they do not benefit at all.

Is it green tea specifically? The Japanese studies keep finding an effect for green tea and none for black. The UK Biobank found a clear effect in a population drinking almost entirely black tea. Singapore found both. The American study of black tea found nothing.

How much? This is the argument that has shifted most while I was compiling this page. Ōsaki found the benefit at five or more cups. Saku found it at two to three, with no significant benefit at four or more. The Aichi study is the striking one: it found essentially the same reduction at one cup a day, at two to three, and at four or more, with the numbers barely moving across the three groups. The imaging study saw lesion volume improve with intake but nowhere near proportionally. Only the big pooled analyses report a steady per-cup decline, and pooled analyses are the design least able to detect a curve.

Three separate human studies now point at a plateau rather than a ladder. That does not prove there is one, but it is more support than I expected to find for the idea that a moderate habit gets you most of whatever is on offer.

There is no long randomised trial. There probably never will be, because you cannot blind people to drinking tea for twenty years and no one will fund it.

Almost nobody has measured what people actually drank, as opposed to what they said they drank on a questionnaire, once, years ago.

There is very little on brewing method, leaf type or strength, which means the entire literature treats a weak bag of bottled tea and a strong cup of sencha as the same exposure.

And almost all of it is in East Asian and European populations, drinking their own local teas.

Reading the set rather than the highlights, the pattern is a consistent, modest, slow effect that shows up more reliably in larger studies and longer follow-ups, most likely operating through the brain's small blood vessels, with no clear answer yet on dose or on who benefits.

That is a good deal weaker than most green tea marketing implies, including some that has appeared on this site in the past. It is also considerably stronger than nothing, and it is getting firmer rather than shakier as the studies get bigger.

For the fuller version of this argument, see our main article on green tea and dementia.

If you are worried about your own memory or someone else's, talk to a doctor rather than a tea merchant.

¹ Noguchi-Shinohara M, Yuki S, Dohmoto C, et al. "Consumption of green tea, but not black tea or coffee, is associated with reduced risk of cognitive decline." PLoS One. 2014;9(5):e96013. https://doi.org/10.1371/journal.pone.0096013

² Tomata Y, Sugiyama K, Kaiho Y, et al. "Green Tea Consumption and the Risk of Incident Dementia in Elderly Japanese: The Ohsaki Cohort 2006 Study." American Journal of Geriatric Psychiatry. 2016;24(10):881-889. https://doi.org/10.1016/j.jagp.2016.07.009

³ Matsushita N, Nakanishi Y, Watanabe Y, et al. "Association of coffee, green tea, and caffeine with the risk of dementia in older Japanese people." Journal of the American Geriatrics Society. 2021;69(12):3529-3544. https://doi.org/10.1111/jgs.17407

⁴ Kaise R, Kitamura K, Watanabe Y, et al. "Green tea consumption and dementia risk in community-dwelling Japanese people aged 40-74 years: A 12-year cohort study." Journal of Nutrition, Health and Aging. 2025. https://doi.org/10.1016/j.jnha.2025.100615

⁵ Feng L, Chong MS, Lim WS, et al. "Tea Consumption Reduces the Incidence of Neurocognitive Disorders: Findings from the Singapore Longitudinal Aging Study." Journal of Nutrition, Health and Aging. 2016;20(10):1002-1009. https://pubmed.ncbi.nlm.nih.gov/27925140/

⁶ Koreki A, Nozaki S, Shikimoto R, et al. "A longitudinal cohort study demonstrating the beneficial effect of moderate consumption of green tea and coffee on the prevention of dementia: The JPHC Saku Mental Health Study." Journal of Alzheimer's Disease. 2025;103(2):519-527. https://doi.org/10.1177/13872877241303709

⁷ Shibata S, Noguchi-Shinohara M, Shima A, et al. "Green tea consumption and cerebral white matter lesions in community-dwelling older adults without dementia." npj Science of Food. 2025;9:2. https://doi.org/10.1038/s41538-024-00364-w

⁸ Uchida K, Meno K, Korenaga T, et al. "Effect of matcha green tea on cognitive functions and sleep quality in older adults with cognitive decline: A randomized controlled study over 12 months." PLoS One. 2024;19(8):e0309287. https://doi.org/10.1371/journal.pone.0309287

⁹ Kuriyama S, Hozawa A, Ohmori K, et al. "Green tea consumption and cognitive function: a cross-sectional study from the Tsurugaya Project." American Journal of Clinical Nutrition. 2006;83(2):355-361. https://pubmed.ncbi.nlm.nih.gov/16469995/

¹⁰ Ng TP, Feng L, Niti M, Kua EH, Yap KB. "Tea consumption and cognitive impairment and decline in older Chinese adults." American Journal of Clinical Nutrition. 2008;88(1):224-231. https://pubmed.ncbi.nlm.nih.gov/18614745/

¹¹ Feng L, Gwee X, Kua EH, Ng TP. "Cognitive function and tea consumption in community dwelling older Chinese in Singapore." Journal of Nutrition, Health and Aging. 2010;14(6):433-438. https://doi.org/10.1007/s12603-010-0095-9

¹² Feng L, Langsetmo L, Yaffe K, et al. "No Effects of Black Tea on Cognitive Decline Among Older US Men: A Prospective Cohort Study." Journal of Alzheimer's Disease. 2018;65(1):99-105. https://doi.org/10.3233/JAD-180103

¹³ Hu HY, Wu BS, Ou YN, et al. "Tea consumption and risk of incident dementia: A prospective cohort study of 377 592 UK Biobank participants." Translational Psychiatry. 2022;12:171. https://doi.org/10.1038/s41398-022-01923-z

¹⁴ Kakutani S, Watanabe H, Murayama N. "Green Tea Intake and Risks for Dementia, Alzheimer's Disease, Mild Cognitive Impairment, and Cognitive Impairment: A Systematic Review." Nutrients. 2019;11(5):1165. https://doi.org/10.3390/nu11051165

¹⁵ Li F, Liu X, Jiang B, et al. "Tea, coffee, and caffeine intake and risk of dementia and Alzheimer's disease: a systematic review and meta-analysis of cohort studies." Food and Function. 2024;15:8330-8344. https://doi.org/10.1039/D4FO01750A

¹⁶ "Coffee and tea consumption and risk of dementia: a dose-response meta-analysis of cohort and cohort-nested case-control studies." Nutrition, Metabolism and Cardiovascular Diseases. 2026. https://www.sciencedirect.com/science/article/pii/S2950433326000029

¹⁷ Shirai Y, Kuriki K, Otsuka R, et al. "Green tea and coffee intake and risk of cognitive decline in older adults: the National Institute for Longevity Sciences, Longitudinal Study of Aging." Public Health Nutrition. 2020;23(6):1049-1057. https://doi.org/10.1017/S1368980019002659

¹⁸ Ma YH, Wu JH, Xu W, et al. "Associations of Green Tea Consumption and Cerebrospinal Fluid Biomarkers of Alzheimer's Disease Pathology in Cognitively Intact Older Adults: The CABLE Study." Journal of Alzheimer's Disease. 2020;77(1):411-421. https://doi.org/10.3233/JAD-200410