You have started noticing it: the word that is right there and will not come, the reason you walked into the room, gone. That is not a dementia sign, and nothing here is proven to prevent dementia either. What 25 years of following nearly 8,000 people does show is that the habits with the strongest evidence are the dull ones, sleep and heart health in your forties and fifties.
What Has the Longest Evidence Behind It?
Sleep and blood pressure, by a distance.
Following nearly 8,000 British civil servants for 25 years, researchers at University College London and Inserm in Paris found that sleeping six hours or less at age 50 came with a higher rate of dementia, and again at 60. People short on sleep at all three check-ins, at 50, 60 and 70, had about 30% higher risk. That last number is the shakiest: the range around it runs from 1.00 to 1.69, and 1.00 would mean no effect at all, so it might be nothing.
Two limits. It watched people rather than testing anything, and the authors say short sleep may be an early sign the disease has already started rather than a cause. When a genetic study of 446,118 people went looking for cause, it found none, and the signal ran the other way: Alzheimer's risk predicted shorter sleep.
The same cohort scored heart health at 50 (smoking, diet, activity, weight, blood sugar, cholesterol, blood pressure) and followed people for 25 years. Each point better came with about 11% lower dementia risk. Same study design, so the same limit applies: watched, not tested. Not one habit, though. The stack. Our 2-minute vitality check covers the same ground.
Does a Sauna Really Lower Alzheimer's Risk?
The link is strong. The catch is the temperature.
A Finnish team at the University of Eastern Finland followed 2,315 men for a median of 20.7 years (Age and Ageing, 2017). Men using a sauna 4 to 7 times a week were about 65% less likely to be diagnosed with Alzheimer's than men who went once a week. The estimate is loose: the real effect could be small or large. And that study recorded only how often people went.
A second, much larger Finnish study looked at heat and time. Following 13,994 men and women for 39 years, it found the lowest dementia risk at around 80 to 99°C, though not clearly enough to be sure. What was clear: at 100°C and above, risk ran roughly double in the first 20 years, fading across the full 39. How long people stayed in barely mattered.
So: frequent and warm, not occasional and scorching, and both studies only watched people. Sauna is also not for everyone. Skip it with unstable chest pain, a recent heart attack, or severe narrowing of the aortic valve, the heart's main outflow valve, and never combine it with alcohol, which raises the risk of a dangerous blood-pressure drop, an irregular heartbeat, and sudden death.
Does Exercise Actually Grow Your Brain?
In a year-long trial of 120 older adults run at the University of Illinois, regular aerobic exercise grew the front of the hippocampus, the brain's memory centre, by about 2%, while a stretching group's shrank by about 1.4%. Two things have gone wrong with how that study gets retold.
We used to write that memory improved with it. The paper's own words: the exercise group *"did not improve performance above that achieved by the stretching control group."* Both groups improved by a similar amount. BDNF, the growth protein usually credited for this, did not rise more in the exercisers either.
The brain-growth part did not repeat. Pooling eight later trials in 554 people, researchers found no clear effect of exercise on hippocampal volume, and the 2011 study's own first author is a co-author on that null. The largest trial in this area, 1,635 adults over two years, found no overall cognitive benefit, though the oldest and frailest did improve on one measure.
Move anyway. Heart health at 50 is the strongest brain signal in this article, and exercise is how you move it. The hippocampus was never the point.
What About Your Hearing?
The one thing here that has been put to a proper randomised trial.
Hearing loss sits on the Lancet standing Commission's list of changeable things linked to dementia. In ACHIEVE, 977 adults aged 70 to 84 with untreated hearing loss were randomly given hearing aids or health education, then followed three years. Overall, both groups declined the same amount. But among those recruited from an older, higher-risk group, the hearing aids appeared to slow it.
The trial missed its main target and found a signal where risk was highest. That is still better evidence than anything else here, because somebody ran the experiment. If you are straining to follow conversations, get your hearing tested.
Can a Breathing Exercise Change Your Brain?
In one spot, in one trial, and it did not make anyone think better.
USC randomised 193 adults (2022). One group learned to breathe slowly at a pace tuned to their own heart rhythm; the other wore the same sensor but was coached to keep their heart rate flat. Five weeks later, of the 145 with usable scans, the first group had gained volume in one region behind the left eye and the second had lost it. That is two active groups compared with each other, and scan volume is not new brain cells.
The same team tested thinking, and working memory, processing speed and self-control were unchanged. No lab outside that team has repeated the brain finding in a peer-reviewed study. Slow breathing is free and safe, and we cover it in how to calm down in five minutes. Just do not expect a sharper memory.
What About Supplements and Cold Water?
NAD+ and NMN. NMN raises NAD+ in the blood easily. Whether that sharpens a brain has now been tested: three randomised trials gave older adults with early memory trouble a NAD+ booster for 8 to 12 weeks, and thinking scores did not beat placebo in any of them. The cleanest, with no industry funding, found the same nothing as the two with industry ties. A 2026 study of over 300 people also found blood NAD+ does not fall with age, which unsettles the premise under the whole category.
Vitamins. VITAL, run from Harvard, gave 2,000 IU of vitamin D daily to over 4,000 adults aged 60 and up: no slowing of decline, and no difference by starting vitamin D level. D-Health found the same nothing in 3,887 Australians on a monthly high dose. B vitamins are the one real exception: an Oxford trial of 271 people over 70 with mild memory problems slowed brain shrinkage, biggest in those who started high on homocysteine, a blood marker B vitamins bring down, though its lead author holds two university patents on that exact treatment. Oxford's own pooled analysis of 11 trials in 22,000 people found no effect on thinking. Fish oil: a Cochrane review found no benefit in healthy older adults, and the one clearly positive trial was in people already losing memory and moved one test out of several. Eat the fish. Skip the certainty.
Cold water. Cold exposure raises cold-shock proteins. Raising one in mice with Alzheimer-type and prion disease protected the connections between brain cells and lengthened survival (Nature, 2015, MRC Toxicology Unit at Leicester and Cambridge). Mice, and laboratory cooling: in humans that protein has only been measured in blood, in patients cooled by machines in hospital. Meanwhile 15 of 18 studies found being acutely cold makes people think *worse*. And sudden immersion can set off a dangerous heart rhythm even in healthy people. A lead for a future drug, not a reason to stand under cold water.
What to Do Tomorrow
Change nothing yet. Tomorrow night, write down when you got into bed and when you got up. Do that for a week, then subtract, and see what your real number is.
Sleep has the longest evidence here, and it is the habit people guess wrong about themselves. You cannot protect six hours you did not know you were missing. Here is how we check claims like these, and here is how I live with it.
