Cognitive Health: 9 Simple Habits the Research Actually Supports

Why Physical Activity Is the Most Reliable Brain-Health Habit We Have

Of all the levers researchers have studied for protecting cognition, regular movement is the one with the most converging evidence. Aerobic exercise, resistance training, and the simple consistency of daily activity each show up as protective in study after study — not as a dramatic cure, but as a steady, cumulative benefit that adds up over time.

In a well-known randomized trial published in PNAS, one year of aerobic exercise increased the volume of the anterior hippocampus — a brain region central to memory — and improved spatial memory in older adults compared with a stretching control group.

A later systematic review in the British Journal of Sports Medicine of exercise studies in adults aged 50 and older reached a similar conclusion. Combining aerobic and resistance training, in sessions of 45 to 60 minutes, produced the clearest cognitive gains.

For specific targets, the 2018 US Physical Activity Guidelines describe a range that most public health bodies have converged on: roughly 150 to 300 minutes of moderate-intensity aerobic activity each week (or 75 to 150 minutes of vigorous activity), plus muscle-strengthening work on two or more days. That's about thirty minutes of brisk walking on most days, alongside two basic strength sessions a week. The biology lines up with the epidemiology — exercise raises brain-derived neurotrophic factor (BDNF), supports the small blood vessels that feed the brain, and reduces the chronic inflammation that increasingly looks like a driver of cognitive aging.

The effect size in any single trial is modest. What makes the case strong is the consistency of the direction across very different study designs. Physical inactivity is also named explicitly in the Lancet Commission's list of modifiable factors.

Why Chronic Short Sleep Is a Quiet Risk Factor for Cognition

Sleep doesn't always make the brain-health headlines, but the underlying biology is striking. During deep sleep, the brain's glymphatic system — a kind of waste-clearance plumbing — becomes more active, helping flush out metabolic byproducts that accumulate during waking hours. Experimental work in mice published in Science shows that interstitial space in the brain expands during sleep, and that clearance of metabolites — including amyloid-beta, the protein that builds up in Alzheimer's disease — increases substantially compared with the waking state.

Translating that mouse work to humans is where the caveats come in. Glymphatic flow can't yet be measured directly in living people. But the epidemiology points in the same direction. A 25-year analysis of the Whitehall II cohort in Nature Communications, tracking about 8,000 British civil servants, found that people who consistently slept six hours or fewer per night in midlife had roughly a 30 percent higher risk of developing dementia compared with seven-hour sleepers, even after adjusting for cardiovascular, behavioral, and mental health factors.

The practical implication is reassuring rather than alarming. A single short night isn't meaningfully changing anyone's long-term trajectory. The pattern that matters is chronic — months and years of routinely under-sleeping. For most adults, seven to nine hours is the window where the research clusters. Sleep apnea is worth flagging separately. If you or a partner notice loud snoring, gasping, or breath pauses during sleep, that's the kind of thing worth raising with a clinician, because untreated apnea has its own well-documented associations with cognitive decline.

The Most Studied Medical Lever for Protecting the Brain

If there's a single medical variable that has been studied extensively for its effects on cognition, it's blood pressure. Chronic hypertension damages the small vessels that supply the brain, contributing to white-matter changes that, over decades, raise the risk of both vascular cognitive impairment and Alzheimer's-type dementia.

The most rigorous evidence comes from the SPRINT MIND trial, a large randomized study published in JAMA involving more than 9,000 adults over 50 with elevated cardiovascular risk. Participants randomly assigned to intensive blood pressure control — a systolic target below 120 mm Hg — had a meaningfully lower rate of mild cognitive impairment than those treated to the standard target of below 140. The effect on probable dementia alone did not quite reach statistical significance, which is an honest caveat worth naming, but the combined outcome of MCI and dementia was reduced, and the direction of the evidence is consistent.

For context, the American Heart Association defines stage 1 hypertension as a reading of 130/80 mm Hg or higher, with stage 2 starting at 140/90. The takeaway isn't that everyone should chase a sub-120 systolic — those targets need to be individualized, especially for older adults at risk of falls or medication side effects. The takeaway is that knowing your numbers, and treating them when they're persistently elevated, is one of the highest-yield things you can do for your brain over the long run.

A Pattern That's Good for the Heart and Probably Good for the Brain

The diet pattern with the most consistent evidence for cognitive health is the broad family of Mediterranean-style eating: meals built around vegetables, fruits, whole grains, legumes, nuts, olive oil, and fish, with limited red meat and minimally processed food. The MIND diet — a Mediterranean–DASH hybrid specifically tuned for brain health — is the most studied variant.

In a 2015 cohort study in Alzheimer's & Dementia, participants with the highest adherence to the MIND diet had a 53 percent lower rate of Alzheimer's disease over an average 4.5 years of follow-up than those with the lowest adherence. Even moderate adherence was linked to a 35 percent lower rate.

Randomized data pointed in a similar direction at first. The PREDIMED-NAVARRA trial in the Journal of Neurology, Neurosurgery & Psychiatry found that an energy-unrestricted Mediterranean diet supplemented with olive oil or nuts produced better cognitive performance over 6.5 years than a low-fat control diet.

Here is the part you don't see in every brain-health roundup. A more recent randomized trial of the MIND diet, published in the New England Journal of Medicine in 2023, did not find that participants assigned to MIND outperformed those on a control diet with similar mild caloric restriction over three years. Both groups showed small improvements. The fair reading is that overall diet quality matters more than the specific MIND framework, and that short-term randomized differences are hard to detect when both arms are eating reasonably well.

For practical purposes, the lesson is the one you've probably heard before, with more confidence behind it: more plants, more fish, more olive oil, fewer sugar-sweetened drinks, less ultra-processed food. Berries and leafy greens earn specific mentions in the MIND research, if you're looking for places to start.

One of the Largest Modifiable Risk Factors in Midlife

Hearing has emerged, somewhat unexpectedly, as one of the biggest single levers in the dementia-prevention landscape. The 2024 Lancet Commission estimates that midlife hearing loss accounts for a population-attributable fraction of roughly 7 percent of dementia cases globally — among the largest of any single factor on its list.

The observational evidence is what first put hearing on the radar. A prospective study in Archives of Neurology of 639 adults with a mean baseline age in their late 50s found that even mild hearing loss roughly doubled the risk of subsequent dementia compared with normal hearing; moderate loss tripled it; severe loss multiplied it fivefold. That dose-response pattern is part of what makes researchers take the association seriously.

The randomized evidence is more recent and more qualified. The ACHIEVE trial in The Lancet tested whether a best-practices hearing intervention (audiologic care plus hearing aids) would slow cognitive decline over three years in roughly 1,000 older adults with hearing loss. In the overall sample, the answer was no — the difference between hearing intervention and a health-education control didn't reach statistical significance. But in a prespecified higher-risk subgroup drawn from the long-running ARIC study, the intervention did slow decline.

Here's how to read that honestly: the evidence supports treating hearing loss, but the cognitive benefit is clearest in higher-risk older adults rather than as a universal cognitive prescription. The supporting case is even stronger when you factor in the quality-of-life and social-engagement benefits of better hearing, which aren't in dispute. Over-the-counter hearing aids are now widely available in the US, which lowers the barrier for many people considerably.

Social Contact Is Genuinely Protective — Not Just Nice to Have

Loneliness research used to feel like soft science. It doesn't anymore. The evidence linking social engagement to lower dementia risk is now robust enough that the Lancet Commission includes social isolation in its list of modifiable risk factors, and the underlying datasets are substantial.

A systematic review of 19 longitudinal cohort studies in Ageing Research Reviews found that poorer social relationships — isolation, infrequent contact, smaller networks — were consistently associated with higher dementia incidence in community-dwelling adults.

The Whitehall II cohort, which has tracked British civil servants for nearly three decades, published findings in PLOS Medicine showing that more frequent social contact at age 60 was associated with lower dementia risk and better cognitive trajectories years later.

The reasonable skeptical question here is reverse causation: does early cognitive decline make people withdraw socially, rather than the other way around? The most rigorous studies adjust for baseline cognition and follow people for decades, and the association largely holds. Social interaction is also cognitively demanding work — language, memory, attention, reading other people's intentions — and it tends to come bundled with movement, novelty, and meaning. There's no specific frequency that emerges as the magic threshold. The consistent signal is that regular, in-person contact matters.

Cognitive Reserve Is Real — But It's Not Built by Puzzle Apps

The "use it or lose it" idea has more science behind it than the slogan suggests. Decades of work on what researchers call cognitive reserve, summarized by Yaakov Stern in Nature Reviews Neuroscience, describe a real phenomenon: people with more education and more lifelong engagement in mentally demanding activities can often maintain cognitive function longer even when their brains show the structural changes of aging or early disease.

What's less clear is whether commercial brain-training apps build that reserve in a meaningful way. The ACTIVE trial, a major randomized study reported in the Journal of the American Geriatrics Society, trained nearly 3,000 older adults in memory, reasoning, or speed. Improvements in the specific trained skill persisted for up to ten years — impressive on its own terms. But transfer to untrained domains and to everyday function was modest. The honest translation is that brain-training apps generally make you better at the games themselves, not broadly better at cognition.

What seems to work is harder, and that's part of why it works. Genuinely novel learning — a new language, a musical instrument, a craft, a complex hobby — engages the brain in ways that low-effort entertainment doesn't. The difficulty and unfamiliarity are the point. The activity matters less than the engagement: bridge, woodworking, ceramics, learning to read music in your 60s, taking a community college course — all of these plausibly count.

Depression and Chronic Stress Are Both on the Risk List

Mental health doesn't always get billed as brain-health, but the connection runs both ways. The Lancet Commission includes depression among its modifiable risk factors, based on longitudinal evidence that a history of clinical depression is associated with higher subsequent dementia incidence.

There's an important honest caveat. Depression can be an early symptom of dementia rather than a cause of it, so some of the observed association reflects reverse causation. Even so, treating depression — and treating it well, with the full range of effective options — has independent value, and the long-term cognitive picture adds to the reasons. If you've experienced persistent low mood, hopelessness, or loss of interest in things you used to enjoy for two weeks or longer, that's the kind of thing that warrants a real conversation with a clinician.

Chronic stress is a related but distinct story. A Nature Reviews Neuroscience review summarizes evidence that prolonged elevation of stress hormones, particularly cortisol, is associated with hippocampal atrophy and memory impairment across the lifespan. The dementia-prevention evidence for stress reduction specifically is less robust than for some other factors, but the biological plausibility is solid, and the practices that help with chronic stress — regular movement, sleep, social connection, mindfulness-based approaches — overlap heavily with the other habits on this list.

Smoking, Alcohol, and Metabolic Health All Show Up in the Data

A handful of behaviors keep appearing across dementia-risk studies in ways that are worth taking seriously together: smoking, heavy alcohol use, and the cluster of metabolic conditions — particularly type 2 diabetes — that affect cardiovascular and brain health alike.

For smoking, a meta-analysis in the American Journal of Epidemiology found that current smokers had roughly a 1.7-fold higher risk of Alzheimer's disease and faster cognitive decline compared with former smokers, while former smokers' risk was similar to never-smokers. The implication is that the brain benefits of quitting begin to show within years, not decades.

For alcohol, the science has shifted. The older "moderate drinking is protective" finding has not held up under more recent analyses that better account for confounding, including the "sick quitter" problem in which people who stop drinking for health reasons end up incorrectly inflating the apparent benefits of zero consumption. The GBD 2016 alcohol analysis in The Lancet concluded that the consumption level that minimizes overall health harm is zero standard drinks per week. Whatever the right personal answer is, less alcohol is the direction the evidence has moved.

For metabolic health, type 2 diabetes shows up consistently as a risk factor. A 2018 review in Nature Reviews Endocrinology describes a roughly 1.5- to 2-fold increased dementia risk associated with diabetes, with mechanisms ranging from cerebrovascular disease and chronic hyperglycemia to insulin resistance in the brain. Midlife obesity travels with this cluster and is independently flagged in the Lancet list. The brain is, in this respect, a cardiovascular organ — and what protects the heart tends to protect cognition.

A short note, because this comes up so often. Over-the-counter memory supplements — ginkgo, fish oil pills marketed for cognition, proprietary "brain" blends — have largely not held up in rigorous trials. The Ginkgo Evaluation of Memory (GEM) study, a 3,000-person randomized trial reported in JAMA, found that a standardized ginkgo extract did not reduce dementia incidence over more than six years of follow-up compared with placebo. Omega-3 supplements for cognition have produced similarly disappointing results in randomized trials, even though dietary fish intake remains part of the Mediterranean-style pattern.

Brain-training apps are a related story. The improvements they produce tend to be specific to the trained task rather than transferring to real-world thinking, memory, or function. That doesn't make them harmful — they're entertaining, and entertainment has its own value — but they're not a substitute for the kinds of activity, engagement, and learning the rest of this list describes.