Brain health

Why brain health matters#

Dementia is a decline in memory, thinking, and daily function severe enough to take away independence. It is a syndrome with several causes. Alzheimer's disease, defined by amyloid plaques and tau tangles in the brain, accounts for an estimated 60–70% of cases; vascular disease, Lewy bodies, and frontotemporal degeneration cause most of the rest, often in combination. Mild cognitive impairment (MCI) is the stage before dementia: measurable decline that does not yet stop you from running your own life.

Worldwide, 57 million people were living with dementia in 2021, with nearly 10 million new cases a year. Alzheimer's and other dementias killed 2.1 million people in 2023, the fifth leading cause of death. In the US, about 1 in 9 people 65 and older has Alzheimer's, health and long-term care for dementia is projected to cost $409 billion in 2026, and nearly 13 million Americans provide unpaid care.

The lifetime number is higher than most people expect. In a 2025 analysis of 15,043 adults followed in the US ARIC study, the risk of dementia between 55 and 95 was 42%, counting death from other causes as a competing outcome. It was roughly 45% to 60% in women, Black adults, and carriers of the APOE4 gene variant. Because the population is aging, the authors projected new US cases to roughly double, from about 514,000 in 2020 to about 1 million a year in 2060.

The 14 modifiable risk factors#

The Lancet Commission on dementia prevention is an international expert panel that periodically pools the evidence. Its 2020 report listed 12 risk factors. The 2024 report added two, high LDL cholesterol in midlife and untreated vision loss in later life, and estimated that removing all 14 could prevent or delay about 45% of dementia.

The remaining 55% reflects age, genes, and causes not yet identified. Two points stand out. Hearing loss and LDL each match smoking, inactivity, and diabetes combined, and both are cheap to address. And most of the list is the same list that prevents heart attacks and strokes.

APOE4 and genetic risk#

APOE is a gene that helps carry fats in the blood and brain. It comes in three common versions: e2, e3, and e4. Everyone carries two copies, one from each parent. According to the NIA, about 15% to 25% of people carry at least one e4 copy and 2% to 5% carry two. A tiny fraction of Alzheimer's, usually starting before 65, comes from rare mutations in APP, PSEN1, or PSEN2 that are passed on directly; that is a different situation that calls for a genetics clinic.

The largest analysis, a 2023 study of 68,756 people, found the effect of one e4 copy varied by ancestry: an odds ratio of 4.5 in East Asian, 3.5 in White, 2.2 in Black, and 1.9 in Hispanic participants. Two copies are a different category. A 2024 study of more than 13,000 people found that almost all e4 homozygotes had Alzheimer's pathology, with biomarkers rising from about age 55 and abnormal amyloid in nearly all by 65. Symptoms began at an average of 65, which led the authors to call APOE4 homozygosity a distinct genetic form of Alzheimer's.

If you carry APOE4#

The levers are the same; the stakes are higher, which is a reason to start earlier and be stricter about blood pressure, LDL, fitness, sleep, hearing, and alcohol. In the US POINTER lifestyle trial, the extra benefit of the structured program was the same in carriers and noncarriers. No special diet or supplement has been shown to help carriers: in the PreventE4 trial, 2 g of DHA a day for two years did not improve cognition in 365 adults 55 to 80, including carriers. If you later develop symptoms, your genotype matters for drug decisions, because brain swelling from anti-amyloid antibodies is much more common with two copies (see Alzheimer's drugs).

Should you get tested#

For most people serious about prevention, the optimal choice is to know your APOE genotype Weak, with counseling, and after you have bought any long-term care or life insurance you want. Some prevention-focused clinics test it routinely. The case is practical: a carrier result is a reason to start earlier and hold tighter targets, and the fear that the news will do harm has not held up. In the REVEAL trial, 162 adults with a parent with Alzheimer's were randomized to learn their genotype or not; those told had no more anxiety, depression, or distress over a year, and carriers were 2.7 times as likely to change their health habits. The evidence is weak because no trial shows that knowing changes who gets dementia.

Guideline minimum: guidelines do not recommend APOE testing for people without symptoms, and the NIA notes that results cannot tell you who will or won't develop Alzheimer's.

Reasons to test#

• You would use the result to commit harder to prevention, or to plan finances and care.
• You want to join prevention trials, many of which recruit carriers.
• You are already being evaluated for memory problems: the drug labels say APOE testing should be done before starting lecanemab or donanemab.

Reasons to wait or skip it#

• The levers are the same whatever the result; only the urgency changes.
• A two-copy result can cause real distress, and it also says something about your parents, siblings, and children.
• The US genetic nondiscrimination law, GINA, protects health insurance and employment but not life, disability, or long-term care insurance. If that insurance matters to you, think about timing before you test.

If you test, do it through a clinician or genetic counselor, or read the counseling material that comes with a consumer test before you open the result. The labs chapter covers how to order genetic tests and protect your data.

Prevention levers with evidence#

Each of these has its own chapter. Here is what each does for the brain, and how strong the evidence is.

Blood pressure#

This is the best-tested lever Strong. In SPRINT MIND, 9,361 adults with high blood pressure were randomized to a systolic target below 120 or below 140 mm Hg. The intensive group had 19% fewer cases of mild cognitive impairment (hazard ratio 0.81) and 15% fewer cases of MCI or dementia combined. The 17% reduction in probable dementia alone fell short of statistical significance; the trial stopped early and may have been too small to tell. A 2025 cluster-randomized trial in rural China then treated 33,995 people 40 and older with uncontrolled hypertension to below 130/80. Systolic pressure fell 22 mm Hg more than with usual care, and all-cause dementia was 15% lower over four years. The optimal target for the brain is the same as for the heart: under 120/80, checked at home (guideline minimum: under 130/80). Targets and drugs are in Blood pressure.

Hearing#

Hearing loss carries the largest single attributable fraction Moderate. In observational studies of about 127,000 people, hearing aid users had 19% lower risk of cognitive decline; the randomized ACHIEVE trial found no overall effect over three years but slower decline in a higher-risk subgroup. The optimal plan is a baseline hearing test by 40, a repeat every 3 years, and yearly tests from 60 or once loss appears, with hearing aids as soon as they would help; see Getting tested.

Vision and cataract surgery#

In a Seattle cohort of 3,038 older adults with cataracts, those who had cataract surgery had 29% lower dementia risk than those who did not Moderate. Glaucoma surgery, which does not restore vision, showed no such link, making simple "healthier people get surgery" bias less likely. Get an eye exam every 2 years, with a dilated baseline by 40 and yearly exams from 65; see Eyes.

LDL cholesterol#

The Commission added high midlife LDL in 2024 and gave it a 7% share, based mostly on long-term observational studies Moderate. No trial has tested LDL lowering with dementia as the main outcome, but the heart case alone justifies it, so use the heart chapter's optimal plan: ApoB tested yearly, with a target under about 60 mg/dL, lower if your risk is high. Guideline goals depend on 10-year risk and are looser. See ApoB, LDL, and non-HDL cholesterol.

Diabetes and weight#

Diabetes and obesity in midlife each raise dementia risk in cohort studies, possibly through blood vessel damage and insulin resistance in the brain Moderate. Prevention and treatment are in Metabolic health.

Exercise#

Across cohort studies of more than 250,000 people, physically active adults had 20% lower risk of dementia and 14% lower risk of Alzheimer's, including over follow-ups of 20 years or more (which reduces the worry that early dementia makes people less active) Moderate. Exercise also lowers blood pressure, glucose, and depression. See Exercise.

Sleep#

In a British cohort followed for 25 years, people sleeping 6 hours or less at 50 and 60 had 22% to 37% higher dementia risk than those sleeping 7 hours, and persistent short sleep was linked to 30% higher risk Moderate. Sleep apnea is common, treatable, and worth ruling out. See Sleep.

Social and mental engagement#

Less education (5%) and later-life social isolation (5%) together account for a tenth of dementia in the Commission's estimates, possibly by building "cognitive reserve," the brain's ability to keep working despite damage Moderate. Staying connected and mentally challenged is low-risk and good for mood. See Mind and connection.

Depression#

Depression is linked to later dementia, but it can also be an early symptom of it, which makes cause and effect hard to separate Mixed. Whether treating depression prevents dementia is unproven; treating it is worthwhile regardless. See Mind and connection.

Alcohol, smoking, and air pollution#

Heavy drinking, smoking, and long-term exposure to fine-particle air pollution each carry a share in the Commission's table Moderate. The Avoiding harm chapter covers all three, including what to do on bad air days.

Medications that cloud thinking#

Drugs with strong anticholinergic effects (some older antidepressants, bladder antispasmodics, and sedating antihistamines) blunt a brain chemical important for memory Moderate. In a UK study of 284,343 people, the heaviest long-term use was linked to 49% higher odds of dementia. It is observational, but it is a good reason to review long-term medications with your doctor, especially after 65. Never stop a prescribed drug on your own. After 65, a hospital stay or surgery also brings a real risk of delirium, which can speed cognitive decline; see Aging well after 65 for how to prevent it.

Lifestyle and brain-training trials#

Single-factor studies are hard to run for dementia, so researchers have tested bundles of changes. The results are consistently positive and consistently modest.

US POINTER, reported in July 2025 at the Alzheimer's Association conference and published in JAMA in August, had no do-nothing control group. Both arms raised activity and improved diet, so the trial shows that more structure helps a little; it cannot say how much either program beats doing nothing. The ACTIVE finding is intriguing because the training was brief and decades old by the time dementia was counted, but it rests on the subgroup of participants who completed booster sessions, so treat it as promising rather than settled Mixed. The speed training used speeded, adaptive tasks that split attention; no specific consumer app has its own dementia outcome trial.

Vaccines, GLP-1 drugs, and supplements#

Shingles vaccine#

Some of the most surprising brain evidence of the decade comes from a vaccine Moderate. In Wales, eligibility for the older live shingles vaccine (Zostavax) depended on exact birth date: people born on or after September 2, 1933 could get it, people born a week earlier never could. Uptake jumped from 0.01% to 47.2% across that line, creating a near-random experiment. Getting the vaccine reduced new dementia diagnoses over seven years by 3.5 percentage points, a 20% relative reduction, with a stronger effect in women.

The same design has since been repeated. In Australia, eligibility lowered new dementia diagnoses by 1.8 percentage points over 7.4 years; in Ontario, Canada, by 2.0 percentage points over 5.5 years. A follow-up in Wales found the vaccine also reduced diagnoses of mild cognitive impairment and deaths from dementia among people who already had it. The current vaccine, Shingrix, is a different (recombinant) product; in a US study comparing the two, Shingrix recipients had 17% more time free of a dementia diagnosis over six years, about 164 extra days among those who developed dementia. That comparison is observational.

The mechanism is unknown; reducing reactivation of the dormant chickenpox virus in nerves is the leading idea. A randomized trial with dementia as the outcome has not been done. Since Shingrix is already recommended for everyone 50 and older, the decision is easy. See Shingles.

GLP-1 drugs#

Large health-record studies link GLP-1 drugs such as semaglutide to lower rates of dementia in people with diabetes or obesity. The direct test failed Negative. In the EVOKE and EVOKE+ trials, 3,808 people with early Alzheimer's (mild cognitive impairment or mild dementia, average age 72) took oral semaglutide 14 mg or placebo for two years. Decline on the main clinical scale was the same in both groups (differences of −0.08 and +0.10 points), even though some Alzheimer's blood markers improved. A preventive effect years before symptoms has not been tested, but no one should take a GLP-1 drug for brain protection. Their role as possible aging drugs is covered in Longevity drugs.

Supplements#

The multivitamin result is small and comes from older adults in one trial program; a standard, inexpensive multivitamin is a reasonable choice. The supplements chapter grades the rest.

Blood tests and cognitive screening#

Alzheimer's blood tests#

Since 2025, simple blood tests can show whether Alzheimer's amyloid pathology is likely, reducing the need for PET scans or spinal taps. Most measure p-tau217, a form of the tau protein that rises in blood as amyloid builds up in the brain, sometimes as a ratio with amyloid-beta.

Every one of these is cleared only for people with symptoms, and the FDA says they are not screening or stand-alone tests. The Alzheimer's Association's 2025 guideline says that in specialty memory clinics, tests with at least 90% sensitivity and 75% specificity can be used for triage, and tests with at least 90% for both can replace PET or spinal fluid testing. It cautions that many commercial tests do not meet those thresholds and that none replaces a full clinical evaluation.

Cognitive screening#

The optimal plan Weak is a brief baseline cognitive test in your 50s, then a repeat every year from 65, or sooner if you or your family notice a change. Clinicians use tests such as the Mini-Cog or MoCA, which take a few minutes and probe memory, attention, and planning. Scores vary with education and test-taking skill, so the most useful comparison is with your own earlier score, which is why the baseline comes before any worry. A low score is a reason for a workup, not a diagnosis.

Guideline minimum: the US Preventive Services Task Force concluded in 2020 that evidence is insufficient to recommend for or against screening adults 65 and older who have no symptoms (an "I statement"). Medicare's free yearly wellness visit, available once you have had Part B for 12 months, includes a cognitive assessment; before 65, ask for the test at a routine physical. If something looks off, Medicare also covers a separate, longer cognitive assessment and care planning visit, with the usual 20% coinsurance. The care team chapter covers how to make the most of these visits.

Alzheimer's drugs#

Two antibody drugs remove amyloid from the brain and are fully approved for early Alzheimer's: lecanemab (Leqembi) and donanemab (Kisunla). They are for people with mild cognitive impairment or mild dementia who have confirmed amyloid, not for prevention.

CDR-SB (Clinical Dementia Rating, Sum of Boxes) rates memory, orientation, judgment, community affairs, home life, and personal care from 0 to 18. On it, placebo groups worsened by about 1.7 to 2.4 points over the trials, and the drugs trimmed that by about half a point to 0.7 points. Whether a difference that size is noticeable to patients and families is genuinely debated.

The main risk is ARIA (amyloid-related imaging abnormalities): brain swelling (ARIA-E) or small bleeds (ARIA-H), usually seen only on MRI but sometimes causing headache, confusion, seizures, or, rarely, death. In the donanemab trial, three deaths were judged treatment related. Risk depends sharply on APOE4.

Figures are from the Leqembi and Kisunla labels, which carry boxed warnings and call for APOE testing before treatment. In the lecanemab trial, 2 of 79 people taking an anticoagulant had a brain hemorrhage, compared with none on placebo, so blood thinners need a careful conversation.

What changed recently, as of October 2026:

• Gentler donanemab dosing. A slower dose increase cut swelling from 24.2% to 15.6% in the TRAILBLAZER-ALZ 6 trial, and the label now recommends that schedule.
• At-home lecanemab. A subcutaneous autoinjector (Leqembi Iqlik) was approved for maintenance in 2025 and, in July 2026, for starting treatment, as a weekly dose given at home. MRI monitoring is still required.

Treatment means amyloid confirmation, repeated MRIs in the first months, and regular visits. Lecanemab launched at a list price of $26,500 a year for the IV form, before scans and infusion costs. Medicare covers these drugs when the prescribing team enters patients in a registry. A memory specialist who treats many patients is the right person to weigh the trade-offs with.

Parkinson's, stroke, and head injury#

Parkinson's disease#

Parkinson's disease is a progressive loss of dopamine-producing brain cells that causes tremor, stiffness, slowness, and, often later, cognitive decline. There is no proven way to prevent it, but two things shift the odds.

• Exercise Moderate. In a meta-analysis of eight cohorts with 544,336 people, the most physically active had 21% lower risk, and those with the most moderate-to-vigorous activity 29% lower, with stronger associations in men. The SPARX3 trial is testing whether high-intensity treadmill training slows progression in people who already have it.
• Pesticides and solvents. People who spray the weed killer paraquat are more than twice as likely to develop Parkinson's as those using other pesticides. More than 70 countries ban it; it remains legal in the US, where an EPA reassessment begun in 2022 was unfinished as of March 2026, when Syngenta announced it would close its UK paraquat plant. The degreasing solvent trichloroethylene (TCE) is another suspect: Marines stationed at Camp Lejeune, whose water was contaminated with TCE and related solvents, had 70% higher odds of Parkinson's than Marines at a comparison base. The EPA finalized a ban on most TCE uses in December 2024, with some compliance dates since pushed back. If you work with pesticides or solvents, use the protective equipment on the label, and see Avoiding harm for household exposures.

GLP-1 drugs looked promising in small Parkinson's trials, but the larger phase 3 Exenatide-PD3 trial showed no clinical benefit.

Stroke#

Strokes, including small silent ones, cause vascular dementia and often add to Alzheimer's changes. Preventing stroke (blood pressure control, LDL lowering, treating atrial fibrillation, not smoking) protects the brain; the heart chapter covers prevention and the warning signs that need an ambulance.

Head injury and helmets#

In a Danish study of 2.8 million people, a history of traumatic brain injury was linked to 24% higher dementia risk, rising with each additional injury to almost three times the risk after five or more. The practical steps are fall prevention as you age (see Muscle, bone, and mobility), seat belts, and helmets. Among cyclists who crashed, helmet wearers had 51% lower odds of head injury, 69% lower odds of serious head injury, and 65% lower odds of fatal head injury Strong. Wear one for cycling, e-bikes, scooters, skiing, and climbing.

If you or your child takes a hit to the head and has symptoms (headache, confusion, dizziness, memory gaps), stop the activity and get checked by a health care provider. Repeat impacts before the brain has healed raise the risk of further injury and slow recovery.

If you're worried about your memory#

Some forgetfulness is normal. The NIA draws the line like this:

Many memory problems have causes that can be treated or reversed. The NIA lists medication side effects, depression and anxiety, sleep problems, alcohol or drug misuse, low vitamin B12, thyroid, kidney, or liver problems, head injury, and, less often, tumors, clots, or infections in the brain.

What to do:

1. Book a visit with your primary care doctor and say clearly that you want your memory evaluated. Bring someone who knows you well.
2. Bring a full list of medications and supplements, including sleep aids and allergy pills.
3. Expect a brief cognitive test, a physical exam, and blood tests (often B12, thyroid, kidney and liver function, blood count, glucose), plus a hearing and vision check if you have not had one.
4. Ask about a referral to a neurologist, geriatrician, or memory clinic if results are abnormal or unclear. That is where amyloid testing and treatment decisions happen.
5. Plan while you can. Whatever the result, update your advance directive, power of attorney, and financial access. Medicare's care planning visit covers this, and some people with dementia and their caregivers can get extra support through Medicare's GUIDE program.

A diagnosis of mild cognitive impairment is not a dementia sentence: not everyone with MCI develops Alzheimer's, and every lever in this chapter still applies.

Action plan#