Dementia Prevention: What the Neuroscience Actually Shows

Dementia Prevention: What the Neuroscience Actually Shows

Dementia is a symptom, not a disease. It sits underneath several distinct disease processes, Alzheimer's, Parkinson's, frontotemporal degeneration, and each one degrades the brain through different mechanisms that converge on the same surface complaint: memory failure, usually with a second symptom riding alongside it. In Alzheimer's, that second symptom is often a dysregulation of fear or anger. In frontotemporal dementia, language goes first.

That distinction matters because most of what people call early dementia in themselves is something else entirely. If you are losing words, hunting for names, living in tip-of-the-tongue states, that is usually processing speed, and processing speed drops when you are tired and not getting enough deep sleep. True early Alzheimer's hits episodic memory first, the first-person experiential stuff: forgetting what you wore to your own wedding, forgetting the party you planned last year. That is a hippocampal signature, not a fatigue signature.

Here is what the research has converged on over the past few years: a large fraction of dementia risk is modifiable, and the levers are specific enough to act on.

How much dementia is actually preventable?

The 2024 Lancet Commission report, authored by Livingston and colleagues, pooled global evidence and put the ceiling at 45%. Erase 14 modifiable risk factors across the whole population and you would roughly halve dementia incidence worldwide.

These factors spread across the entire life course, not just old age:

• Early life: education level (accounts for about 5% of population risk)
• Midlife: hearing loss (7%), high LDL cholesterol (7%), depression, traumatic brain injury, physical inactivity, diabetes, high blood pressure, obesity, excessive alcohol
• Late life: social isolation (5%), air pollution, untreated vision loss

Midlife carries more of these factors than early or late life combined. Two newer entries from the 2024 update are worth flagging: high LDL cholesterol in your middle years, the oxidative-stress-prone flavor of cholesterol, and untreated vision loss, which joined the long-established hearing loss as a sensory-deprivation risk.

Some factors you cannot rewind. You cannot re-educate your childhood or undo a concussion from 20 years ago. But most of these are workable, and most dementias build over 20 to 30 years before symptoms appear. That long runway is the opportunity. Watch the full breakdown of the 14 factors here: Half of Dementia Is Preventable.

A note on the "symptom-modifying versus trajectory-modifying" distinction. For most diseases of aging we have historically only had treatments that ease symptoms without bending the curve. The exciting shift is that some of these levers appear to flatten the trajectory itself, compressing illness into the last few years of life rather than the last few decades.

How does the brain clear the proteins that drive Alzheimer's?

Your brain is 2% of your body weight and burns about 20% of your energy. That metabolic activity generates waste: amyloid beta, tau fragments, inflammatory byproducts. The rest of your body has a lymphatic system to haul out the trash. The brain does not have a traditional one. It has the glymphatic system, discovered only about a dozen years ago, and it does most of its work while you sleep.

The mechanism: cerebrospinal fluid flows along perivascular channels, water channels running alongside blood vessels, mixes with the interstitial fluid between cells, dissolves out proteins and toxins, and drains the waste-laden fluid out through those same spaces. The whole exchange is gated by a water-channel protein called aquaporin-4 (AQP4), which sits at the end-feet of astrocytes wrapped around blood vessels. When AQP4 is properly polarized and concentrated at those end-feet, fluid moves efficiently. When that polarization degrades with age, flow drops.

The system runs dramatically harder at night. During deep non-REM sleep the interstitial space around the brain expands, by some estimates 60%, physically opening the drainage paths. Delta and slow-wave activity drive oscillatory fluid waves, an agitation cycle, like a washing machine shaking dirt loose.

The clinical link is direct. The proteins the glymphatic system clears, amyloid beta and tau, are the exact proteins that accumulate in Alzheimer's. If clearance fails, the proteins build up, and that buildup appears to start the cascade. Picture a New York City trash strike: two days in, there is a six-foot pile on every block, because the city never stops producing waste.

The human evidence is now solid. Benedetti and colleagues (Brain, 2025) ran a randomized crossover in healthy young people: a single night of normal sleep lowered CSF amyloid beta and tau compared to staying awake, with no change in unrelated markers. Iliff and colleagues (Nature Communications, 2026) followed with 39 adults and found that time spent in non-REM 2 and non-REM 3 sleep directly predicted brain-to-plasma clearance of these biomarkers, explaining roughly half the variance. A 2024 study by Lee (Alzheimer's & Dementia) used a glymphatic-function proxy (DTI-ALPS) and found impaired glymphatic signatures tracked with faster atrophy, more memory complaints, and reduced volume, with correlations around -0.7 in some samples.

Two practical levers fall out of this. Sleep is the obvious one, and you have control over deep sleep specifically: fast before bed, exercise in the afternoon to burn off cortisol while it is naturally low. Ignore your tracker's REM numbers, they are unreliable; deep sleep readings are usually valid. Exercise is the second. The large thigh muscles pump the lymphatic system, walking is potent here, and in Alzheimer's mouse models exercise improved AQP4 polarity and amyloid clearance.

There may be a third lever I work on clinically. I use hemoencephalography (HEG), an infrared device that trains vascular toning by leaning into the natural surges of vasomotion. The pacemaker neurons that set vasomotion of the blood vessels sit right alongside the perivascular water channels. It would be strange if training one left the other untouched. The full walkthrough of this mechanism is here: The Glymphatic System and Alzheimer's.

Does your Alzheimer's gene change what you should eat?

About 30% of people carry at least one copy of APOE4, the variant that raises Alzheimer's risk. APOE is the gene and the molecule it encodes, apolipoprotein E, which moves cholesterol and other fats around the body. Humans carry two copies, drawn from three variants: 2, 3, and 4. Carry two copies of APOE4 and your Alzheimer's risk rises 8 to 12 times the background rate. Among people who develop Alzheimer's, about 70% carry at least one copy.

It is not a mutation. APOE4 is the ancestral variant, the original version our ancestors carried hunting large game across the savannah for hundreds of thousands of years before agriculture. APOE3 and APOE2 emerged later, after diets shifted toward grains and legumes. APOE4 evolved for an animal-heavy diet and now sits in bodies eating a plant-forward, lower-protein modern diet. Evolutionary biologists call that a mismatch, the same pattern we see with lactose tolerance and amylase copy number varying by ancestral diet.

A 2025 study by Norén and colleagues at Karolinska (JAMA Network Open) followed 2,100 Swedish adults aged 60+ for 15 years, adjusting for age, sex, education, and lifestyle. At low meat intake, APOE 34 and 44 carriers had double the dementia risk of non-carriers. In the highest meat-intake group (around 870 grams a week), that excess risk vanished. APOE4 carriers eating the most meat looked like non-carriers.

Two details govern how you use this. First, it was genotype-specific: higher meat intake only helped APOE4 carriers, so this is not a go-carnivore message for everyone. Second, it was unprocessed meat. Restrict the analysis to processed meat and the benefit disappeared.

The mechanism runs through brain energy. APOE4 brains tend to run hyper-metabolic early in life, then shift to hypo-metabolism as glucose metabolism breaks down, the same direction implied by the "type 3 diabetes" framing of Alzheimer's. APOE4 alters carnitine-dependent fat oxidation, the pathway that burns fat for fuel. Animal foods are rich in the precursors, creatine, carnitine, complete amino acids, heme iron, B12, and can generate ketones nutritionally. APOE4-dominant mice upregulate ketone-burning enzymes and perform better on ketogenic diets. A supervised clinical case series showed ketogenic diets improving cognition in APOE4 carriers with mild Alzheimer's.

If you carry APOE4: do not fear quality unprocessed protein, fish, eggs, meat; avoid processed meat; prioritize omega-3s and DHA; stay within a Mediterranean pattern. If you do not carry APOE4, the standard advice holds and the heavy protein emphasis matters less. An APOE blood test is quick and tells you where you fall. This is precision brain health, and the same error of one-size-fits-all applies to diet as it does to medication or neurofeedback. The full discussion is here: Your Alzheimer's Gene Might Need a Different Diet.

Why do nutrient interventions only work in the right context?

The single most important lesson in the supplement literature for brain aging is that nutrients rarely act alone. The matrix matters.

The clearest human example is the VITACOG trial (American Journal of Clinical Nutrition), 168 adults over 70 with mild cognitive impairment, randomized to high-dose B vitamins (folate, B6, B12) or placebo for two years, with brain atrophy tracked on MRI.

First, the baseline rates. Most adults past midlife lose a quarter to half a percent of brain volume per year. MCI roughly doubles or triples that, to 1 to 1.5%. Alzheimer's pushes it to 2 to 4% per year, and the atrophy spreads beyond the medial temporal regions.

The intervention group had 40% less atrophy. That is close to erasing the MCI penalty, pulling an accelerated brain back toward a normal aging curve.

The kicker: the effect was not uniform. When researchers segmented by serum omega-3 levels, the 40% benefit held only for participants who already had high baseline omega-3s. With low omega-3s, the B vitamins did essentially nothing.

The biology is coherent. B vitamins support one-carbon metabolism and methylation and keep homocysteine down. Omega-3s, especially DHA, are structural components of neurons; at birth, a large fraction of your brain is DHA, and it remains a major membrane component. Without DHA in place, the B vitamins had no structure to act on. Together they slowed atrophy.

A follow-up review found that across 12 B-vitamin trials in older adults, only this one showed benefit, and only in subgroups. A 2022 post-hoc analysis (Gong) found B vitamins slowed decline only in people with greater left frontal atrophy at baseline, the approach-oriented region, perhaps the part of the brain leaning in most.

So do not think single bullets. Think stacks. A couple of additional offsets worth stacking: regular meditation, where Sara Lazar's work shows reduced age-related cortical thinning and regularity matters more than duration, and pulling oxidative fuel off the fire by reducing dietary sugar. The omega-3 details, including the high-absorption LPC-DHA forms and the rancidity problem with most shelf fish oil, are covered here: 40% Less Brain Atrophy.

Could lithium be an early lever against Alzheimer's?

Recent work points at a mechanism most people never associate with Alzheimer's: lithium, not as a high-dose psychiatric drug, but as a natural brain process that goes wrong.

In Alzheimer's pathology, lithium appears to get sequestered into amyloid, an early feature in both humans and mouse models. Researchers tested this by artificially lowering lithium in Alzheimer-prone mice, which produced a progressive Alzheimer-like phenotype. Then they reintroduced lithium in the water and restored memory function, with the mice running mazes better and showing reduced amyloid plaque.

The proposed mechanism runs through an overactive enzyme, GSK-3 beta, which contributes to amyloid plaque and tau tangle formation. Lithium dampens that enzyme and soothes overactivated glial cells driving an inflammatory, immune-activated environment. Amyloid is not purely demonic, it is part of the innate immune system, so the goal is balance, not elimination.

Two details make this more than a mouse curiosity. The researchers used lithium orotate, which crosses into the brain well at far lower doses than lithium carbonate. Converted to human scale, the effective dose lands in the range of the lithium orotate already sold as a supplement, single-digit milligrams. And the human psychiatric literature already shows low-dose lithium carbonate, 3 to 6 milligrams rather than the hundreds used for mood stabilization, having real effects on mood and suicidality. Historically, towns with naturally lithium-rich water show statistically lower depression rates, a signal we have known about for decades.

Standard caution applies. Mouse-to-human translation for brain and aging findings averages around 10%, though confidence rises when the mechanism is shared and well-characterized, as it appears to be here. This is not medical advice, and lithium does affect sodium handling and the EEG (with an effective washout near a week given lithium orotate's pharmacokinetics).

What this adds is a new candidate for the biohacker's pharmacopeia, sitting alongside citicoline for processing speed and the omega-3 and B-vitamin stack, as something to consider early, before degradation, if you have significant family history. The full discussion of dosing, the natural-spring evidence, and the open questions is here: Lithium & Alzheimer's: What the New Research Really Says.

What should you actually do?

The picture across all of this converges. Whether you come at Alzheimer's through impaired waste clearance, oxidative stress and glycation, vascular flow, or metabolic dysfunction, the data point the same direction, and the levers overlap.

A reasonable order of operations:

1. Protect deep sleep. This is the highest-leverage single move, because it drives glymphatic clearance of the exact proteins that accumulate. Fast before bed, exercise in the afternoon, treat sleep apnea, and double any tracker deep-sleep number that seems too low.
2. Move the large muscles. Walking pumps the lymphatic system and improves the brain's drainage machinery.
3. Know your APOE status. A quick blood test tells you whether to prioritize quality unprocessed protein and ketogenic-leaning patterns (APOE4) or to follow standard Mediterranean advice (everyone else). Avoid processed meat either way.
4. Stack nutrients in context. B vitamins need adequate omega-3s to work. Build the matrix, do not chase single inputs.
5. Address the midlife factors now. Correct hearing and vision loss, manage LDL and blood pressure, treat depression, stay active. Midlife carries more of the modifiable risk than any other window.

If you want to look at your own regulatory features, sleep, stress, attention, and how they show up in your brainwave patterns, a QEEG brain map can show you where to push. For the broader set of evidence-based moves after 50, see the top 7 brain aging interventions, and if you want to understand why your brain starts aging at 44 rather than 70, read about the critical aging window.

Most dementias build over decades and announce themselves halfway into the disease process. That long runway is exactly why these levers matter. Pick one, start tonight, and start with sleep.