Expert Answer
Is beta amyloid the cause of Alzheimer's?
Quick Answer
Every expert in our dataset who covers this says no, or at least not primarily. The agreement is unusually one-directional: amyloid plaque tracks poorly with how impaired someone actually is, it shows up in plenty of cognitively healthy older adults, and the drugs built to clear it have not delivered matching clinical improvement. What none of them claim is that amyloid is irrelevant. The working position across the panel is that it looks like a marker or a consequence rather than the root cause. This is an active scientific dispute, not settled fact, and it is a conversation for a neurologist rather than a self-directed decision.
What the Panel Says
With neurologist David Perlmutter, argues the amyloid hypothesis was flawed and points at a Cochrane analysis of the drugs built on it. His frame is immunometabolism - microglia, the brain's immune cells, flipping from maintenance to destruction.
“This was an analysis of the 17 top studies that evaluated the effectiveness of the drugs that do indeed lower beta amyloid in the brain.”
David Perlmutter, on Mark Hyman's show, on the record @ 19:58
Has made the case from three directions across his corpus - that plaque burden tracks poorly with symptom severity, that all three antibody drugs clear plaque without reliably improving patients, and that pulling amyloid out of blood vessel walls is itself the mechanism behind the bleeding and swelling those drugs cause.
With Dale Bredesen, reads amyloid as a protective response to a threat rather than the primary villain - part of an innate immune defense, including against herpes viruses.
With Alan Castel, brings the Nun Study to it - brains found full of plaques and tangles at autopsy whose owners had normal cognition to the end.
No direct stance on amyloid causation in the analyzed videos.
How this moved
5 updates since June 2026. Each one is a dated statement from a named expert, timecoded to the video.
-
Latest David Perlmutter, on Mark Hyman's show
A Cochrane review calls the benefit trivial
The hardest number in the thread. 17 studies, 20,342 people, about 18 months each. The benefit was described as trivial, while 20 to 25 percent of people on the drugs developed brain bleeding, swelling, or died, against a cost of roughly $40,000 a year.
on the record @ 18:47 -
Peter Attia
Clearing the marker is not treating the disease
States the lesson the anti-amyloid drugs keep teaching, that removing the marker does not reliably help the patient. Arrives the same week Hyman platforms a claim of reversing p-tau 217, an Alzheimer's blood marker, by treating underlying infection.
on the record @ 01:23 -
Gayatri Devi, on Peter Attia's show
Do not diagnose off a blood amyloid test
The sharpest turn in the thread. Amyloid shows up in plenty of asymptomatic, functionally healthy older adults, so testing for it that way manufactures patients. In the same week Patrick and Huberman land independently on the Nun Study and on amyloid as an immune response.
on the record @ 37:41 -
Mark Hyman
Reframed as a metabolic and inflammatory disease
Treats the brain as downstream of metabolism, pointing at vitamin D and homocysteine as markers that both track Alzheimer's risk and move. Attia's corpus backs the Vitacog trial result but only above 11-13 umol/L homocysteine, and Patrick notes the fix is thinner than the marker.
on the record @ 09:02 -
Matthew Walker, on Andrew Huberman's show
The clearance mechanism, not the cause
The thread opens on plumbing rather than blame. Deep sleep is when the brain clears beta amyloid and tau, which frames the proteins as waste to be removed rather than as the disease.
on the record @ 33:01
Nobody has answered
Which marker to act on first. Every expert here agrees amyloid is the wrong target and each offers a different replacement - metabolic markers, homocysteine, p-tau 217, infection, microglial inflammation. Nobody on the panel has put those candidates in a rank order, and nobody has run a head-to-head. Bryan Johnson has no coverage of the question at all.
Detailed Answer
Ask most people what causes Alzheimer's and they will say the buildup of beta amyloid in the brain. That has been the field's organizing idea for decades, and it is the idea every expert in our dataset who covers the topic now argues against.
The case against it does not rest on one study. It rests on a mismatch that keeps turning up. How much plaque a person has tracks poorly with how impaired they are.
Plaque turns up in the brains of older adults whose cognition was normal to the end, which is the finding the Nun Study is famous for and which Huberman's guest Alan Castel brings to it directly. If the protein were the disease, those brains should not exist.
The second line of evidence is the drugs. Three antibody drugs now clear amyloid from the brain effectively. On Attia's show, neurologist Gayatri Devi's account is that clearing the plaque does not reliably show up as the patient doing better.
Devi also supplies the mechanism for the harm those drugs cause, which is the part that rarely makes the coverage: the antibodies pull amyloid out of blood vessel walls, and that tearing is the fluid leaking and the bleeding. The side effect is not incidental to the drug working. It is the drug working.
What the panel puts in its place
This is where the agreement stops being useful and starts being a list. Hyman, with neurologist David Perlmutter, points at immunometabolism: microglia, the brain's own immune cells, flipping from maintenance into a destructive state. Patrick, with Dale Bredesen, reads amyloid as a protective response to a threat, produced as part of an innate immune defense.
Attia's corpus keeps returning to vascular health and to sleep as the clearance mechanism. Hyman elsewhere makes the metabolic case through vitamin D and homocysteine.
Each of those is plausible and each is cited. None of them has been tested against the others, and that is the honest state of the question. The panel is united on what the target is not, and unresolved on what it is.
What this does not mean
It does not mean amyloid is irrelevant, and none of these experts says that. It does not mean Alzheimer's is not a real disease with real pathology. And it is not a reason to stop a prescribed treatment.
The claim on the record is narrower than the headline version: amyloid looks like a marker, possibly a consequence, and a drug that removes a marker is not the same thing as a drug that treats a disease.
Related Questions
Do the new Alzheimer's drugs work?
The experts in our dataset say the clinical benefit does not match the plaque clearance. On Hyman's show, Perlmutter cites a Cochrane analysis of 17 studies covering 20,342 people that described the benefit as trivial while 20 to 25 percent developed brain bleeding or swelling. Devi, on Attia's show, reaches the same conclusion across all three antibody drugs.
Why do the amyloid drugs cause brain bleeding?
Gayatri Devi's explanation on Attia's show is mechanical. The antibodies clear amyloid from blood vessel walls, and that clearing tears the walls, which produces the fluid leaking called ARIA-E and the bleeding called ARIA-H. The side effect is a consequence of the drug doing its job, not a separate toxicity.
If not amyloid, what causes Alzheimer's?
The panel does not agree. Hyman and Perlmutter argue for immunometabolism and microglial inflammation, Patrick and Bredesen for amyloid as an immune response to an underlying threat, and Attia's corpus emphasizes vascular health and sleep-driven clearance. Nobody has ranked these against each other.
Should I get an amyloid blood test?
Devi, on Attia's show, argues against diagnosing from one, because amyloid shows up in plenty of asymptomatic, functionally healthy older adults and testing that way creates patients out of people who are fine. That is a decision for a neurologist, not a self-ordered test.
More Questions About Alzheimer's & Amyloid
Precis Pro
Ask the experts anything
The answer above is the whole answer — every expert's position, where they disagree, and the timecode behind each one.
Pro adds
- Expert-by-expert breakdowns on any topic in the corpus
- Chat across 3,700+ expert videos
- A timestamp on every claim, so you can check it yourself
$9/mo · Cancel anytime