Does high-dose DHA protect your memory?

Disclosure up front: I run a functional-mushroom supplement company, and "supplements for your brain" is a category I have a commercial stake in. I have no connection to this trial or its authors. The full disclosure is at the bottom.

Why this trial is different

Omega-3s and the brain have a frustrating evidence record. Observational studies, where researchers track what people eat and who declines, keep finding that more omega-3 goes with less cognitive decline. Randomized trials, where you actually hand people capsules or placebo, keep coming back mixed or null.

When a supplement trial fails, there's always a rescue argument: maybe the dose was too low, or maybe the nutrient never reached the target. Earlier DHA prevention trials used 1 g a day or less, and no trial had confirmed that a high dose actually reaches the brain in older adults who don't have dementia. So nobody could say whether "no benefit" meant the idea was wrong or the delivery was.

This trial was built to close that gap. Hussein Yassine and colleagues at the University of Southern California ran it from September 2018 to May 2024 and published it in eBioMedicine in June 2026 (registered as NCT03613844). It was funded by the National Institute on Aging and the Alzheimer's Drug Discovery Foundation, not by a supplement company.

Who got what

The team randomized 365 adults aged 55 to 80. Everyone was cognitively normal, ate little DHA (under 200 mg a day by food questionnaire), and had at least one risk factor such as obesity, high blood pressure, high cholesterol or low physical activity. Almost half (47%) carried APOE4, the strongest genetic risk factor for late-onset Alzheimer's. Mean age was 66.4.

The active group took 2,000 mg of DHA a day as five 400 mg capsules with meals, for two years. The placebo capsules held a 1:1 mix of corn and soybean oil, matched for look, taste and smell. Everyone, in both groups, also got a standard B-complex vitamin.

About half the participants agreed to lumbar punctures so the researchers could sample cerebrospinal fluid, the fluid that bathes the brain and spinal cord. That measurement was the trial's primary outcome: did the ratio of DHA to arachidonic acid (another fatty acid) in spinal fluid go up at six months?

The DHA arrived

It did, clearly. In the 167 people with spinal fluid data, the DHA-to-arachidonic-acid ratio rose by 0.17 on DHA and fell by 0.02 on placebo, a large and highly significant difference. The paper summarizes it as a 17% increase in spinal-fluid DHA. In red blood cells, the omega-3 index went from 4.9% to 11.0%. The effect was the same whether or not people carried APOE4, which surprised the authors: they'd expected carriers to absorb it less well.

So the dose question is answered. At 2 g a day, DHA gets into the central nervous system of older adults who start low.

The brain didn't change

After two years, DHA had no significant effect on hippocampal volume, the memory-related brain region that shrinks with aging and Alzheimer's. The left side shrank slightly more on DHA, the right side slightly less, and neither difference was meaningful.

Cognition looked the same. On the RBANS, a standard battery of memory, attention, language and visuospatial tests, the total score rose by 2.76 points on DHA and 2.67 on placebo. The difference was 0.09 points, with a confidence interval running from about -2.2 to +2.4. Both groups improved a little, which the authors attribute partly to how healthy and well educated this group was. APOE4 status didn't change the answer.

Safety was unremarkable: adverse events were reported by 40.3% on DHA and 40.9% on placebo, and serious adverse events were actually less common on DHA (2.8% versus 7.2%), with none attributed to the treatment.

One correction to the coverage

USC's press release, widely reposted, says the team "found an average 17% increase of DHA levels in patients' brains." That's a slight stretch. The trial measured cerebrospinal fluid, not brain tissue. Spinal fluid is a reasonable window into the central nervous system and the best you can do in living volunteers, but it is a proxy. The paper itself calls it "CNS target engagement," which is the right phrase.

The release also calls the capsules "fish oil supplements." The paper describes the intervention as DHA, and I didn't find the source oil stated in the text I read.

What the trial doesn't settle

The authors are clear about the limits, and they matter.

It lost a lot of people. Dropout reached 38% by 24 months, mostly because of COVID-19, and only 225 people completed the trial. The ones who left were more likely to be Hispanic, less educated, lower-scoring at baseline and lower in plasma DHA, which is not a random slice.

Cognition wasn't the main question. The trial was sized for the spinal-fluid outcome. Hippocampal volume was a secondary outcome and cognitive scores were exploratory. The authors estimated the expanded trial could detect a cognitive effect of about 0.35 standard deviations, so a smaller real benefit could have been missed.

The population was healthy. Hippocampal shrinkage was under 0.5% a year and cognition improved in both arms. If nobody is declining much over two years, there's little decline for a supplement to prevent. The authors suggest future trials target people with early biomarker signs of neurodegeneration.

It didn't test eating fish. One capsule, one nutrient, one dose. It says nothing about fish as part of a diet, which is where much of the observational evidence comes from.

How it fits the rest of the evidence

I ran a literature search to check whether this trial is an outlier. It isn't, but the wider picture is less uniform than "omega-3s don't work." The trial's own review of prior work says only 5 of 24 randomized DHA trials in people without dementia showed a cognitive benefit, and all 4 trials in established Alzheimer's were null. A 2023 scoping review of 78 omega-3 trials (I read the abstract only) found 43.6% reported some positive cognitive outcome, with a higher share in people who already had mild cognitive impairment, though in just 9 trials. And a 2023 meta-analysis of observational cohort studies (abstract only) linked higher dietary omega-3 intake, especially DHA, with roughly 20% lower risk of dementia or cognitive decline. Association in food records is not the same thing as a capsule changing outcomes, and this trial is the latest to show the gap.

How I'd read it

The most useful thing about this study is the measurement design. Plenty of supplement trials can't tell you whether the ingredient reached its target. This one can, and it did, and the outcome still didn't move. That's a stronger null than "we gave people pills and nothing happened," because it takes away the delivery excuse.

It doesn't prove DHA is useless for the brain. DHA is genuinely structural in neurons, and the authors' own conclusion is that supplementation alone is insufficient, not that omega-3 doesn't matter. What it does undercut is the simple pitch on a bottle: take this, it reaches your brain, your memory is protected. Reaching the brain happened. Protection, in this group over two years, didn't.

What would change my read: a well-powered trial in people with mild cognitive impairment or early biomarker changes, a longer trial than 24 months, or a trial combining DHA with diet and exercise that isolates the DHA's contribution. If one lands, I'll update this post in place.

Not medical advice. This is educational analysis, not a recommendation — a study is not a prescription. Talk to a qualified clinician before acting on anything you read here. Full disclaimer →

Disclosure, plainly: I founded and run Shroombiosis (a company I run), which formulates and sells functional-mushroom supplements, so I have a commercial stake in how people think about supplements and the brain. I have no relationship with this trial, its authors or its funders. Nothing here is sponsored and no link earns a commission; here's the full policy. A recommendation with no stake at all: for performance nutrition, Die Tryin Co. is a fellow combat-veteran-owned brand I'm glad to point people to. I don't own it and earn nothing from the link.