Omega-3 and Brain Health: What a 2026 Alzheimer's Study Got Wrong

Table of Contents

• The Study Landed
• What The Study Didn't Measure — The Bottle
• What The Study Didn't Measure — The Co-Factor
• What The Evidence Suggests
• References

In 2024, a video titled Your Brain on Omega-3 argued that omega-3 combined with B-vitamins was one of the most defensible nutritional strategies for brain health. In early 2026, a study landed that appeared to find the opposite [1]. Older adults who took omega-3 declined faster — on every cognitive test the researchers measured [1].

The authors' proposed mechanism was not subtle. They argued that the omega-3 in commercial fish oil is going rancid — and damaging the brain's mitochondria. But when those findings are placed alongside the existing literature, two critical variables the study never measured change the picture considerably.

The Study Landed

The paper is from Liao and colleagues, published in The Journal of Prevention of Alzheimer's Disease in 2026 [1]. They used the ADNI database — the Alzheimer's Disease Neuroimaging Initiative — which is one of the best longitudinal cognitive datasets in the world.

They took 273 people in the database who reported taking omega-3, and paired each of them with two similar non-users — matched for age, sex, baseline memory status, and even genetic Alzheimer's risk. Both groups were then followed for a median of five years. On every cognitive test the researchers ran, the omega-3 group declined faster [1].

The most immediate objection to findings like these is reverse causality. People often begin taking a supplement because they have noticed something is wrong — a slipped memory, a partner's concern. The decline already underway continues, and the supplement gets the blame.

The authors did try to address this. They looked at the omega-3 users' memory test scores and brain scans from before they ever started taking the supplement and compared them to non-users at the same time point. The two groups were tracking the same way — no difference before the supplement entered the picture [1].

That is a real check, and the authors deserve credit for running it. But the limitation they themselves concede is that this kind of pre-supplementation analysis cannot detect the subtle cognitive trajectories that prompt someone to walk into a pharmacy in the first place. These ADNI participants typically enter the study because they are already concerned about their memory. Residual reverse causality cannot be fully excluded [1].

Now look at what the imaging actually shows — because this is where the study becomes particularly interesting. The authors examined four imaging biomarkers: amyloid, tau, gray matter atrophy, and FDG-PET, which measures brain glucose metabolism. Three of the four were completely flat. There was no association between omega-3 supplementation and amyloid, tau, or gray matter volume [1].

Just one biomarker moved — the one connected to glucose. The brain runs on glucose as its primary fuel. PET scans showed the omega-3 group's brains were burning less of it over time, a pattern associated with cognitive decline [1]. But why would the brains of people who take omega-3 supplements burn less glucose?

What The Study Didn't Measure — The Bottle

Start with the molecule itself.

DHA — the omega-3 fatty acid the brain preferentially enriches — is the most chemically vulnerable fat in common dietary use. It has structural weak points where oxygen attacks readily. In direct laboratory measurements, DHA oxidises around five times faster than linoleic acid (the primary fat in seed oils) and far faster than the monounsaturated fats in olive oil [2].

Gray matter — the region of the brain responsible for the majority of cognitive processing — is around 14% DHA by weight. The brain's preferred structural fat is also its most chemically vulnerable [3].

This is the mechanism the authors of the new study explicitly propose: if the DHA in supplements is already rancid when consumed, or becomes rancid in the body, it could degrade mitochondrial function — producing exactly the kind of drop in brain energy use the FDG-PET scans showed [1].

That is a biologically coherent hypothesis. But it rests on a significant assumption: that the omega-3 supplements used by participants were already oxidised, or that participants were not metabolising them well. And this is where a paper the ADNI study itself cites becomes central to the conversation.

Ben Albert and colleagues at the University of Auckland conducted what remains one of the most cited analyses of fish oil product quality. They went into a New Zealand pharmacy, pulled 32 fish oil supplements off the shelf, and tested them for oxidation. Three had what the label said. Eighty-three percent exceeded the oxidised limit. Half exceeded the total oxidation limit [4].

To be fair, the supplement industry challenged Albert's methodology, and a follow-up study found fewer oxidised products. Claiming that all fish oil is rancid would be an overclaim. But ConsumerLab — which independently tests supplements — consistently identifies fish oil as the single most quality-variable category it assesses.

The critical limitation of the ADNI study is that it measured none of this. It did not record oxidation status of the supplements being used. It did not ask which brand participants were taking. It did not ask whether the product was in ethyl ester or triglyceride form. It did not account for whether a bottle had been sitting on a kitchen counter for 18 months at room temperature. The entire exposure variable is a single self-reported binary: omega-3 user, yes or no [1].

What that means practically: supplement quality matters. Third-party-tested fish oil products, ideally with documented low peroxide values, are a different exposure from an unlabelled bargain bottle left open in a warm cupboard. Algal-derived omega-3 is, in principle, less prone to oxidation because of how it is extracted and concentrated — although no head-to-head cognitive trial has directly compared algal against fish-derived omega-3 on brain outcomes.

But the bottle is only half the story.

What The Study Didn't Measure — The Co-Factor

The other half comes from the VITACOG trial — a body of work that is directly relevant to interpreting the ADNI result.

VITACOG was a randomised controlled trial in older adults with mild cognitive impairment. Participants received high-dose B-vitamins — folic acid, B6, and B12 — or placebo, for two years. The primary outcome was brain atrophy measured on MRI. What emerged from a series of follow-up analyses was striking: whether the B-vitamins had any effect on brain shrinkage depended almost entirely on the participants' omega-3 status at baseline [5].

In participants with low blood omega-3 levels, the B-vitamins produced no meaningful effect on brain health outcomes. In participants with adequate omega-3 levels, the B-vitamins significantly slowed brain atrophy on MRI. This interaction was replicated in an independent study the following year [6].

The picture that emerges from the VITACOG work is that omega-3 and B-vitamins are not independent variables. They appear to function as a pair: B-vitamins protect the brain more effectively when omega-3 levels are adequate, and omega-3 is more likely to benefit the brain when homocysteine is being kept in range by adequate B-vitamins.

The ADNI study did not measure homocysteine levels. It did not measure B-vitamin status in any participant.

This creates a substantial interpretive gap. It is entirely plausible — and statistically likely for at least a fraction of the 273 omega-3 users in the cohort — that some participants were B-deficient, with elevated homocysteine, at the time they were taking omega-3. By the framework established in the VITACOG literature, that is precisely the wrong nutritional environment for omega-3 to produce cognitive benefit. Omega-3 in a B-deficient context is a fundamentally different exposure from omega-3 in a person with replete B-vitamin status.

The ADNI study treated omega-3 as an isolated variable. The VITACOG evidence says it is not.

What The Evidence Suggests

The 2026 ADNI study is a meaningful signal that warrants attention. An observational finding of faster cognitive decline in omega-3 users across a well-matched, five-year follow-up is not something to dismiss.

But the two unmeasured variables — supplement quality and B-vitamin co-factor status — are not minor details. They are, based on existing mechanistic and trial data, the variables most likely to determine whether omega-3 supplementation helps, is neutral, or potentially causes harm.

The current evidence base, taken as a whole, still supports the combination of a high-quality, low-oxidation omega-3 supplement at around 1 gram of EPA/DHA per day, taken alongside adequate B-vitamins to keep homocysteine in a healthy range. This combination is supported by the VITACOG interaction analyses [5,6] and by the cardiovascular evidence base from the VITAL trial [7] and a Mayo Clinic meta-analysis [8].

What the ADNI study does usefully underscore is that supplement quality matters and that omega-3 does not function in isolation from the rest of nutritional status. Taking a third-party-tested, low-oxidation fish oil product alongside adequate B-vitamin intake is a meaningfully different intervention from simply self-reporting as an "omega-3 user."

As with any supplement, individual circumstances vary. Discussion with a healthcare provider before starting or continuing omega-3 supplementation is advisable.

References

1. https://doi.org/10.1016/j.tjpad.2026.100569

2. https://pmc.ncbi.nlm.nih.gov/articles/PMC3656724/

3. https://www.mdpi.com/2072-6643/3/5/529

4. https://www.nature.com/articles/srep07928

5. https://pubmed.ncbi.nlm.nih.gov/25877495/

6. https://pubmed.ncbi.nlm.nih.gov/26757190/

7. https://www.nejm.org/doi/full/10.1056/NEJMoa1811403

8. https://www.mayoclinicproceedings.org/article/S0025-6196%2820%2930985-X/fulltext