Omega-3 and Cognition: How Big Is the Effect, Really?

Abstract

Omega-3 supplements are the best-selling ingredient in the cognitive category, and a 2025 dose-response meta-analysis of 58 randomised trials reports effects on cognition larger than anything else this library has covered.

That is the reason to look closely rather than to celebrate. A standardised mean difference above 1.0 for global cognition, from a nutrient, would be roughly two and a half times the effect of exercise across 2,724 trials.

It would also be graded low certainty by the authors themselves — and the abstract contains a printed inconsistency that a reader should know about before quoting any of it.

1.What Was Pooled

The analysis, published in Scientific Reports in 2025, searched PubMed, Scopus and Web of Science to December 2024, included randomised controlled trials only, and estimated the change in cognitive function per 2,000 mg per day increment of omega-3.

Fifty-eight studies met the criteria. Effects were expressed as standardised mean differences with random-effects pooling, and — importantly — certainty was assessed with GRADE rather than left to the reader.

The dose-response framing is the useful part of the design. Rather than asking whether omega-3 works, it asks how the answer changes as the dose rises, which is the question a person standing in front of a shelf actually has.

2.The Numbers, and What They Would Mean

Per 2,000 mg per day, the reported improvements were: attention 0.98, perceptual speed 0.50, language 0.98, primary memory 0.87, visuospatial function 0.86, and global cognitive abilities 1.08.

Some context is required, because these numbers do not look like nutrient effects. In our series on exercise, an umbrella review across 133 reviews and 2,724 randomised trials found a standardised mean difference of 0.42 for general cognition. The pooled effect of L-theanine on attention tasks, across the whole randomised literature, is measured in milliseconds.

A standardised mean difference of 1.08 means the average supplemented participant scored better than about 86 per cent of controls. From fish oil. Sustained across trials.

The authors' own GRADE ratings are the first check on this, and they are not reassuring about the biggest numbers. Global cognition is rated low certainty. Attention is rated low. Language is rated low. The three moderate ratings go to the smaller effects — perceptual speed at 0.50, primary memory, visuospatial function.

That pattern is worth naming. The largest effects carry the weakest certainty, which is what one expects when large effects are driven by small or heterogeneous trials.

3.The Printed Inconsistency

Attention is reported as SMD 0.98, with a confidence interval of 0.41 to 1.54. Language is reported as SMD 0.98, with a confidence interval of 0.41 to 1.54.

Two different cognitive domains, pooled from different subsets of trials with different sample sizes, producing point estimates identical to two decimal places and confidence intervals identical at both bounds.

That does not happen. The overwhelmingly likely explanation is a duplication when the abstract was assembled, and one of the two figures belongs to something else.

We are not alleging anything about the underlying analysis, which may be entirely sound, and the full text may resolve it. We are reporting that as printed, two of the six headline numbers cannot both be right, and a reader quoting the attention figure has a one-in-two chance of quoting the language figure by mistake.

This is the same class of problem we documented in two sports-nutrition trials on theanine and caffeine, and it matters for the same reason: the abstract is what travels.

4.The Finding That Complicates the Marketing

The dose-response analysis produced two non-linear relationships, and both cut against the intuition that more is better.

Episodic memory decreased as omega-3 dose rose, before appearing to rise again at higher doses — a U-shape, with a test for non-linearity at p = 0.01.

Global cognitive ability rose with dose and then declined — an inverted U, non-linearity at p = 0.008.

Taken at face value, that means there is a dose beyond which the benefit to global cognition diminishes, and a range in which episodic memory is worse than at lower intake. Neither appears on any packaging we have seen.

Taken with appropriate scepticism, non-linear dose-response curves fitted across heterogeneous trials are fragile, and these are being estimated from studies that were not designed to sit on a dose curve together. We would not build advice on either shape. We would want anyone citing the linear benefit to acknowledge that the same analysis produced these.

5.What Adjacent Reviews Show

The wider literature is more sober. A 2024 review in BMC Medicine examining n-3 polyunsaturated fatty acids and cognitive function, and a 2024 synthesis in Neuropsychopharmacology Reports covering omega-3, DHA, EPA and a medical food in Alzheimer's disease, both describe a field where results depend heavily on population, baseline status and outcome measure.

The recurring theme across all of this is that baseline matters. Effects appear more readily in people with low omega-3 status or existing impairment than in well-nourished healthy adults — which is the same boundary condition that appears throughout nutritional cognition research, and the one most often left out when a supplement is sold to the general population.

Our honest summary: omega-3 supplementation probably produces a modest cognitive benefit in some people, most plausibly those who are deficient. The specific numbers in the 2025 meta-analysis are larger than we find credible, and the authors' own certainty ratings suggest they would not defend them strongly either.

Editorial Comment

MindHeaven® does not sell omega-3 and has no product containing it. We have no commercial reason to want this literature to be strong or weak.

We cover it because it is the ingredient a customer is most likely to be taking alongside anything of ours, and because it is a clean demonstration of a habit worth acquiring: when an effect size is much larger than comparable interventions, that is information about the study rather than good news about the compound.

Exercise, across a quarter of a million participants, manages 0.42. A capsule reporting 1.08 is making an extraordinary claim, and the authors themselves grade it low certainty.

There is an authorised health claim covering DHA and brain function in the European Union, tied to a specified daily intake. We are deliberately not quoting its wording here, because we have not re-verified the current register entry and quoting claim wording from memory is precisely the error this library exists to avoid. What is worth knowing without the exact text is its shape: the authorised claim concerns maintenance of normal function in the general population, not improvement of cognition — which is considerably narrower than what omega-3 is usually sold on.

If cognitive decline is the concern, this is a conversation for a doctor rather than a shelf. Several causes of cognitive change are treatable and identifying them matters more than any supplement in this article.