Creatine and Cognition: What the 2024–2026 Evidence Actually Shows

What the 2024 meta-analysis actually found

The single most-cited paper behind creatine's brain reputation is Xu, Bi, Zhang and Luo (2024), published in Frontiers in Nutrition. It pooled 16 randomised controlled trials covering 492 adults with a mean age range of 20.8 to 76.4 years, all using creatine monohydrate. These are the numbers, and as far as we can tell they appear on no consumer-facing page.

Negative values are improvements here, because attention and processing speed were measured as response times — less is faster. A standardised mean difference of 0.31 is, by convention, a small effect: the sort of difference that shows up reliably in a pooled analysis of hundreds of people and is close to invisible in one person's Tuesday.

Two 2024 reviews reached opposite conclusions

This is the part that makes the topic honest, and the part the supplement blogs omit.

McMorris and colleagues confirmed that supplementation genuinely raises creatine concentrations in the brain, then found the cognitive consequences equivocal, and argued future research must actually measure brain creatine rather than assume it changed. Their title is unusually blunt for a journal paper: creatine research "fails to support the theoretical basis for an effect on cognition."

The European Food Safety Authority's NDA Panel assessed a formal health-claim application in November 2024. It accepted that creatine is well characterised and that improved cognitive function is a genuine beneficial effect — then concluded the evidence did not establish cause and effect. Its reasoning matters: acute working-memory effects appeared at 20 g a day but not at lower doses, nor with 5 g a day for six weeks; only one of ten intervention studies showed an effect on response inhibition, at 20 g a day for seven days; and three studies in patient populations did not support an effect either.

Then in 2026, a commentary in the same journal that published the meta-analysis argued Xu et al. had treated multiple correlated subtests from the same participants as independent effect sizes — a unit-of-analysis error. In some included studies seven or more memory subtests from one sample were each counted separately, so the number of observations exceeded the number of people randomised, inflating apparent precision. Crucially, the commentary notes that when the earlier Prokopidis meta-analysis was corrected the same way, its positive findings became non-significant except in older adults. That is the most decision-relevant sentence in the literature and it is on no commercial page we could find.

Dose: the number in the tub is not the number in the studies

Almost every discussion of creatine quotes 3 to 5 grams a day. That figure comes from sports nutrition, where it is the maintenance dose that keeps muscle stores topped up after loading. The brain is a different compartment, and the doses look different.

The 20 g and 8.7% figures come from the "Heads Up" review by Candow and colleagues in Sports Medicine; the loading and maintenance figures from the International Society of Sports Nutrition position stand. The gap between them is the most useful thing on this page: the dose people take is not the dose shown to change brain creatine.

Who is most likely to respond — and why it is probably not you

Creatine is not exotic. Your body makes roughly half of what it needs from arginine and glycine, and the rest comes from food — about 1 to 2 grams from a pound of uncooked beef or salmon. If you eat meat and fish regularly and sleep properly, your stores are already close to saturated and there is little headroom for a supplement to fill. That is the qualifier nobody states, and it inverts the marketing: the people most likely to respond are the people least likely to be reading a nootropic blog.

• Vegetarians and vegans — little or no dietary creatine, so the largest gap to close.
• Older adults — the Prokopidis meta-analysis found a memory effect of SMD 0.88 in adults aged 66–76 against 0.03 in adults aged 11–31.
• Acutely sleep-deprived people — Gordji-Nejad et al. (2024) gave 15 healthy adults a single 0.35 g/kg dose during 21 hours of partial sleep deprivation and reported roughly 10% better word memory and faster processing, alongside measurable changes in cerebral high-energy phosphates. Note the size: 15 people, one night, one enormous dose.
• Healthy, well-slept omnivores — the group with the least to gain, and the group the advertising is aimed at.

Why creatine is not in nootropic capsules

Here is a formulation fact that explains more than any amount of ingredient debate: 5 grams of creatine is roughly ten standard capsules. A typical size 00 capsule holds around 0.5 to 0.8 grams of powder. The brain-relevant 20-gram dose would be forty capsules a day.

That is why creatine is sold as a scoop of powder and always has been. It is also why you should be suspicious of any multi-ingredient capsule listing creatine among a dozen others: whatever is in there is a fraction of a gram, orders of magnitude below anything studied. It is there for the label, not for you.

The corollary is that capsule-format nootropics have to work with compounds active in milligrams rather than grams — caffeine at tens of milligrams, huperzine A at hundredths of a milligram, standardised extracts at a few hundred. That constraint is arithmetic, not conspiracy, and it quietly determines what can be in a capsule at all. Our review of huperzine A, ginseng and caffeine covers the compounds that clear that bar. And if you do add a creatine powder alongside a capsule, timing is one fewer thing to worry about — unlike caffeine, creatine has no acute onset and no bedtime problem, which we cover in our guide to when to take a nootropic.

What the evidence does not show

Four limits, and they matter more than the headline effect size.

No regulator accepts the claim. EFSA's 2024 opinion is the clearest statement available from a body with no product to sell: cause and effect has not been established. That is not the same as "creatine does nothing" — it means the evidence submitted was not good enough.

The headline meta-analysis is under methodological challenge. The 2026 unit-of-analysis critique is technical but consequential, and EFSA independently flagged the same concern. Until a corrected re-analysis appears, SMD 0.31 is a provisional figure, not a settled one.

The trials are small and short. 492 participants across 16 trials averages about 31 people per study; the sleep-deprivation study everyone quotes had 15. These are pilot-scale samples, and small studies systematically overstate effects.

Early clinical work is not evidence for you. Small feasibility studies giving creatine to people with diagnosed cognitive conditions have had press coverage. Those are early-stage studies in patient populations, run to see whether a larger trial is feasible. They are not evidence that creatine treats, prevents or slows anything, they say nothing about healthy adults, and a dietary supplement is not a treatment for any disease.

The unglamorous version: creatine is cheap, well tolerated, genuinely useful for muscle, and its cognitive case rests on small effects in populations that are mostly not you. The highest-confidence lever on your own memory remains the one nobody is selling — sleep and memory consolidation — followed some way behind by omega-3 DHA, which at least has a long-duration trial base.

Safety and who should be careful

Creatine monohydrate is among the more thoroughly studied supplement ingredients and is generally well tolerated in healthy adults at standard doses. Three practical points. It raises serum creatinine, a routine blood marker used to estimate kidney function — this reflects the supplement, not kidney damage, but it produces a confusing result, so tell your doctor you take it before a blood test. Anyone with existing kidney disease, or on medication affecting the kidneys, should ask a clinician first. And high single doses, particularly during loading, commonly cause bloating and gastrointestinal upset.

Frequently asked questions about creatine and cognition