In October 2003, Rae, Digney, McEwan and Bates published a short paper in the Proceedings of the Royal Society of London. Six weeks of oral creatine, they reported, improved two measures of thinking in a group of 45 people on vegetarian or vegan diets. Over the two decades that followed, cognitive claims for creatine became increasingly visible in both the scientific literature and supplement marketing aimed at the general public.
Data Manifest
- Primary Investigation: How cognitive benefits reported in creatine research compare with later clinical trials, brain-scan evidence, regulatory review and current supplement marketing.
- Key Anomalies Documented: Positive findings in selected early trials were not consistently replicated; large differences in dietary creatine were not matched by differences in measured brain creatine; a proposed 3-gram-per-day health claim was supported largely with studies using 20 grams per day.
- Primary Sources Utilised: Human intervention trials from 2003 to 2023, the Solis brain-scan study and the 2024 UKNHCC and EFSA scientific opinions.
Glossary
- Crossover trial: A study design in which every participant receives both the treatment and the placebo in turn.
- Fluid intelligence: The ability to solve unfamiliar problems and recognise patterns, measured in the 2003 study using Raven's Advanced Progressive Matrices.
- Proton magnetic resonance spectroscopy: A scanning technique used to measure chemical concentrations in living brain tissue.
The 2003 Creatine Study and Its Intended Scope
Rae and colleagues recruited 45 young adults who did not eat meat. 27 vegetarians and 18 vegans.
Their design was double-blind, placebo-controlled, and crossover, meaning every participant took both real creatine and a dummy pill in turn, with neither the participant nor the researcher knowing which was which at the time. The dose was 5 grams per day for six weeks.
Two established tests were used. The ‘Backward Digit Span’ asks participants to hear a string of numbers and repeat them in reverse, a standard measure of working memory. Raven’s Advanced Progressive Matrices is a visual pattern test used as a proxy for what psychologists call fluid intelligence. Both tests showed clear improvement in participants on creatine compared with placebo.
They did not describe vegetarians as cognitively impaired at baseline. They chose vegetarian participants because vegetarians tend to have lower creatine levels in muscle tissue, having no dietary meat source of the compound. Starting from that lower baseline made any effect of supplementation easier to detect. The paper also anticipated that most omnivores would experience some cognitive benefit from creatine as well, though that prediction was not tested in the trial itself.
Citation counts vary by database, with Timothy Bates’s Google Scholar record showing more than 670 citations.
Priority Briefings
New investigations, evidence checks, and unresolved questions from Veriarch, sent directly to your inbox.
Subsequent Clinical Trials and Null Results
In September 2008, Rawson and Lieberman ran a follow-up trial in Physiology & Behaviour. They tested 22 young adults on a battery of six neurocognitive measures, including simple reaction time, code substitution, logical reasoning, mathematical processing, running memory and memory recall. The dose was 0.03 grams per kilogram of body weight per day for six weeks. No test showed a significant effect.
Rawson and his colleagues offered a different explanation. They did not say the earlier positive results were limited to people with impaired baseline cognitive processing. They explicitly classified the Rae vegetarian subjects as not temporarily cognitively impaired, contrasting them with sleep-deprived subjects in other studies. Their proposed explanation was physiological, not cognitive. Vegetarians, having lower tissue creatine to begin with, had more room for creatine to accumulate. Omnivores, already closer to saturation, had less room to accumulate more.
The Rawson paper accumulated between 68 and 135 citations, substantially fewer than the 2003 paper.
December 2010, Benton and Donohoe reported in the British Journal of Nutrition on a trial directly comparing vegetarians and omnivores. They used a high loading dose of 20 grams per day for five days. Vegetarians showed a memory benefit. Meat eaters did not. The paper has around 229 citations.
In November 2023, Sandkühler and colleagues published the largest replication attempt to date. They tested 123 participants, split between vegetarians and omnivores, at the same 5 grams per day for six weeks that Rae had used. They did not replicate the 2003 findings for either group on the Backward Digit Span or the Raven’s Matrices test. A Bayesian statistical analysis reported strong evidence against the effect size seen in the 2003 paper.
Twenty Years of Creatine Cognition Trials
-
October 2003
Rae et al.
Forty-five vegetarian or vegan young adults received 5 grams per day for six weeks in a double-blind, placebo-controlled crossover trial. Backward Digit Span and Raven's Advanced Progressive Matrices both improved on creatine.
-
September 2008
Rawson et al.
Twenty-two young adults received 0.03 grams per kilogram per day for six weeks. None of six neurocognitive measures showed a significant effect.
-
December 2010
Benton and Donohoe
A five-day loading protocol of 20 grams per day produced a memory benefit in vegetarians but no corresponding benefit in meat eaters.
-
November 2023
Sandkühler et al.
The largest replication attempt tested 123 vegetarians and omnivores at 5 grams per day for six weeks. It did not replicate the 2003 findings on either principal cognitive test.
Brain Scan Evidence and Dietary Assumptions
Some commercial explanations of creatine’s cognitive effects rest on the assumption that lower dietary creatine leads to lower brain creatine levels and that supplementation can make up the difference.
That assumption was tested directly. Solis and colleagues published a proton magnetic resonance spectroscopy study, using a scanning technique that measures chemical concentrations in living brain tissue. The paper appeared online in the British Journal of Nutrition on 29 November 2013 and in print in April 2014.
The team scanned 14 healthy vegetarians against 14 age-matched omnivores. Dietary intake of creatine differed sharply. Vegetarians took in 0.03 grams per day. Omnivores took in 1.34 grams per day, roughly forty times as much. Yet the total creatine measured in the posterior cingulate cortex, a region routinely used in this kind of scanning work, was 5.999 international units in vegetarians and 5.917 international units in omnivores. That difference was not statistically significant. The authors concluded that dietary intake did not influence the measured brain creatine content in healthy adults under normal conditions, consistent with local synthesis in the brain.
This paper accumulated between 68 and 88 citations.
Dietary Creatine Gap, Brain Creatine Match
Solis and colleagues compared 14 healthy vegetarians with 14 age-matched omnivores.
Dietary Creatine Intake
The omnivore group consumed roughly forty times as much dietary creatine.
Measured Brain Total Creatine
The measured difference in the posterior cingulate cortex was not statistically significant.
Measured Result
No significant brain creatine difference
The authors concluded that dietary intake did not influence measured brain creatine content in healthy adults under normal conditions, consistent with local synthesis in the brain.
Regulatory Review and Health Claim Rejections
In March 2024, the German supplement manufacturer Alzchem Trostberg GmbH submitted an application to the UK Nutrition and Health Claims Committee and to the European Food Safety Authority. The applicant asked for permission to state that ‘daily creatine supplementation can contribute to improved cognitive function’. The proposed condition of use was fixed at 3 grams per day for the general healthy adult population.
The UKNHCC opinion was published on 14 August 2024. The committee reviewed ten human intervention studies provided by the applicant. Nine of them used loading doses of 20 grams per day, more than six times the proposed daily dose of 3 grams. The committee therefore excluded those studies from the evidence it considered pertinent to the proposed conditions of use.
That left one pertinent study within the requested dose range. It was Rawson 2008. Because Rawson 2008 reported no significant effects on any of its cognitive measures, the UKNHCC concluded that no cause-and-effect relationship had been established.
The EFSA opinion of 30 October 2024 was published online on 19 November 2024. The panel reviewed 21 human intervention trials and concluded that any short-term memory effects observed at 20-gram loading doses did not persist at lower continuous doses. The panel described the evidence for the biological pathway by which creatine might exert a cognitive effect in healthy adults as weak.
The Dose Problem in the 2024 Health Claim Review
| Evidence Point | Dose or Scope | Regulatory Finding |
|---|---|---|
| Alzchem Trostberg proposed the claim that daily creatine supplementation could contribute to improved cognitive function. | 3 grams per day for the general healthy adult population | This was the condition of use regulators were asked to assess. |
| Nine of ten human intervention studies submitted to UKNHCC used loading protocols. | 20 grams per day | UKNHCC excluded these studies as not pertinent to the proposed 3-gram-per-day condition of use. |
| Rawson 2008 was the remaining pertinent study within the requested dose range. | Approximately 2.2 grams per day | No significant cognitive effects were reported, and UKNHCC concluded that no cause-and-effect relationship had been established. |
| EFSA reviewed 21 human intervention trials. | Included high loading doses and lower continuous doses | Short-term memory effects at 20 grams per day did not persist at lower continuous doses. |
Current Market Trends and Industry Claims
Public-facing marketing material is available, but the brands’ internal legal or compliance records are not.
Thorne’s product pages describe its creatine as supporting ‘cognitive function’, ‘brain function’ and ‘mental clarity and focus’. The pages are aimed at athletes and the general public and do not attach a dietary qualifier. In June 2025, Thorne’s ‘Take 5 Daily’ blog published a long-form article, ‘The Research on Creatine for Brain Health Support’, by the dietitian Joel Totoro. In October 2025, the same platform published ‘Mayo Clinic: How Creatine Supports Brain Health’ by Brent Bauer.
Myprotein’s blog contains an article, ‘What Is Creatine?’, published in April 2026, and a further piece on foods for brain function and memory published in July 2026. The site cites a 2024 systematic review and meta-analysis to assert that creatine supplementation is linked to improvements in overall cognitive function and memory. The July 2026 article, credited to Richie Kirwan, PhD, asserts that vegetarians and vegans naturally maintain lower baseline levels of creatine in both their muscles and their brains. The brain point was measured directly by Solis and colleagues, whose scan data found no statistically significant difference in the posterior cingulate cortex of healthy adults.
Momentous issued a press release in June 2026 introducing what it called its Signature Spec Creatine. Its chief executive, Jeff Byers, is quoted in that release as describing creatine as ‘a proven daily essential for both body and brain’.
It is not established when the phrase ‘in vegetarians’ disappeared from the marketing materials of any specific brand.
There is no public record in the supplied material of ASA, MHRA or FTC enforcement action against these brands’ cognitive claims following the August 2024 UKNHCC opinion or the November 2024 EFSA opinion. The absence of a public record in the material held does not establish that no such action has been taken.
Commercial Explanation vs Direct Measurement
| Source | Public-Facing Position | Evidence in the Article |
|---|---|---|
| Myprotein, July 2026 | The article states that vegetarians and vegans naturally maintain lower baseline creatine levels in both their muscles and their brains. | The brain portion of that statement conflicts with the direct measurement reported by Solis and colleagues. |
| Solis et al. | Fourteen healthy vegetarians were compared with fourteen age-matched omnivores using proton magnetic resonance spectroscopy. | Total creatine in the posterior cingulate cortex was 5.999 IU in vegetarians and 5.917 IU in omnivores, with no statistically significant difference. |
Source Box
Sources include: Rae et al. (2003), ‘Oral creatine monohydrate supplementation improves brain performance: a double-blind, placebo-controlled, cross-over trial’; Rawson et al. (2008), ‘Creatine supplementation does not improve cognitive function in young adults’; Benton and Donohoe (2010), ‘The influence of creatine supplementation on the cognitive functioning of vegetarians and omnivores’; Solis et al. (2014), ‘Brain creatine depletion in vegetarians? A cross-sectional ¹H-magnetic resonance spectroscopy (¹H-MRS) study’; Sandkühler et al. (2023), ‘The effects of creatine supplementation on cognitive performance: a randomised controlled study’; the United Kingdom Nutrition and Health Claims Committee’s 2024 ‘Scientific Opinion for the substantiation of a health claim on creatine supplementation and improved cognitive function’; the European Food Safety Authority’s 2024 ‘Creatine and improvement in cognitive function: Evaluation of a health claim pursuant to Article 13(5) of Regulation (EC) No 1924/2006’; and public-facing creatine material from Thorne, Myprotein and Momentous cited in the investigation.
Claim-Source Matrix
| Core Finding | Primary Source Document | Status |
|---|---|---|
| Creatine improved Backward Digit Span and Raven's Matrices performance in 45 vegetarian or vegan young adults given 5 grams per day for six weeks. | Rae et al. (2003) | Confirmed |
| A six-week trial in 22 young adults found no significant effect across six neurocognitive measures. | Rawson et al. (2008) | Confirmed |
| The largest replication attempt did not reproduce the 2003 effects on Backward Digit Span or Raven's Matrices. | Sandkühler et al. (2023) | Confirmed |
| Healthy vegetarians and omnivores had sharply different dietary creatine intake but no significant difference in measured total brain creatine in the posterior cingulate cortex. | Solis et al. (2014) | Confirmed |
| Nine of ten studies submitted for a proposed 3-gram-per-day health claim used 20-gram-per-day loading doses; the one pertinent study within scope reported null cognitive results. | UKNHCC Scientific Opinion (2024) | Confirmed |
| A five-day loading protocol of 20 grams per day produced a memory benefit in vegetarians but no corresponding benefit in meat eaters. | Benton and Donohoe (2010) | Confirmed |
| Short-term memory effects observed at 20-gram loading doses did not persist at lower continuous doses, and the proposed biological pathway was described as weak. | EFSA Scientific Opinion (2024) | Confirmed |
What We Still Do Not Know
- The internal legal or compliance basis on which Thorne, Myprotein and Momentous currently authorise their cognitive claims following the 2024 UKNHCC and EFSA opinions.
- No supplied record establishes whether the ASA, MHRA or FTC has issued a post-August 2024 warning, enforcement notice or private correspondence to the three brands concerning creatine cognitive claims.
- Why Alzchem Trostberg submitted nine studies using 20 grams per day for a proposed claim conditioned on 3 grams per day, or which alternative studies within that dose range were considered but not filed.
- Current evidence does not show whether internal copywriting procedures at leading supplement brands instruct staff how to treat null-result trials when substantiating claims.
- The date on which the qualifier 'in vegetarians' disappeared from the marketing materials of any named leading creatine brand has not been established.
- Whether the current cognitive claims at Thorne, Myprotein and Momentous received internal compliance sign-off, by whom, and when is not established in the supplied material.
Comments (0)