The Creatine Question: Brain Fuel, or Just Hype in a Tub?

In the last two weeks I've had three separate conversations about creatine, two with clients and one with one of my own coaches. From curing Alzheimer's to causing brain fog, the wellness conversation is everywhere. So I thought I'd write a piece cutting through the hype to the evidence, so you know when to use it, how much to take, and why you might reach for it for reasons other than muscle.

Today we'll cover how creatine works in the brain, the notoriously difficult blood-brain barrier, what the trials actually show, the sleep deprivation and stress data, a few myths, and of course the protocols you need.

Let's dive in.

The how

The brain weighs just 2% of your total body weight but uses 20% of your total energy (Raichle & Gusnard, 2002). It runs roughly 86 billion neurons (Azevedo et al., 2009) and more connections between them than there are stars in the Milky Way. All of that takes energy. And the reason creatine works for your muscles is the same reason it works for your neurons. Energy.

Creatine and phosphocreatine sit at the centre of how your cells manage energy. The 'energy currency' of a cell is a molecule called ATP, which the cell breaks apart to release energy where it's needed. Normally ATP is produced and broken down in a steady cycle. But when the energy demand of a cell suddenly spikes, that normal production can't keep up. This is where creatine earns its place.

In the body, creatine is mostly made in the liver and kidneys and shipped out from there. Once inside a cell it acts as a sort of battery. When ATP is plentiful, the cell stores the charge as phosphocreatine. When demand suddenly outstrips supply, that phosphocreatine regenerates ATP fast. It's a buffer for the moments when a cell needs energy over and above what the normal machinery can make.

In a muscle, that means producing force for longer, or producing greater force. In a neuron, it means holding function at full capacity when demand climbs. Rae et al. (2003) found that the brain's energy capacity directly influences its performance (doi). Supplementation reliably raises muscle creatine by around 20%, and more in people who start low (Hultman et al., 1996; Harris et al., 1992). So what's the story for the brain?

The brain's bouncer: the blood-brain barrier

The brain has unique demands, and to make sure only what's needed crosses over from the blood, evolution built a highly selective filter. Think of it as a bouncer that decides what gets in.

Getting creatine from your gut into your blood works well, and getting it into muscle works well too (Harris et al., 1992). Crossing into the brain is where it slows right down. Uptake is gated by a specific transporter (SLC6A8) that runs near saturation, and here's the part most podcasts skip: the brain makes much of its own creatine locally, so it's far less reliant on what you swallow. Dechent et al. (1999) fed volunteers 4 × 5 g a day for four weeks and lifted whole-brain creatine by just 8.7%, with a 4.7% rise in grey matter (doi). Note that was six people, so hold it loosely, but the direction is clear and it's been echoed since.

The tidy idea of "saturate the muscles and the overflow goes to the brain" is a nice story, but the data don't support it. Brain uptake is slow and heavily gated, and most studies show it takes weeks of consistent dosing to shift brain levels at all.

There's one striking exception. Gordji-Nejad et al. (2024) gave a single very large dose, 0.35 g/kg, which is about 28 g for an 80 kg adult, and managed to raise brain creatine and improve cognitive test scores in one sitting (doi). More on that shortly, because the context matters. A dose that size is also widely reported to cause stomach upset and diarrhoea, so it isn't a casual thing to try.

So what does more brain creatine actually buy you?

What the trials show

The good news is that over the past decade the interest in creatine and cognition has grown, and some well-conducted studies have been done. So will creatine carry you through the next strategic manoeuvre? The answer, like most things in human performance: it depends.

Xu et al. (2024) reviewed 16 randomised controlled trials (RCTs, the gold standard for testing supplements) covering 492 participants, looking at memory, executive function, attention, processing speed, and overall cognition (doi). They found small positive effects on memory and on the time taken on attention and processing-speed tasks, and no significant effect on executive function or overall cognition. A published correction the following year clarified that the raw attention and processing-speed scores, as opposed to the time, showed no significant effect (corrigendum), so read those two as speed, not accuracy. Digging into subgroups, the benefits clustered in people with a diagnosed disease. Graded honestly against the GRADE certainty scale, the memory finding sits at moderate certainty, and everything else at low.

Prokopidis et al. (2023) ran a meta-analysis on memory in healthy people and found a small overall effect, concentrated in older adults aged 66 to 76, with no clear effect in younger ones and no dose-response across 2.2 to 20 g a day (doi). Worth knowing: that paper drew a formal Letter to the Editor arguing its statistics double-counted and produced a false positive (doi), and the older-adult signal rested on only two small studies, one of which found nothing. So it's a real thread to pull, but not a settled one by any means. Notice too that Xu's benefit clustered in the 18 to 60 group while Prokopidis put it in the over-66s. The "who benefits by age" question is genuinely unresolved.

Avgerinos et al. (2018) reviewed RCTs in healthy people and found short-term memory and reasoning may improve, with other domains conflicting, performance in the young unchanged, and the benefit again pointing towards ageing and stressed people (doi).

Then the biggest test to date. Sandkühler et al. (2023) ran 123 participants on 5 g a day for six weeks in a pre-registered crossover design and found a small effect at best, with more reported side effects than placebo (doi). Importantly, this study was built as a direct replication of Rae's famous 2003 result, with nearly three times the people. It found strong statistical evidence against Rae's large effect being the real number. In plain terms, the improvement worked out to somewhere between 1 and 2.5 IQ points, and even that wasn't robust across every way they cut the data. The authors' honest summary was that creatine "might have a small beneficial effect", well worth investigating given how safe and cheap it is, because a small effect scaled across many people and a long time can still add up.

So my read of the evidence:

Creatine probably has a small effect in healthy people, and the field's most exciting early result didn't survive scaling up.
It's more likely to help older, stressed, or unwell people.
The safety profile means it's worth experimenting with if your gut tolerates it.

Knowing my audience, let's go to the stress and sleep deprivation data, which is early but genuinely interesting.

Creatine for sleep and stress

Conceptually, a buffer matters most when the system is under load. So the moment creatine should earn its keep is when the brain is energy-stressed. A brain running on no sleep is exactly that.

The study I mentioned above (Gordji-Nejad et al., 2024) gave sleep-deprived volunteers a single high dose and ran them through a battery of cognitive tests (doi). Compared with placebo, creatine improved word memory, cut the time taken on language, logic, and numeric tasks, sharpened reaction time, and reduced subjective fatigue. Not every measure moved, and the effect peaked around four hours after the dose. The catch, and it's a big one, is that this was 15 people. Fifteen.

A year later the same group reanalysed that same cohort with different methods and reported that the sleep-deprived brain shows a left-right asymmetry in energy use that creatine partly rebalanced (Gordji-Nejad et al., 2025, Frontiers in Neuroscience, doi). Interesting mechanism, but ultimately the same fifteen brains.

More usefully, in 2026 the group ran a fresh cohort of 29 people at a lower, more tolerable dose of 0.2 g/kg, roughly 16 g for an 80 kg adult, and again saw reduced deterioration under sleep deprivation, with improvements of up to 12% and a larger effect in women (Gordji-Nejad et al., 2026). Less dramatic than the mega-dose, but a lot easier on your stomach, and it means the sleep-deprivation finding now has a second, independent cohort behind it rather than just one.

So where does that leave us? Two small cohorts and a reanalysis, all from a single lab, and one more thing worth flagging in the spirit of honesty: the creatine was supplied by a manufacturer, which is a conflict of interest to keep in view. The signal is promising and the mechanism is plausible, but it is early. Given creatine's safety, if your gut tolerates it, it's a reasonable thing to experiment with. Just don't file it under settled.

A note for the plant-based crowd

This is where it gets interesting, and as someone who's been vegan for years, it isn't the story I expected.

The logic seems airtight. Creatine comes almost entirely from meat and fish, plant-based eaters carry lower muscle and blood creatine, so surely they'd get the biggest brain boost from topping up. And the early data played along. Rae's 2003 study was run entirely in vegetarians and found a large effect. Benton and Donohoe (2011) found a memory benefit in vegetarians but not meat-eaters (doi). Avgerinos' review saw the same lean.

Then the better data complicated it. When Solis et al. (2017) actually measured the tissue with spectroscopy, vegetarians did respond more than omnivores, but only in muscle. Brain creatine barely moved in anyone, diet regardless (doi). And Sandkühler's large trial, half of whom were vegetarians, found no vegetarian advantage at all.

The honest read: plant-based eaters can be genuinely creatine-depleted, and they respond where that depletion actually lives, in muscle. The brain looks like a separate system. It makes much of its own creatine and seems to defend its levels whatever you eat. So if you're plant-based, creatine is well worth taking for the muscle, training, and energy benefits, and I take it myself. Just don't assume your brain is running on empty and waiting to be filled. It probably isn't.

Myth busting

Load the muscles, then load the brain. As above, the brain penetration data don't support this. A muscle loading phase (20 g a day for a week) does work, it's just not necessary, since 3 to 5 g a day gets you to the same muscle levels within a few weeks (Hultman et al., 1996). And for the brain, loading buys you nothing anyway.
Creatine brain fog. No trial evidence for it, and no plausible mechanism. Most likely an attribution error, dehydration being the usual suspect.

Protocols

Protocol 1: the four-week trial (main recommendation)

What to do: creatine monohydrate, 5 g once a day. Any consistent time works. There's no good evidence that timing around meals or training changes brain uptake either way, so pick a time you won't forget and mix it into water, coffee, or a smoothie. Skip the loading phase; for the brain it does nothing.
How long: give it at least four weeks, ideally four to eight, before you judge it. Remember, brain creatine climbs over weeks, unlike muscle.
Who should skip or check first: anyone with kidney disease or reduced kidney function, anyone pregnant or breastfeeding, or anyone on medication affecting the kidneys should check with their doctor first. General information, not medical advice.
Safety: monohydrate is among the most studied supplements going and is well tolerated in healthy people. Mild stomach upset can happen; taking it with food and water helps.
Evidence: dose and duration from Sandkühler (2023) and the meta-analyses (Xu 2024, Prokopidis 2023). Expect a small effect, larger if you're older, stressed, or run down.

Protocol 2: the sleep-deprivation buffer (emerging, handle with care)

What to do: the evidence that creatine sharpens a sleep-deprived brain comes from a single lab using large one-off doses (0.35 g/kg in the first study, a gentler 0.2 g/kg since). That's a research protocol, not a wellness habit. In the trials creatine was well tolerated, but single doses this size are widely reported elsewhere to cause stomach upset and diarrhoea, so the caution stands.
Who should skip: anyone in the "check with your doctor" group above, and anyone prone to gut issues.
My take: for most people the daily 5 g in Protocol 1 is the sensible baseline, and I wouldn't reach for a 25 to 30 g single dose lightly. Just know the reason isn't that daily dosing keeps your brain topped up. It doesn't; routine supplementation barely shifts brain creatine (Solis et al., 2017). The acute effect seems to work through something different: a stressed, energy-hungry brain pulling creatine in when availability is high. So treat the single dose as an experiment for genuine high-stakes, sleep-deprived windows, not a daily tool, and fix the sleep itself where you can.
Evidence: Gordji-Nejad 2024, 2025, and 2026. Two small cohorts and a reanalysis, one lab, manufacturer-supplied product. Promising, not proven.

Creatine does not replace sleep. But it can blunt the edge of sleep deprivation and stress when they hit. I take it at around 5 to 10 g a day in my protein smoothie as a buffer against a bad night's sleep or a heavy stretch, not as a nootropic and not as a substitute for rest. And as always, track your own attention, energy, and sleep metrics along side your protocol to assess your own mileage.

Let's run a study of our own. I'd love to know if you take it. If so, how much, for how long, and have you noticed a difference? Drop a note in the comments or send me an email.

Is there a supplement or protocol you'd like me to cover? Send me an email.