TallowTallow
Tallow Research

Does magnesium L-threonate cut brain age by 7.5 years? What the 6-week trial actually found

A 2026 trial reported a 7.5-year drop in estimated brain age on Magtein. What the 100-person study measured, what did not change, and who paid for it.

Emily Harper

Several blister packs of white and beige tablets spread on a blue-gray surface.

Short answer first, because it's the question behind the viral clip: the study does not show that magnesium made anyone's brain 7.5 years younger. A 2026 randomized trial of magnesium L-threonate (the branded form is Magtein) reported a 7.5-year reduction in "estimated brain cognitive age" after six weeks, but that number is calculated from a cognitive test score, not measured on a brain scan, and it comes from 100 people who were relatively young and already scoring above average. The real findings are smaller, and more interesting, than the headline.

What the study was

The trial, by Adrian Lopresti and Stephen Smith, was published in Frontiers in Nutrition in January 2026 (PubMed, full text on PMC).

  • Who: 100 adults aged 18 to 45 who reported being dissatisfied with their sleep. Average age was 37.
  • What they took: 2 grams of Magtein per day, taken in the evening about two hours before bed, or a look-alike placebo. That is 145 mg of elemental magnesium, the amount of actual magnesium in the compound.
  • How long: 6 weeks. Randomized, double-blind, placebo-controlled, with 50 people in each group.
  • How they were measured: the NIH Toolbox cognitive tests on a computer, a reasoning test (Raven's), sleep questionnaires, a reaction-time test, and an Oura Ring for sleep data.

A correction published in October 2026 fixed one sentence in the statistics section (the paper had called the significance threshold one-tailed when the tests were two-tailed). The journal says it does not change any reported p-value or conclusion (Frontiers correction).

What it found

Compared with placebo, the magnesium group improved more on:

  • Overall cognition. The NIH Total Cognition Composite rose 8.40 points on Magtein versus 5.60 on placebo (p = 0.043). The effects were larger for working memory and episodic memory.
  • Reaction time (p = 0.031).
  • A self-reported measure of sleep-related impairment (how much poor sleep interferes with daytime functioning, p = 0.043). In a subgroup with more severe sleep problems, sleep disturbance scores also improved (p = 0.031).

It did not find a difference in:

  • The Raven's reasoning test (p = 0.953).
  • Self-reported sleep disturbance in the whole group (p = 0.316), restorative sleep (p = 0.439) or general wellbeing (p = 0.436).
  • Sleep outcomes from the Oura Ring.

So the supported claim is "modest improvement in some memory and reaction-time tests, and some self-reported daytime sleep problems," not "better sleep" or "smarter."

Where the "7.5 years" comes from

The authors reported a 7.5-year reduction in estimated brain cognitive age. That figure is the paper's own conversion of cognitive test performance into an age-equivalent, not an MRI or any measure of brain structure. When a short clip says "brain age dropped 7.5 years," the claim is really "the test scores moved enough that, on an age-norm scale, they looked like a younger person's." That is a different statement, and a much weaker one.

Why to be careful with it

  • The funder makes the product. The paper states the study was funded by Threotech Inc., which also provided the investigational product and the intellectual property. The funder helped with study design. One author runs a contract research organization that has received funding from nutraceutical companies, and the other works there.
  • Young, already sharp volunteers. The authors point out that participants were relatively young with baseline cognition already about one standard deviation above the mean, and say benefits are likely to be larger in older people. That is a hypothesis in the paper, not something this trial showed.
  • Six weeks, one trial. One study of 100 people is a starting point. The authors themselves call for larger trials with a broader set of cognitive tasks.
  • The result sits near the line. The headline cognition result is p = 0.043 against a 0.05 cutoff.
  • Safety looked fine. Treatment-related adverse events were reported by 8.0% of the placebo group and 10% of the Magtein group, with no significant difference.

What this means if you're shopping for magnesium

The trial tested one specific branded compound at one dose. It says nothing about the cheaper forms (oxide, citrate, glycinate), and nothing about the generic "magnesium" in a multivitamin. Anyone selling magnesium L-threonate on the strength of this paper is borrowing a result that applies to the exact product that was tested.

It's also worth reading the label on whatever you buy: the dose of elemental magnesium, the form, and the "other ingredients" line. Supplements are not tested by the FDA before sale, so an independent lab test is the only way to know what is actually in the bottle.

Check the supplement with Tallow

Tallow scores products using independent lab results for things like heavy metals and plastics, so you can compare magnesium products before you buy them. Search for the supplement in the Tallow app, or scan the barcode, and see what the lab data says.

For a plain-English summary of the trial's findings by the supplement press, see Nutritional Outlook.