MindSave Official Website › Label Desk › Eight Names On One Label, Eight Very Different Evidence Records

Eight Names On One Label, Eight Very Different Evidence Records

By The Mindsaves Editorial Desk · Reviewed September 19, 2026 · 8 min read · Reading a label

An ingredient list is typographically democratic: every name gets the same size, the same weight and the same line. The published research behind these eight is nowhere near that even, and the gaps are the interesting part.

Why a list flattens what a bibliography does not

Set eight names in a column and they acquire an equality they have not earned. Each occupies a line and takes the same amount of ink. Nothing in the typography says that one has six trials behind it in the right population while another has almost nothing outside a clinical one. Scale is about quantity and arrives printed on the panel in the carton; this unevenness is about evidence, it is free to read, and no label would ever show it.

Every amount below is the amount a published study gave. The MindSave sales page prints no amount for any of the eight names it lists, so none of these figures describes what is in the capsule.

The strongest record on this list

Bacopa monnieri is the one name here whose main evidence base is healthy adults rather than a diagnosed population. The systematic review collected six randomised controlled trials at 300 to 450 mg a day, every one of them twelve weeks long. Performance improved on nine of seventeen tests of memory free recall, with little evidence of enhancement in any other cognitive domain. The meta-analysis pooled 437 subjects, would only admit trials lasting twelve weeks or more, and found gains of 17.9 milliseconds on the Trail B test and 10.6 on choice reaction time.

A twelve-week trial gave 300 mg a day to 54 participants aged 65 and over with no clinical signs of dementia. Delayed word recall and the Stroop test improved, the divided attention and digit tasks did not, and adverse events were few and primarily stomach upset. That is what a decent record looks like here: narrow, slow, replicated, and honest about what did not move.

The weakest record on this list

Huperzine A attracts the strongest claims elsewhere and has the thinnest usable record of the eight.

Cochrane searched for randomised, parallel-group, placebo-controlled trials of huperzine A in mild cognitive impairment and found none that were eligible. That is not a null result; there was nothing to measure. The companion review in vascular dementia found one small trial of fourteen participants, no significant effect on the MMSE, and a confidence interval wide enough to include both clinically significant benefit and clinically significant harm. Vascular dementia is a diagnosed condition, and that population is a limitation rather than a credential.

The largest placebo-controlled trial randomised 210 people at 200 and 400 micrograms twice a day for at least sixteen weeks and its primary analysis failed, and here too the participants had a dementia diagnosis. An overview of six systematic reviews concluded that the evidence is insufficient to support effectiveness, while recording that none of the six reported serious side effects.

Two things about that record are worth stating rather than implying. The unit is micrograms, a thousandth of a milligram, which makes it the smallest quantity on this panel by three orders of magnitude, and writing mg where mcg belongs would be a thousand-fold error read as a recommendation. And the mechanism is what makes a medication question specific: huperzine A inhibits acetylcholinesterase reversibly and selectively, human toxicity data are described as inadequate, and animal work reports mild to moderate cholinergic side effects at therapeutic doses.

The name the literature is still arguing about

GABA is the entry where the disagreement is the honest answer.

The transport work injected radiolabelled GABA into rat parietal cortex and watched it leave with a half-life of about seventeen minutes, with efflux clearance across the blood-brain barrier at least sixteen times greater than influx clearance. That is a clear measurement pointing one way, and it is a rat microinjection study rather than a study of anything swallowed.

The systematic review of oral GABA in people gathered fourteen placebo-controlled trials at 20 to 300 mg per dose and concluded, in its own words, limited evidence for stress and very limited evidence for sleep. On the barrier it reports that there are no data showing GABA’s permeability in humans at all. Eleven of its fourteen included studies had authors employed by industrial companies, and an industry-affiliated review reaching the word limited argues against overstatement rather than for it. Researchers who think oral GABA does something generally propose that it acts on the gut’s own nervous system rather than inside the skull.

The pharmacopeia’s safety review is more settled: GABA has been associated with a transient fall in blood pressure of under ten per cent, so concurrent use with anti-hypertensive medication could conceivably raise the risk of hypotension, and no studies were available in pregnancy or lactation.

Three records with a catch in each

L-Tyrosine. The positive findings cluster in short, acutely demanding conditions. Twenty-one cadets on a combat training course took 2 g a day for five days and did better on a memory task and a tracking task, while twenty-four healthy adults aged 61 to 72 given 150 mg per kilogram showed no behavioural change, the effect turning increasingly unfavourable as age rose. The military’s graded assessment managed only a weak recommendation for cognitive stress. The catch is the setting: a week of extreme demand is not an ordinary Tuesday.

Phosphatidylserine. The catch is the raw material. The 1991 trial gave 100 mg three times a day for twelve weeks to 149 people with age-associated memory impairment and found improvement against placebo on learning and memory tasks. That material came from bovine cortex and is not what is sold today. The trial of the soy-derived form now sold gave 300 or 600 mg a day for twelve weeks to 120 people over 57 and found no significant difference from placebo on any outcome, while its safety arm found the material well tolerated.

Alpha GPC. Forty-eight healthy college-aged men took 250 or 500 mg a day for seven days with no differences on isometric strength or psychomotor vigilance. A crossover trial in twenty trained men gave 315 or 630 mg as a single dose and found the Stroop score improving while the Flanker and N-back tests and physical performance did not; its author is a paid scientific advisor to the sponsor that supplied the alpha GPC.

The safety question has two published sides and both travel together. A cohort of more than twelve million South Koreans found the 108,877 prescribed alpha GPC had a higher ten-year rate of total stroke after matching. That is prescription alpha GPC in a country that licenses it as a medicine, given to people prescribed it for cognitive complaints, and the authors ask for study of a mechanism rather than establishing one. A 2025 cohort from the same database followed 508,107 patients newly diagnosed with mild cognitive impairment and found users at lower risk of progression to dementia. Europe’s food safety authority assessed 203.7 mg a day as a supplement level and found no safety concern under the conditions it examined. Two cohorts pointing opposite ways with a regulator between them is an open question rather than a finding, and it is set out again on the side effects page.

The two nutrients, and the body that says no

Niacin and vitamin B6 are the only entries here that are nutrients with reference intakes, and their evidence on cognition is the least flattering of the eight. The National Center for Complementary and Integrative Health states that vitamin B6, vitamin B12 and folic acid have not been shown to improve cognitive functioning in older adults, which points away from the reason a B vitamin is usually there.

What the two do carry is the clearest numbers on the shelf. Supplemental niacin has an adult recommended intake of 16 mg NE for men and 14 for women and a tolerable upper intake level of 35 mg a day. Vitamin B6 has a recommended intake of 1.3 mg to age 50, then 1.7 mg for men and 1.5 for women, with an upper level of 100 mg a day from food and supplements together.

Eight records, and the thing none of them covers

An editorial ranking by weight of published evidence, not by weight in a capsule, and it says nothing about how much of anything any product holds.

One gap runs under the whole table. Nothing above is evidence about these eight taken together: no trial of the combination has been published and the finished product has never been tested in one. The only study here that gave two of these ingredients to the same people is a four-arm trial in 48 healthy volunteers which found GABA alone, and GABA with tyrosine, made task switching worse; the authors call it preliminary. The National Institute on Aging’s sentence is the ceiling over all of it: at this time, no vitamin or supplement is recommended for preventing Alzheimer’s or other forms of cognitive decline.

None of that makes an ingredient list worthless. It makes it eight names with eight different amounts of evidence behind them, which is more useful than a ranking nobody can support. The ingredients page carries each name with its study beside it. Three names on this table have a fuller page of their own: alpha GPC's two separate research records, why tyrosine's effect depends on who takes it, and why the label specifies niacinamide rather than plain niacin.