The ones with randomised human evidence
| Ingredient | Dose used | What was found |
|---|---|---|
| Standardised ginkgo biloba | 240 mg daily, six weeks | Better recall of appointments in 188 healthy middle-aged adults. No effect on recognition memory. No effect on dementia rates in two much larger prevention trials. |
| Lutein with zeaxanthin | 12 mg daily, twelve months | An advantage on one demanding attention task in older adults. |
| Bioavailable curcumin | Twice daily, eighteen months | Memory and attention scores ahead of placebo in forty adults aged 51 to 84. A pilot. |
| L-theanine with caffeine | Single doses, crossover | Faster and more accurate task switching, better resistance to distraction, within two hours. |
| Omega-3 (DHA) | Varies | Inconsistent. Looks better in people with low baseline intake. |
| B vitamins | Varies | Meaningful where there is a deficiency. Close to nothing where there is not. |
That is most of the defensible list. Any daily brain health supplement claiming a great deal more is either citing animal work, citing a mechanism rather than an outcome, or citing nothing.
The dose question nobody asks
Presence on a label is not the same as presence at a useful amount. The single most revealing check is to find the dose used in the trial being cited and compare it with the panel. A fashionable ingredient at a tenth of the studied quantity is there to be read, not to be absorbed.
Against a proprietary blend this check is impossible, which is generally the point.
Form matters as much as quantity
Three examples where the name alone tells you almost nothing. Ginkgo standardised to 24 percent flavone glycosides and 6 percent terpene lactones is a specification; “ginkgo 120 mg” is a weight of leaf. Curcumin on its own absorbs so poorly that a large dose of plain turmeric never reaches trial levels. And lutein needs dietary fat present to be taken up at all.
What is usually there for other reasons
Not necessarily harmful, and worth recognising. Ingredients with mechanistic plausibility but no cognitive trials — our own bamboo silica and D-glucosamine are exactly this, and we say so on every page they appear on. Ingredients present at trace amounts to lengthen a label. And proprietary complexes whose composition is undisclosed.
The comparison that still outranks the list
Everything above concerns choosing between bottles. The larger comparison is between the bottle and the alternatives: hearing, blood pressure, movement, sleep, company and diet. The evidence behind those is firmer than the evidence behind anything in the table, which is inconvenient to publish here and still true.