Methodology
Every A, B, C, and Emerging grade on this site is assigned against one rubric, applied the same way whether an ingredient ends up in Signal State Core or gets graded out. This page is that rubric, made public. It is not a claim about any ingredient's effect, and not a claim about the unlaunched product.
Scope
Educational, methodology content: not medical advice and not a product or efficacy claim. Evidence grades describe our conservative reading of the published human research for an ingredient's studied role. They are unrelated to whether Signal State Core, once launched, will carry any specific claim, which depends on separate regulatory review.
The five statuses
A, B, C, and Emerging are ordered by strength of human evidence, not by how promising an ingredient's story sounds. Graded Out is different in kind, not just in strength: it sits off that scale entirely. Here is what each one actually requires.
Counts are the 54 ingredients currently in the library, by the status each one carries today.
Who does the review
Every grade and ingredient/article page is written and checked in-house against this published rubric by the Signal State Labs editorial team, and re-checked when a page's "Updated" date changes.
This is an editorial process, not a credentialed clinical review. We have not yet engaged a registered dietitian, pharmacist, or physician to sign off on this content, and we say so plainly rather than imply otherwise.
Why nothing is a Grade A
Most stimulant-free cognitive ingredients simply don't have the multi-trial, right-population, consistent-effect evidence base that a caffeine-grade claim would need. Rather than lower the bar to fill the top grade, we leave it empty.
Right now every active ingredient under consideration for Signal State Core v1 clears Grade C or better. No lower grade is used for a cognitive active in the formula. Grade C and Emerging ingredients in the library are either excluded from v1 entirely, or included only for a separate, well-established non-cognitive role (see the foundational layer on the Ingredient Library).
We grade ingredients out
A brand that only shows you what it sells can't be second-guessed. We publish the ingredients we researched and chose not to use, with the same evidence note we'd give anything else.
Graded out, not merely low-graded. The strongest human data is 1960s–70s pediatric hyperactivity trials, since abandoned; there are essentially no adequate positive RCTs for memory or focus in healthy adults, and the acetylcholine mechanism is not well substantiated. A Cochrane review of the related cholinergic approach (Tammenmaa-Aho 2018) found no evidence of benefit. An NTP prenatal study (DART-04) reported equivocal evidence of developmental toxicity in rats. That is not a clear-cut signal on its own, but it adds a precautionary reason against use on top of the null-efficacy picture, and the related prescription compound Deanol was withdrawn from the U.S. market over unproven efficacy. That combination (a null/failed efficacy picture plus a precautionary developmental-safety flag and a regulatory withdrawal for unproven efficacy) is what "Graded Out" means on this site.
Read the full DMAE breakdownGraded Out isn't the same as grading low. Some Grade C and Emerging ingredients stay in the library for education and comparison, because they still have a real, if limited, positive signal. DMAE carries the Graded Out status specifically because the acetylcholine mechanism it's marketed on is poorly substantiated, its strongest positive human data is decades-old pediatric research since abandoned, a Cochrane review of the related cholinergic approach found no evidence of benefit, and a modern developmental-toxicity signal argues for avoidance rather than inclusion. That is a failed efficacy picture plus a safety veto, which fails the standard outright, not just the grade.
See every researched ingredient, in or out, in the Evidence Explorer, filtered by Graded out.
Claim-specific grading
The same ingredient can honestly score differently depending on what it's being asked to do: the trial base often supports one use far better than another. Rather than invent B+ or C− notation, we keep the letters simple and scope each grade explicitly to its claim.
Stress and cortisol evidence is roughly Grade B: 60-day RCTs of standardized extract lowered perceived stress and serum cortisol versus placebo. The direct-cognition evidence is thinner and smaller, so the cognition claim doesn't inherit the stress grade.
Evidence is more consistent for symptomatic mild cognitive impairment than for healthy adults, where controlled trials are mixed and large prevention trials were negative, so the healthy-adult use case can't borrow the clinical signal.
Most positive cognition trials are in dementia or post-stroke recovery populations, and a large observational cohort flagged a stroke-risk association. That is an association, not proven causation, but it is enough to grade the healthy-adult use case conservatively.
The persuasive trials used a bovine-cortex source in memory-impaired older adults; a direct RCT of the soy-derived form actually used in modern supplements failed to beat placebo at 300 or 600 mg/day. Wrong source, wrong population: an honest Grade C. We held PS to that grade, and it's the reason we cut it from the Core formula rather than carry a below-B active for appearances.
The split can run even within a genus: an independent Cochrane review found no convincing cognitive effect for Asian Panax ginseng, which grades C for cognition here, while American ginseng (Panax quinquefolius) carries a separate, narrower mental-fatigue signal and grades B on its own page. Reading one grade across to the other would be dishonest, so we don't.
Studied dose vs. label dose
Every ingredient page states a studied dose range (what human research actually used) separately from any dose a real product might carry. A high grade at 500 mg says nothing about a product carrying 50 mg.
“Fairy dusting” (including an ingredient at a fraction of its studied dose so it can appear on a label without meaningfully contributing) is the most common way a legitimate evidence grade gets borrowed for a formula that doesn't deserve it. We treat studied dose as part of the grade, not a footnote: check the dose transparency breakdown in Ingredient count vs. studied dose, and check any ingredient's own studied range on its library page.
FAQ
A grade from A to Emerging is our own conservative reading of the published human research behind a single ingredient, for the specific claim we're discussing. It is not a product claim, a promise of effect, or a regulatory approval.
A is reserved for multiple independent human trials, in the right population, with a consistent effect and no safety flags. Nothing in our current library clears that bar. Right now the library tops out at Grade B (7 ingredients), with most entries at C or Emerging.
Graded Out is its own evidence status, not just a low grade. It means we researched an ingredient to the same standard as everything else and it actively failed that standard: either the dominant evidence is a failed replication or a null result on independent review, or a safety signal vetoes it regardless of how the efficacy data reads. It doesn't sit on the A–Emerging strength scale at all. DMAE is the clearest example: a poorly substantiated mechanism, no adequate positive human data in healthy adults, and an animal developmental-toxicity signal.
A studied dose is the amount used in the actual research behind an ingredient. A label dose is what a product actually delivers per serving. If the label dose is well below the studied dose, the ingredient's evidence grade doesn't transfer to the product. That's why every ingredient page states a studied dose range, not a recommendation.
Yes. Each ingredient page carries an updated date, and a grade is revisited as new trials, meta-analyses, or safety data appear. It isn't a one-time label.
See the standard applied to every ingredient in the Evidence Explorer, read the full Ingredient Library, or see how grading connects to what we will and won't claim in the Claims Policy.