Evidence Updates
What changed, and why.
A grade that never moves isn't rigorous. It's just old. This is the dated, public record of every change we've made to an ingredient's evidence status, with the reasoning attached. Downgrades included; especially downgrades.
- Logged changes
- 118
- Review dates
- 10
- Last update
- October 24, 2026
What this log records
Changes to our evidence reads: how strong we judge the human research to be for an ingredient's studied role. Grades are educational, not product claims, and a studied dose is never a recommended dose. See The Evidence Standard for the rubric behind every decision here.
October 24, 2026
2 changes- Citation review5-HTP
Added the largest placebo-controlled trial we found: in 166 adults with inactive inflammatory bowel disease, 100 mg twice a day for 8 weeks did no better than placebo on fatigue, depression, anxiety or stress. Also added a small 12-week sleep trial in older adults that had no placebo group. C holds.
Review context: Primary-source review for our article on how long 5-HTP takes to work, October 2026.
- Citation reviewLion's Mane
Corrected our description of the 2019 Japanese fruiting-body trial. Only its general cognitive screening test improved versus placebo, at the top of that test's scale; its two memory tests showed no difference. We had said both memory tests improved. Emerging holds.
Review context: Primary-source review for our article on when to take lion's mane, October 2026.
October 12, 2026
30 changes- Safety note5-HTP
Added the Canadian monograph's driving caution and the symptoms that mean stop. The antidepressant and scleroderma contraindications are unchanged. C holds.
Review context: Regulator safety guidance re-read against the primary monographs, October 2026.
- Citation reviewAmerican Ginseng
Corrected our summary of the Bell 2022 trial. We said daily use was largely null; its primary attention measure improved after 14 days, while several memory measures did not. Naturex, the maker, funded it. C holds.
Review context: Primary-source review of cited studies, October 2026.
- Citation reviewAnthocyanins
Corrected our description of the blueberry pilot: nine people took blueberry, and the comparison group of seven was recruited separately and not randomized. Word-list recall did not improve; paired-associate learning did. C holds.
Review context: Primary-source review of cited studies, October 2026.
- Safety noteAshwagandha
Following the TGA's July 2026 alert: avoid ashwagandha with current or past liver problems, and stop and seek medical advice for signs of liver injury (yellow skin or eyes, dark urine, nausea, unusual fatigue) or for sudden vomiting or diarrhea after the first doses. C and the exclusion from Core hold.
Review context: Regulator safety guidance re-read against the primary monographs, October 2026.
- Citation reviewAshwagandha
Corrected our summary of the Zhu 2026 meta-analysis: memory and attention improved, but the executive-function result was low certainty and the visuospatial result was not significant. In the 8-week safety trial, 5.2% of people on ashwagandha and 7.8% on placebo reported side effects, a difference that was not significant. Shri Kartikeya Pharma funded it. C holds.
Review context: Primary-source review of cited studies, October 2026.
- Safety noteB-Complex (B6, B9, B12)
Added the Canadian B6 warning: stop and seek advice for numbness, tingling or pain in the hands or feet. It applies at 10 mg a day or more, counting B6 from every supplement you take. C holds.
Review context: Regulator safety guidance re-read against the primary monographs, October 2026.
- Citation reviewB-Complex (B6, B9, B12)
Corrected our description of the VITACOG trial: raised homocysteine defined a subgroup, not who could enroll, and its brain-scan result is not evidence about energy or dementia prevention. C holds.
Review context: Primary-source review of cited studies, October 2026.
- Citation reviewBacopa Monnieri
Corrected the Morgan and Stevens trial's population to adults aged 55 to 86 (we had 40 to 65) and named its extract, BacoMind. Named US Pharma Lab as the funder and capsule supplier of the Lopresti and Smith trial, whose main results were null. B holds.
Review context: Primary-source review of cited studies, October 2026.
- Safety noteCiticoline
Added the Canadian monograph's cautions: ask a practitioner before use if you take dopaminergic or cholinergic drugs, or if you have ADD or ADHD and are using it for sustained attention. C holds.
Review context: Regulator safety guidance re-read against the primary monographs, October 2026.
- Citation reviewCiticoline
Added what the Nakazaki trial actually found: its primary memory measure did not improve, while secondary memory measures did, in adults aged 50 to 85 with age-related memory complaints. Kyowa, the maker, funded it and helped design it. C holds.
Review context: Primary-source review of cited studies, October 2026.
- Safety noteCurcumin (lipidated)
Added the TGA's liver warning for medicinal curcumin: avoid it with current or past liver problems, and stop for signs of liver injury. The warning covers curcumin products generally; it gives no rate for this specific form. C holds.
Review context: Regulator safety guidance re-read against the primary monographs, October 2026.
- Safety noteGinkgo Biloba
Strengthened the surgery and blood-thinner warnings to do not use, kept other medication questions as ask-a-practitioner cautions, and added the monograph's dose-dependent limits on how long to use it without advice. C holds.
Review context: Regulator safety guidance re-read against the primary monographs, October 2026.
- Citation reviewGlycine
Removed an informal "B-minus" description; the published grade is Emerging. Also removed our claim that Ajinomoto funded most of the research, which we had not checked study by study. Emerging holds.
Review context: Primary-source review of cited studies, October 2026.
- Citation reviewHoly Basil
Corrected the reason we gave for its B grade. We said its stress trials replicated across different manufacturers. They don't: Natural Remedies developed and supplied OciBest and funded and supplied Holixer. The B grade is under review.
Review context: Primary-source review of cited studies, October 2026.
- Safety noteHuperzine-A
Added the Canadian monograph's advice to ask a practitioner before using it beyond four weeks, and the symptoms that mean stop: dizziness, unusual muscle cramps, agitation, nausea, vomiting, diarrhea or excessive salivation. The 8 to 16 weeks studied in trials is not a recommended period of self-use. C holds.
Review context: Regulator safety guidance re-read against the primary monographs, October 2026.
- Citation reviewL-Theanine
Corrected our summary of the Haskell 2008 trial. The calm rating for theanine alone was an exploratory result, not a clear effect, and most of the positive findings came from theanine combined with caffeine. B holds.
Review context: Primary-source review of cited studies, October 2026.
- Citation reviewLavender (oral)
Corrected our description of the Dold 2023 meta-analysis. We called it independent; Schwabe, the maker of Silexan, supported it financially, and its five trials enrolled people with anxiety disorders, not healthy adults under everyday stress. The B grade is under review.
Review context: Primary-source review of cited studies, October 2026.
- Citation reviewLion's Mane
Updated several trial summaries after reading the full papers. In the 2026 preprint, the registered main result was null; one task improved. In La Monica 2023, the gains were within the lion's mane group and did not beat placebo. We also reached the Černelič Bizjak paper we had listed as unavailable, and recorded its funding. Emerging holds.
Review context: Primary-source review of cited studies, October 2026.
- Citation reviewMagnesium Glycinate
Separated the one bisglycinate sleep trial from the meta-analysis we cited, which pooled other forms of magnesium in older adults. That sleep-onset estimate is not specific to glycinate. C holds.
Review context: Primary-source review of cited studies, October 2026.
- Safety noteMelatonin
Corrected the Canadian adult limit: it is 10 mg total per day, not per dose, and it is a ceiling, not a starting dose. Added the monograph's as-needed directions, no driving for five hours after a dose, and do not use in pregnancy or breastfeeding. B holds.
Review context: Regulator safety guidance re-read against the primary monographs, October 2026.
- Citation reviewMelatonin
Corrected our description of the Ferracioli-Oda meta-analysis as non-sponsor research. It disclosed academic funding and support linked to Eli Lilly, though not from a melatonin maker, and the funders had no role in the analysis. B holds.
Review context: Primary-source review of cited studies, October 2026.
- Citation reviewMultivitamin-Mineral
Corrected our description of the Docherty trial as independent: Vitabiotics, a multivitamin maker, funded it and approved its design, though not its analysis. Its main wellbeing result was null. C holds.
Review context: Primary-source review of cited studies, October 2026.
- Citation reviewPanax Ginseng
Removed a stale reference saying American ginseng is graded B for mental fatigue. It is graded C, and it is a different species with its own profile. C holds.
Review context: Primary-source review of cited studies, October 2026.
- Citation reviewPhosphatidylserine
Clarified that the Kato trial tested soy-derived PS, and that its main memory measure did not improve across the whole group. It says nothing direct about sunflower PS. C holds.
Review context: Primary-source review of cited studies, October 2026.
- Safety noteRhodiola Rosea
Strengthened the bipolar warning to do not use, and added the Canadian monograph's cautions: ask a practitioner first if you take antidepressants, HRT or birth control, stop if irritability or insomnia develops, and avoid taking it at bedtime. The profile now separates research doses from the monograph's daily and per-dose limits. C holds.
Review context: Regulator safety guidance re-read against the primary monographs, October 2026.
- Citation reviewRhodiola Rosea
Corrected our 17 September summary of the 2025 crossover trial. We said its Stroop result was compared against no capsule, not a blinded placebo. That was wrong: both doses improved total Stroop scores against blinded placebo, though individual subtests were mixed. The trial was registered after the fact, used a different extract from the early fatigue trials and studied 27 athletes. C holds while we reassess this outcome.
Review context: Primary-source review of cited studies, October 2026.
- Citation reviewRhodiola Rosea
Corrected our description of the 2012 Ishaque review: it covered 10 randomized trials and one non-randomized controlled trial, not 11 randomized trials, and its fatigue results were mixed rather than consistent. C holds.
Review context: Primary-source review of cited studies, October 2026.
- Citation reviewSpearmint Extract
Corrected our description of the Falcone trial: 900 mg a day for 90 days, with agility tests on days 7, 30 and 90. It was not a same-day attention study. Kemin, the maker, funded it. C holds.
Review context: Primary-source review of cited studies, October 2026.
- Citation reviewVitamin K2 (MK-7)
Clarified that the bone meta-analysis pooled MK-4 and MK-7, so it is not an MK-7 result, and that the artery-calcification analysis covered 304 of the trial's 389 men. C holds.
Review context: Primary-source review of cited studies, October 2026.
- Citation reviewZynamite
Removed two claims the cited trial did not test: that it works through COMT inhibition in people, and that it avoids the heart-rate effects of stimulants. The trial was sponsored by Nektium, the maker. C holds.
Review context: Primary-source review of cited studies, October 2026.
September 17, 2026
20 changes- Citation reviewAnthocyanins
Added a 2026 meta-analysis restricted to healthy participants (14 trials, 632 people): blueberry intake improved flow-mediated dilation but produced no significant change in executive function or mood. It is the clearest counterweight to the large 2025 meta-analysis that put this profile under monitoring, and it points the same way as the six-month blueberry trial already cited. C holds, still under monitoring.
Review context: Reassessed during our September 2026 deep ingredient and formula review.
- Safety noteAshwagandha
Added the two largest pieces of new evidence and kept the grade where it is. An independent meta-analysis of 20 trials in 1,249 adults now reports consistent cognitive gains across memory, attention, executive function and visuospatial ability, which is a stronger efficacy base than this profile had. A 1,002-person, 8-week randomized safety trial found no change in liver-function tests and fewer adverse events than placebo. Neither moves the grade: the cap here is the LiverTox rating and the TGA alert, and an 8-week trial does not test the rare idiosyncratic injury they describe.
Review context: Reassessed during our September 2026 deep ingredient and formula review.
- Citation reviewBacopa Monnieri
Checked whether the trigger for moving this grade to C had fired: it has not. No second adequately powered null trial at 300 mg has appeared since the 2025 one this profile already cites. Added a disclosure to that citation: the trial was run by a contract research organisation whose managing director is the first author, and its competing-interests statement refers to a funder it does not name. B holds, still under monitoring.
Review context: Reassessed during our September 2026 deep ingredient and formula review.
- Citation reviewCiticoline
Re-checked both trials this grade is waiting on. NCT05000190 (PepsiCo, 250 mg) and NCT06906848 (Kirin) still had no posted results on 2026-09-17, and neither registry record has been updated since 2022 and 2025 respectively. No new healthy-adult citicoline trial has appeared. C holds, still under monitoring.
Review context: Reassessed during our September 2026 deep ingredient and formula review.
- Citation reviewCiticoline
Corrected how three trials' sponsorship is described. Spiers 1996 was counted as independent healthy-adult evidence; Archives of Neurology published a correction supplying a financial-disclosure statement the paper had omitted, setting out authors' financial ties to Interneuron and Ferrer, companies with commercial interests in citicoline. ICTUS, one of the two large null trials, was described as independent; it was funded by Ferrer, a citicoline manufacturer. Nakazaki 2021 was described as manufacturer-run; Kyowa funded it and took part in its design, but the paper reports Kyowa was not involved in the intervention, data collection or statistical analysis. Knott 2015 remains the independent healthy-adult trial. The grade is unchanged at C.
Review context: Source audit after the September 2026 re-review.
- Citation reviewCocoa Flavanols
Added the strongest positive synthesis this ingredient has — a 2026 GRADE-assessed meta-analysis reporting faster Trail Making Test times and better verbal fluency at 500 mg flavanols/day or more — and flagged this grade for monitoring rather than moving it. The paper is paywalled, so whether the COSMOS cognitive sub-studies are among its 16 trials is unknown. C holds either way: a positive pooled estimate does not displace the null result in the largest and longest trial programme.
Review context: Reassessed during our September 2026 deep ingredient and formula review.
- Citation reviewCreatine Monohydrate
Added two 2026 sources that separate the performance claim from the cognitive one: a trial in young active adults where creatine improved Wingate cycling performance but not cognition, and a scoping review that searched 8,317 records for creatine trials in dementia and found no randomized controlled trials at all. C holds.
Review context: Reassessed during our September 2026 deep ingredient and formula review.
- Citation reviewErgothioneine
Added two sources and the reason neither moves the grade. A 12-week trial in 80 healthy older adults reported better delayed recall and lower negative mood, but it fed participants oyster mushrooms rather than ergothioneine, so the result cannot be attributed to the compound; and a 2025 systematic review of 19 studies found the case still resting mainly on observational and preclinical work. Emerging holds.
Review context: Reassessed during our September 2026 deep ingredient and formula review.
- Citation reviewGinkgo Biloba
Re-grounded this profile on the 2026 Cochrane review of 82 studies in 10,613 participants, which supersedes the older reviews cited here. In mild cognitive impairment ginkgo probably has little or no effect at six months on global status, cognition or daily function; in dementia there may be small-to-moderate benefits at low certainty with very high heterogeneity; in people with subjective cognitive complaints the effect is uncertain and adverse events may be more common than on placebo. C holds for a healthy-adult use case.
Review context: Reassessed during our September 2026 deep ingredient and formula review.
Downgraded from a conditional B to C on the same rubric point as vitamin C. The meta-analysis behind this grade found the clearest gains in participants who were anaemic or iron-deficient to begin with, and this profile already stated that the benefit does not extend to iron-replete adults and that iron should be screened for rather than included by default. Evidence concentrated in a deficient population is a C under our standard. Nothing about the cautions or the screening guidance changes.
Review context: Reassessed during our September 2026 deep ingredient and formula review.
- Citation reviewL-Theanine
Corrected a one-sided description. This profile reported that the 2025 five-trial meta-analysis found no effect on simple reaction time or the Stroop test, which is true but leaves out that the same analysis found a dose-dependent improvement in rapid visual information processing and recognition reaction time. Both findings are now stated and the meta-analysis is cited in full, alongside a larger 2025 tea review whose Lipton and Unilever funding is disclosed. B holds.
Review context: Reassessed during our September 2026 deep ingredient and formula review.
- Citation reviewLion's Mane
Deepened the lion's mane review to every human trial we could find, rather than the three this profile cited. Four more are now cited: a 31-person 12-week fruiting-body trial (Saitsu 2019), an acute 1 g crossover run by a contract research organisation (La Monica 2023), a null trial at 10 g/day for four weeks (Grozier 2022), and a 33-person pilot of an erinacine-A-standardized extract (Černelič Bizjak 2024) in which cognition was a secondary outcome and the gain held only after adjusting for baseline score, age, sex and a biomarker. The 2026 supplier-funded trial was still an unreviewed preprint when we re-checked, and three further trials have completed without posting results. The trials that report a benefit each use a different preparation, dose and outcome measure, so none replicates another. Emerging holds.
Review context: Reassessed during our September 2026 deep ingredient and formula review.
- Citation reviewMagnesium Glycinate
Added a 2026 systematic review of every randomized trial of magnesium as a stand-alone sleep intervention: results were inconsistent across subjective scales, EEG, actigraphy and wearables, certainty was low to very low, and the review concludes magnesium is not supported as a routine treatment for insomnia. C holds, on firmer ground than before.
Review context: Reassessed during our September 2026 deep ingredient and formula review.
- Citation reviewMagnesium L-Threonate
Funding and conflicts now disclosed for the three positive trials behind this grade: Liu 2016 (funded by Neurocentria, which analysed the cognitive outcomes and wrote the paper), Hausenblas 2024 (commissioned and funded by AIDP Inc., with three of six authors paid by AIDP) and Lopresti & Smith 2026 (funded by the ingredient's owner). A 2025 corrigendum to Hausenblas 2024 supplied its omitted competing-interest declaration and disclosed that the trial was registered retrospectively. Two more trials are now cited: Zhang 2022, a commercially funded combination-formula trial that cannot isolate this ingredient, and Ni 2023, a publicly funded trial after breast-cancer surgery that found no significant difference from placebo in sleep, mood or cognition. The grade is unchanged at C.
Review context: Source audit after the September 2026 re-review.
- Safety noteMelatonin
Updated the heart-failure monitoring note. The 2025 conference abstract still has no peer-reviewed publication, and the peer-reviewed work published since points the other way: an NHANES analysis found no significant association between recent melatonin use and prevalent cardiovascular disease or heart failure once baseline differences were accounted for, and a meta-analysis of 14 randomized trials in cardiovascular patients found no harm signal. Neither settles the question, but together they make the abstract weaker evidence than when we first logged it. B holds.
Review context: Reassessed during our September 2026 deep ingredient and formula review.
- Citation reviewMultivitamin-Mineral
Added a 228-person, 12-week trial in adults over 70 that missed its primary wellbeing outcome, with positive findings confined to sex-split secondary measures. It supports reading this grade as C rather than higher. C holds.
Review context: Reassessed during our September 2026 deep ingredient and formula review.
- Citation reviewRhodiola Rosea
Added the newest cognitive data and the reason it does not move the grade: a 2025 crossover trial in 27 resistance-trained adults reports better Stroop performance against a no-capsule condition rather than against its own placebo, so the cognitive comparison is not blinded. C holds.
Review context: Reassessed during our September 2026 deep ingredient and formula review.
- Citation reviewSaffron
Added a 2026 GRADE-assessed meta-analysis of 34 trials in 1,769 adults and sharpened how this profile describes the mood evidence. Self-rated depression and anxiety scores improved; the clinician-rated scales and the POMS did not, and heterogeneity was above 90% on both pooled outcomes. The mood evidence is now described as stronger on self-report than on clinician-rated measures. The cognition grade is unchanged at C.
Review context: Reassessed during our September 2026 deep ingredient and formula review.
Downgraded from B to C on our own rubric rather than on new evidence. The rubric places evidence that sits in a deficient population under C and reserves B for effects tied to a context such as stress, fatigue or sleep loss. Vitamin C is the first case: the supporting review is largely observational, with 5 randomized trials among 50 studies, and the cleanest supporting trial enrolled only adults whose serum vitamin C was already below 50 micromoles per litre. The benefit described — correcting an inadequate baseline — has not changed; the grade now matches it.
Review context: Reassessed during our September 2026 deep ingredient and formula review.
- Citation reviewVitamin E
Added a 2026 systematic review of 43 clinical studies of vitamin E and cognition. Only one tested a vitamin E supplement on its own; the rest measured dietary intake or multivitamins containing vitamin E. That is a further reason the dietary associations cannot be read across to supplementing, and it leaves the safety veto behind this status untouched.
Review context: Reassessed during our September 2026 deep ingredient and formula review.
September 14, 2026
14 changes- Citation reviewAnthocyanins
Held at C and placed under monitoring after a 2025 meta-analysis of 59 RCTs reported a cognitive benefit; the pooled trials are heterogeneous, and a six-month blueberry RCT found no cognitive effect.
Review context: Reassessed during our September 2026 evidence re-review.
- Citation reviewB-Complex (B6, B9, B12)
Cognition grade held at C. Added a meta-analysis showing a small reduction in perceived stress with B vitamins in healthy and at-risk adults, with no effect on anxiety or depressive symptoms.
Review context: Reassessed during our September 2026 evidence re-review.
- Citation reviewBacopa Monnieri
Held at B and placed under monitoring. A 2025 twelve-week trial at 300 mg/day found no difference from placebo on verbal learning, attention or working memory, and a 2026 network meta-analysis placed the clearest working-memory gains at 600 mg/day or more.
Review context: Reassessed during our September 2026 evidence re-review.
Downgraded from B to C. The trials behind the branded form's attention claim are all funded by or co-authored with its manufacturer; the healthy-adult work from outside them is small and concentrated in low baseline performers (Knott 2015, and Spiers 1996, whose authors' financial ties to citicoline companies were disclosed in a later correction); EFSA's 2024 review found no cause-and-effect relationship for memory; and two large RCTs were null (ICTUS in stroke, funded by the citicoline manufacturer Ferrer; COBRIT in brain injury, sponsored by the NIH). Two completed healthy-adult trials had no posted results when we checked.
Review context: Reassessed during our September 2026 evidence re-review.
Downgraded from B to C. The COSMOS cognitive sub-studies found no overall cognitive benefit from 500 mg/day of cocoa flavanols, and the memory signal was confined to people with the lowest baseline diet quality — so the acute blood-flow finding no longer carries a B.
Review context: Reassessed during our September 2026 evidence re-review.
- Citation reviewCreatine Monohydrate
Reviewed for an upgrade and held at C. EFSA's 2024 opinion found no established cognitive effect, a corrected re-analysis of the main memory meta-analysis was not significant outside older adults, and the largest healthy-adult trial was borderline.
Review context: Reassessed during our September 2026 evidence re-review.
New profile at Emerging. The larger randomized trial, funded by the ingredient's maker, showed no between-group benefit on its primary memory outcome, and the other is a 19-person pilot in mild cognitive impairment; the rest of the case is observational and animal research.
Review context: Added during our September 2026 evidence re-review.
- Citation reviewL-Theanine
Held at B with two new meta-analyses added: a 2026 review of 31 RCTs found a single 200 mg dose improved attention and reaction time, and a 2025 review found improvements in self-rated, but not objectively measured, sleep.
Review context: Reassessed during our September 2026 evidence re-review.
- Citation reviewLion's Mane
Reviewed for an upgrade and held at Emerging. A 2025 acute trial found no overall effect, and the strongest new result is an unreviewed, sponsor-funded 2026 preprint in which one of several cognitive tasks improved.
Review context: Reassessed during our September 2026 evidence re-review.
Upgraded from Emerging to C. Randomized trials published in 2024 and 2026 reported modest improvements in cognition or self-rated sleep, alongside an earlier small cognition trial, but all three were commercially funded and none has been independently replicated, so it stops at C.
Review context: Reassessed during our September 2026 evidence re-review.
- Safety noteMelatonin
Monitoring note added. A 2025 American Heart Association conference abstract associated a year or more of prescribed melatonin use with higher heart-failure rates in adults with insomnia. It is observational, not peer-reviewed, and captures only prescribed use, so we record it as information to monitor rather than a safety signal that caps the grade; the grade is unchanged.
Review context: Reassessed during our September 2026 evidence re-review.
New profile at claim-specific Grade C. The COSMOS meta-analysis found small cognitive benefits in adults 60 and older, but the evidence sits in older adults and comes from one parent trial, which our standard does not count as independent replication.
Review context: Added during our September 2026 evidence re-review.
- Citation reviewRhodiola Rosea
Grade unchanged at C. Added product-authenticity evidence: a UK market survey found some Rhodiola products contained no rosavins and most held less than registered products, so identity testing now appears in the label guidance.
Review context: Reassessed during our September 2026 evidence re-review.
- Citation reviewSaffron
Cognition grade held at C. Two 2025 trials added: a 202-person trial found a modest benefit on low-mood scores, and a smaller publicly funded trial missed its primary mood endpoint. Effects are shrinking as trials improve, so the mood evidence stays short of A.
Review context: Reassessed during our September 2026 evidence re-review.
July 12, 2026
10 changesNew profile at Grade C, with explicit serotonergic-interaction cautions.
Review context: Added while expanding our sleep and recovery coverage.
New profile at Grade C for sleep/relaxation, on limited human data.
Review context: Added while expanding our sleep and recovery coverage.
New profile at Grade C for sleep/mood-adjacent roles, with conservative interaction cautions.
Review context: Added while expanding our sleep and recovery coverage.
New profile at Grade C — a well-tolerated, well-absorbed magnesium form with limited direct human cognition/sleep trials at this specific form.
Review context: Added while expanding our sleep, recovery, and micronutrient coverage.
New profile, published at Grade B on independent (non-sponsor) evidence: a meta-analysis of 19 randomized trials found melatonin reduced sleep-onset latency and modestly increased total sleep time. Graded for sleep onset specifically — not as a daytime nootropic.
Review context: Added while expanding our sleep and recovery coverage and introducing the Hormone category.
New profile at Grade C for sleep/relaxation, on limited human data.
Review context: Added while expanding our sleep and recovery coverage.
Published straight into Graded Out on a safety-and-null-evidence basis. Randomized trials in replete older adults show no cognitive or dementia-prevention benefit (SELECT, PREADViSE), and the window between adequate and excessive intake is narrow.
Review context: Part of a broader review of our micronutrient coverage.
New profile at Grade C for sleep; trials are mixed and heterogeneous in preparation.
Review context: Added while expanding our sleep and recovery coverage.
Published straight into Graded Out on a safety veto. In replete adults the cognitive-prevention evidence is null (PREADViSE), and high-dose supplementation carries a documented safety signal — repletion of a true deficiency is not a nootropic effect.
Review context: Part of a broader review of our micronutrient coverage.
New profile at Grade C — essential for cognition when deficient, but supplementation in replete adults is a repletion story, not a nootropic one.
Review context: Part of a broader review of our micronutrient coverage.
July 11, 2026
6 changesDowngraded from B to C. Acute cognitive effects appear in small trials but do not replicate consistently enough across doses and populations to hold a B.
Review context: Reassessed during a full-library evidence and data refresh.
Downgraded from B to C. The cognition signal comes from a small number of trials with a specific bioavailable form; the broader, independent replication needed for a B is not there yet.
Review context: Reassessed during a full-library evidence and data refresh.
Downgraded from B to C. Cognitive-benefit trials are promising but limited in number and largely in specific populations, short of the consistent replication a B requires.
Review context: Reassessed during a full-library evidence and data refresh.
Downgraded from B to C. Acute dosing raises cerebral blood flow but did not improve cognition in healthy adults; the only positive cognition data come from a narrower postmenopausal-women population.
Review context: Reassessed during a full-library evidence and data refresh.
Downgraded from B to C. The working-memory effect is real in the trials that exist but rests on a single branded extract with limited independent replication.
Review context: Reassessed during a full-library evidence and data refresh.
Downgraded from B to C. The attention/memory result rests on a single acute-dose trial in healthy adults; a B needs independent replication that does not yet exist.
Review context: Reassessed during a full-library evidence and data refresh.
July 8, 2026
2 changesInitial profile published at Grade C for cognition specifically: a 24-trial meta-analysis found only a small effect concentrated in deficient groups, with no cognitive lift in replete adults; its bone, calcium, and immune roles grade stronger separately.
Initial profile published at Grade C: a 16-RCT meta-analysis supports bone and works-with-D roles in postmenopausal women, but the trial base is still limited and mixed.
July 7, 2026
19 changesInitial profile published at Grade C: most positive cognition trials are in dementia or post-stroke populations with sparse healthy-adult data; the profile carries a stroke-risk association caveat.
Initial profile published straight into Graded Out: the sponsor-run Madison Memory Study failed its pre-specified primary endpoints, and an orally-digested protein has little plausible route to the brain.
Initial profile published at Grade C: 60-day RCTs lower perceived stress and cortisol, but direct healthy-adult cognition signals remain thin; the profile carries a liver-safety caution.
Initial profile published at Grade B (qualified): acute cerebral-blood-flow and attention effects are demonstrated, while the large chronic trial (COSMOS) was null on its primary global-cognition endpoint.
Graded out, not merely low-graded. The strongest human data are 1960s–70s pediatric 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.
Review context: Reassessed during our full-library evidence audit.
Initial profile published at Grade Emerging: a replicated pre-sleep sleep-quality and next-day-alertness signal, held below B because the trial base is concentrated with a single amino-acid manufacturer.
Initial profile published at Grade B: stress and cortisol RCTs replicate across more than one manufacturer's standardized extract — cross-sponsor replication rather than repeated single-sponsor trials.
Initial profile published at Grade B (conditional): the cognition benefit is specific to correcting iron deficiency or iron-deficiency anemia, not a general enhancer in replete adults.
Initial profile published at Grade B: multiple small human trials show acute calm and attention effects, often studied alongside caffeine, with modest and variable effect sizes.
Initial profile published at Grade B: five or more double-blind RCTs plus independent meta-analyses support the Silexan extract on anxiety-scale endpoints, with no sedation or dependence signal on its studied endpoints.
Initial profile published at Grade B: replicated acute calm-under-load findings, extended beyond single-dose use by a 2023 subchronic RCT from a different sponsor.
Downgraded from B to C. Acute cognitive effects appear in some trials but replicate inconsistently across doses and extracts, so the evidence is limited rather than moderate.
Review context: Reassessed during our full-library evidence audit.
Downgraded from B to C. The stronger data sit in age-related decline and soy-derived PS trials; the healthy-adult, focus-specific evidence at label-realistic doses is thinner than a B implies.
Review context: Reassessed during our full-library evidence audit.
Initial profile published at Grade Emerging: positive human trials are all single-sponsor on one branded ingredient, with no independent replication published yet.
Downgraded from B to C. The human trials are mostly small, short, and heterogeneous in outcome measure, and several are sponsor-linked — not enough consistent, independent replication to hold a B.
Review context: Reassessed during our full-library evidence audit, when we introduced the Graded Out status.
Initial profile published at Grade Emerging: small, sponsor-linked Zembrin trials with limited independent replication.
Initial profile published at Grade Emerging: human data is biomarker-level (brain phospholipid precursors on 31P-MRS), not a cognition-outcome trial in healthy adults.
Initial profile published straight into Graded Out on legal and safety grounds: the FDA has stated vinpocetine is not a lawful dietary ingredient and flagged a reproductive-harm risk.
Initial profile published at Grade B: a status-correction story — benefit is tied to correcting an inadequate baseline rather than a universal cognitive lift in already-replete adults.
July 5, 2026
10 changesInitial profile published at Grade C: the only Cochrane cognition review is in a dementia population, with no dedicated evidence base in healthy, unimpaired adults.
Initial profile published at Grade C: benefit is concentrated in older adults with elevated homocysteine (VITACOG), while a ~22,000-person meta-analysis found no cognitive effect in unselected adults.
Initial profile published at Grade C: a 6-RCT systematic review suggests short-term memory and reasoning gains concentrated in stressed or low-baseline states.
Initial profile published at Grade C: healthy-adult enhancement trials are mixed and large prevention trials (GEM, GuidAge) were negative; the evidence is more consistent only for symptomatic mild cognitive impairment.
Initial profile published at Grade C: positive RCTs cluster in dementia patients and high-risk-of-bias student cohorts, and the signal largely disappears in healthy adults.
Initial profile published at Grade B: acute-dose RCTs improve working memory and cognitive flexibility mainly when catecholamines are depleted by stress or demand, not at rested baseline.
Initial profile published at Grade Emerging: the mechanism rests on rodent work (Slutsky 2010), with only a small developer-linked human trial to date.
Initial profile published at Grade C for cognition in healthy adults: a Cochrane review and AREDS2 found no cognitive benefit; its cardiovascular and triglyceride roles grade stronger separately.
Initial profile published at Grade C: RCTs support mood, stress, and sleep at 28 mg/day affron, but direct healthy-adult cognition data are thin and mostly extrapolated from clinical populations.
Initial profile published at Grade C: small acute trials show recall and attention effects via cholinesterase inhibition, with the strongest branded trial sponsor-funded.
July 4, 2026
5 changesInitial profile published at Grade C: a growing but still limited human evidence base on cognitive endpoints specifically.
Initial profile published at Grade B: multiple 8–12 week human trials report memory and learning gains, with a characteristically cumulative (not acute) time course.
Initial profile published at Grade B: several human trials report attention and memory measures, though both principal supporting trials are tied to the branded-form manufacturer — a sponsor-dependence caveat behind the grade.
Initial profile published at Grade Emerging: some small human pilot trials plus considerable preclinical work, short of the consistent healthy-adult replication a higher grade needs.
Initial profile published at Grade C: human cognition trials are concentrated in one branded extract and largely one research group, so the results do not automatically generalize.
Get the changes as they ship.
We publish an update whenever a grade moves, an ingredient is graded out, or a new profile lands. Follow the RSS feed, or join the waitlist below. Evidence-update notes go out to the list, never spam.
Less noise. More signal.
Get The Signal by email
New research write-ups, evidence-grade changes and safety notes, about every two weeks. Free, with the Clean Focus Stack Guide when you join, and first word if Signal State Core launches.
About every two weeksFree guide on signupUnsubscribe anytime