How Synaptik checks: sources, evidence, funding

Who is behind Synaptik

Synaptik is an independent project by Tilman Finley Winter, a nutritional scientist with a focus on biochemistry (B.Sc. Nutritional Science, University of Hohenheim). Forums and shops make big claims about many compounds. Synaptik explains where and how they act in the body and how well that is supported.

At the centre of every profile in the library are the claims made for a compound in advertising and online forums. Each one is checked against the studies on its own. The first is the main claim; for nootropics, for example, better cognitive performance in healthy people.

Research and drafts are produced with AI assistance. Every source with a DOI or PMID is resolved automatically via Crossref and PubMed, and the main claims are also checked by hand. The rules on this page apply to the research as much as to the checking.

Funding and interests

So far, Synaptik is paid for out of the author’s own pocket. At present, Synaptik carries no advertising, no shops and no affiliate links. No profile recommends a product or brand to buy.

Scientific objectivity

All pharmacological profiles, mechanisms of action and receptor data on Synaptik are based strictly on published scientific literature, first and foremost primary literature. Future partnerships, selected product placements and Synaptik Labs’ advanced tools help develop the platform further and have no influence whatsoever on the biochemical evaluation of the evidence.

As soon as there is advertising, product placement or a partnership, it will always be clearly labelled as advertising and kept separate from the profiles. It changes no level of evidence and no result.

Declaration of interests

Tilman Finley Winter has no conflicts of interest. He receives no money and no products from manufacturers of supplements or medicines, or from laboratories, and he is not employed by or a consultant to any of them.

Any professional work by the author related to supplements, medicines, health coaching or health advice will be disclosed on this page. At present there is none.

Interests behind the studies themselves are stated in the profile whenever the evidence depends on them: funding by the manufacturer, patents, or conflicts of interest declared by the researchers.

Research and sources

Synaptik relies on primary studies, systematic reviews and meta-analyses, and guidelines from professional societies and agencies such as the German Nutrition Society (DGE), EFSA and the WHO. These take precedence over secondary sources. Text from other websites is never copied or paraphrased.

Citations follow the Vancouver style: sources are numbered in order of first mention and carry their DOI or PubMed ID (PMID) where one exists. If a rating rests on the abstract alone because the full text is not freely available, the profile says so („nur Abstract geprüft“).

Before a profile is added, a checking script is run on it. It checks required fields, vocabulary and reference numbers, and resolves every identified source online: DOIs via Crossref or, if they are not listed there, via doi.org; PMIDs via PubMed, including a check that the PMID and DOI belong to the same article; and trial registry numbers (NCT) at ClinicalTrials.gov. The working rule is that a profile is only accepted once the script reports no errors. The script is run by hand before every change; it does not run automatically when the website is built.

Approval and legal status are based on legal sources: EU legislation with its CELEX number, such as Regulation (EU) No 432/2012 on health claims; German statutes and ordinances such as the Narcotics Act (BtMG), the Prescription Ordinance (AMVV) and the Anti-Doping Act; the BfArM drug database AMIce; EMA documents, including its Article 57 product database; and the European Commission’s Novel Food Catalogue. Database queries are cited with the date of access. The Novel Food Catalogue is guidance issued by the Commission, not legislation, and it is cited as such. The checking script recognises these legal sources as evidence but does not resolve them online; their wording and status are checked by hand against the legal text or the agency’s page. Anything that cannot be documented is marked as not verified in the profile („nicht verifiziert“).

Popular claims from forums, shops or catalogues are not adopted. They are rated instead: supported, partly supported, not supported or refuted.

Level of evidence and result

Every claim gets two ratings that are kept apart. The level of evidence says how well the claim was tested, that is, with which study design. The result says what the tests found. A randomised trial (level 4) can therefore end in “no benefit”, and an animal study (level 2) does not show a benefit in humans, however clear the effect was in the animals.

Levels of evidence 0 to 4

0No dataKeine Daten: No study that counts for the claim. Often there is only the claim itself; sometimes existing studies test something other than what is promised, such as a different form, a different design or only a marker.
1LaboratoryLabor: Findings from the test tube, cell culture, isolated tissue or computer models (in silico). They show targets, not the effect in the body.
2Animal studyTierversuch: Findings in living animals. The profile always names the species; an animal finding never counts as evidence for a claim in humans.
3Human, non-randomisedMensch, nicht randomisiert: Human studies without random allocation: open-label or uncontrolled studies, case series and cohort studies. They show associations but not cause and effect.
3Case reports onlyNur Fallberichte: Level 3, but from individual case reports only: observations in single people without a comparison group. They show that something can happen, not how often or whether the compound caused it.
4Human, randomisedMensch, randomisiert: Human studies that are randomised and/or placebo-controlled. Causal wording such as “lowers” or “improves” is used for randomised studies only.

What the studies found

Proven benefitNutzen belegt: Human studies show the promised benefit.
Small benefitkleiner Nutzen: The promised effect is there, but it is small.
Only under strainnur unter Belastung: An effect appears only under strain, such as sleep deprivation. Under ordinary conditions it is not shown.
Only tested against induced deficitsnur bei künstlichem Defizit geprüft: The studies only tested whether the compound eases an artificially induced deficit or disturbance, such as memory impairment caused by scopolamine or the effects of a low-boron diet. Whether it helps healthy people without that deficit is open.
Only proven in patientsnur bei Kranken belegt: A benefit is shown only in people with a disease or a proven deficiency. Nothing follows from this for the healthy people the claim is aimed at.
Only markers provennur Marker belegt: Only for separate secondary claims about a marker, such as a blood level: the marker changes in a consistent, well-supported direction. That does not show a clinical benefit.
Single unconfirmed findingEinzelbefund, unbestätigt: A single finding that no other study has confirmed. This includes a small pilot study that missed its primary endpoint.
Mixed resultsuneinheitlich: Studies at the same level contradict each other.
No benefitkein Nutzen: Studies tested the claim and found no effect.
Worse performanceVerschlechterung: The studies show the opposite of the claim, such as worse performance.
Not studiednicht untersucht: No study tests the claim the way it is made: in humans, in the promised form and in the group it is aimed at.

In the profiles the scale ends at 4. Systematic reviews and meta-analyses are cited and discussed in the text; the evidence ladder on the home page shows in general terms what each type of study can and cannot tell you.

Rating rules

Who was studied

Every claim states the population of the studies, in one of four forms: healthy („gesund“), healthy with a detail such as older adults („gesund (…)“), patients with the condition named („Patienten: …“), or mixed, when healthy people and patients were studied together („gemischt: …“).

Findings in one group do not automatically apply to another. If a benefit is shown only in patients or in deficiency, the result is “only proven in patients”; if only a lab value changed, in a consistent and well-supported direction, this is listed as a separate marker claim rated “only markers proven”. Conversely, a weak study in a different group, such as young children, is not enough for “no benefit” in adults.

Lab values and health claims

A higher blood level, greater excretion or a change in enzyme activity does not show a clinical benefit. If a clinical claim rests on such markers alone, it stays at level 0, “not studied”, and the marker finding is listed as a separate secondary claim.

Authorised EU health claims (mainly under Regulation (EU) No 432/2012) are a legal statement, not a level of evidence. They appear in the profile as a separate label, “EU claim authorised” („EU-Claim zugelassen“), in some older profiles still only in the legal status and the text, and raise neither the level nor the result. If there is no study in healthy people on the content of the claim for that form, the label adds “legally permitted, not studied in healthy people” („rechtlich zulässig, bei Gesunden nicht untersucht“).

Vitamins and minerals: each form on its own

Vitamins and minerals have one profile per chemical form, for example magnesium citrate or zinc bisglycinate. The level and result are based only on studies that used exactly that form, and for salts exactly that salt. Studies with a different form, or that do not say which form they used, are flagged in the text but do not count. A main claim without a study on that form remains “not studied”, even if the nutrient itself is well researched.

There is one documented exception. If a study only says “pyridoxine” or “vitamin B6”, “thiamine” or “vitamin B1”, or “riboflavin” or “vitamin B2”, it counts for pyridoxine hydrochloride, thiamine hydrochloride or free riboflavin respectively, as long as the source names no other form. For each of these three vitamins only one form is in common use. All other nutrients follow the strict rule; “glycinate” without “bis-”, for example, does not count as bisglycinate.

When a study finds no effect

A null result is rated according to what the study tested and how large it was. Since 26 October 2026 the rules are:

If a study tests the claim in a different design from the one promised, such as a single dose when a training effect is promised, the result is “not studied” (level 0). The finding is described in the text.
If a small pilot study tests the claim directly and misses its primary endpoint, the result is “single unconfirmed finding”, at the level of its design, until a larger study is available.
If a properly planned randomised trial with an adequate sample size finds no effect, the result is “no benefit”.

A few ratings made under the earlier reading are being brought into line with this version.

Further rules

Studies of combination products do not count for the individual compound.
Cell and animal findings are labelled as such in the text, animal findings with the species.
If a claim rests on individual case reports only, it is shown as level 3 with the note “case reports only” („nur Fallberichte“), and the result is at most “single unconfirmed finding”.

Notes on certainty and numbers

Claims from level 3 upwards carry small notes when the evidence is less certain than the level alone suggests: “small studies” when every supporting study has fewer than 100 participants; “studies disagree” when supporting studies point in opposite directions; “manufacturer-funded” when, according to their funding or conflict-of-interest statements, all supporting studies were paid for by the manufacturer or list its employees as authors. A note appears only when the source supports it; the levels 0 to 4 stay as they are. Synaptik gives numbers in absolute rather than only relative terms, with the same denominators, and side effects as “x in 100 over y weeks” compared with placebo. New and revised texts follow this rule at once; older ones are adjusted step by step, each only after the primary source has been checked.

What Synaptik does not do

Synaptik gives no doses, no dosing schedules and no places to buy, and it makes no purchase recommendations. Figures from studies do appear in the profiles, such as binding affinities (Ki, IC50), the number of participants, effect sizes or bioavailability in per cent.

For recreational drugs and potent medicines, the profiles are limited to mechanism, evidence and risks. Amounts, ways of use, preparation and descriptions of what an effect feels like are left out.

Risks and warnings appear in a profile only with evidence and a level of evidence. Unsupported warnings are rated like any other claim.

Synaptik is a reference work, not personal advice. Its content does not replace medical advice, diagnosis or treatment. Anyone considering a product should discuss it with a doctor or pharmacist.

The figures are schematic. Every number in a figure is referenced, usually to a source in the compound’s profile, otherwise to the reference list of the page. Anything a model merely illustrates is marked as illustrative.

Review dates and corrections

Every profile shows its review date, the date it was last checked, below the reference list, in the section „Wie dieses Profil geprüft wurde“. Pages like this one carry the date they were last reviewed. For now the library is in German only; an English version is planned.

If you find a mistake, read a source differently or know of a newer study, please email tilman@synaptiklabs.org, ideally with the name of the profile, the passage and the source. A statement is only changed once the primary source has been checked.

Last reviewed: 26 October 2026 · This information does not replace medical advice.