MindHeaven Evidence Standard

MHES is an independent evidence-assessment framework developed by MindHeaven.

It draws methodological principles from internationally recognised systems for evidence synthesis, trial reporting, risk-of-bias assessment and research transparency.

These include:

• PRISMA for transparent reporting of systematic reviews;
• GRADE principles for assessing certainty in bodies of evidence;
• Cochrane Risk of Bias 2 for randomised trials;
• ROBINS-I for non-randomised intervention studies;
• AMSTAR 2 for evaluating systematic reviews;
• CONSORT principles for evaluating the completeness of randomised trial reporting;
• ICMJE principles concerning authorship, trial registration and conflict-of-interest disclosure;
• EFSA and European Commission frameworks concerning nutrition and health claims.

PRISMA 2020 provides reporting guidance and a checklist for systematic reviews, while Cochrane recommends RoB 2 for assessing bias in randomised trials. GRADE assesses certainty separately for important outcomes rather than treating an intervention as having one universal evidence rating.

CONSORT provides structured reporting standards for randomised trials; ROBINS-I addresses risk of bias in non-randomised intervention studies; and AMSTAR 2 is designed to critically appraise systematic reviews containing randomised or non-randomised intervention studies.

MHES is not PRISMA, GRADE, Cochrane, CONSORT or an EFSA methodology.

MindHeaven does not imply endorsement by any of these organisations.

Where MHES uses principles originating from established scientific frameworks, those principles are adapted for the specific purpose of transparent evidence communication in supplementation and cognitive performance.

Principle 1 - Evidence before formulation

A formulation decision should follow evidence assessment whenever reasonably possible.

The preferred sequence is:

1. Scientific literature→
2. Evidence assessment→
3. Safety assessment→
4. Claim assessment→
5. Formulation rationale→
6. Product development

Marketing considerations must not determine the scientific interpretation of evidence.

Principle 2 - Claims are graded, not ingredients

MHES does not normally assign a single evidence grade to an ingredient.

Evidence is attached to a defined claim.

For example:

L-theanine improves cognitive performance.

is too broad.

A more appropriate assessment would be:

Acute administration of L-theanine combined with caffeine improves selected measures of attention in healthy adults compared with caffeine alone.

The evidence supporting attention, stress, anxiety, sleep, reaction time and long-term cognitive function may differ substantially for the same compound.

Therefore:

One ingredient may legitimately hold several different MHES ratings.

Principle 3 - Human outcomes outrank biological plausibility

A mechanism may explain why an effect could occur.

It does not demonstrate that the effect occurs in humans.

Evidence from receptor studies, cell cultures, biochemical pathways, animal models and pharmacokinetic modelling may contribute to biological plausibility.

It must not be presented as equivalent to demonstrated human benefit.

Principle 4 - Negative evidence must be actively sought

MindHeaven does not merely search for evidence supporting a hypothesis.

Every substantive MHES review requires a disconfirmation search.

Researchers must actively look for:

• null findings,
• unsuccessful replications,
• contradictory trials,
• adverse findings,
• withdrawn or retracted studies,
• outcomes that failed to reach statistical significance,
• failed primary endpoints,
• studies challenging proposed mechanisms.

Negative evidence must not be buried beneath positive evidence.

Principle 5 - Replication matters more than novelty

One impressive study is not equivalent to a replicated body of evidence.

Novel findings may be scientifically important while remaining uncertain.

MHES therefore places substantial weight on:

• independent replication,
• consistency across research groups,
• consistency across populations,
• consistency across outcome measures,
• consistency across study designs.

Principle 6 - The tested population matters

Evidence obtained from one population must not automatically be extrapolated to another.

Particular care is required when moving between:

• deficient and non-deficient individuals,
• clinical and healthy populations,
• young and older adults,
• sleep-deprived and rested individuals,
• trained and untrained populations,
• children and adults.

Evidence that a nutrient corrects cognitive impairment caused by deficiency does not automatically establish that additional intake enhances cognition in already sufficient individuals.

Principle 7 - The tested dose and form matter

Evidence applies most directly to the chemical form, dose, route of administration, duration and formulation actually studied.

Evidence concerning one molecular form should not automatically be transferred to another.

For example, evidence concerning a parent compound does not by itself establish equivalent efficacy for a salt, ester, precursor, derivative, complex or alternative delivery system.

Pharmacokinetic similarity may support an inference.

It does not replace outcome evidence.

Principle 8 - Safety and efficacy are evaluated separately

An ingredient may have promising efficacy evidence but insufficient safety evidence, or strong safety evidence but weak efficacy evidence.

These questions must not be conflated.

Absence of observed harm is not automatically proof of long-term safety.

Principle 9 - Commercial conflict must be visible

MindHeaven is a commercial organisation that formulates and sells supplements.

This creates an inherent potential conflict when MindHeaven evaluates substances that appear in its own products.

MHES therefore requires explicit disclosure of that relationship.

ICMJE emphasises disclosure of both financial and non-financial relationships that may influence professional judgement.

Commercial interest does not automatically invalidate evidence.

It does increase the need for transparency.

Principle 10 - Conclusions are provisional

Scientific conclusions may change.

MHES conclusions are therefore never presented as permanent truth.

Every assessment must contain:

• the date of the literature search,
• the date of the latest review,
• the evidence version,
• and, where appropriate, the evidence that could change the conclusion.

Changing a conclusion because the evidence changed is not a failure of the system. It is evidence that the system is working.

4.1 Evidence Note

A concise analysis of one study, claim or scientific issue.

Typical use:

• interpretation of a newly published study,
• clarification of a viral claim,
• statistical anomaly,
• methodological observation.

An Evidence Note is not a systematic review.

4.2 Evidence Dossier

The principal MHES format.

An Evidence Dossier is a structured assessment of the available evidence concerning a defined ingredient or claim.

It contains:

• question definition,
• literature search,
• evidence table,
• negative evidence,
• risk-of-bias assessment,
• evidence interpretation,
• MHES grade,
• safety considerations,
• uncertainties,
• references.

An Evidence Dossier may be extensive without being called a systematic review.

4.3 MindHeaven Evidence Report

A major evidence synthesis addressing a broad question.

Whenever a report is described as a systematic review, MindHeaven must meet substantially higher methodological and reporting requirements, including a predefined protocol, reproducible search methodology and transparent study-selection process consistent with systematic-review practice.

PRISMA specifically provides reporting guidance for systematic reviews and includes a checklist and flow diagrams for documenting identification, screening and inclusion of evidence.

MindHeaven will therefore not use the expression systematic review merely because many studies were reviewed.

The basic unit of evidence assessment is a claim.

Whenever possible, a claim should define:

P - Population Who was studied?

I - Intervention What compound, formulation and dose?

C - Comparator Compared with what?

O - Outcome What was measured?

T - Time For how long?

This produces a PICOT-style evidence question.

Example:

In healthy adults aged 18-45, does acute administration of 200 mg L-theanine combined with caffeine improve objective attention outcomes compared with caffeine alone within four hours of ingestion?

This is substantially more scientifically useful than:

Does L-theanine work?

MHES uses evidence hierarchy as guidance, not as an automatic scoring system.

A poorly performed randomised trial may be less informative than a well-conducted observational study for some questions.

Nevertheless, for questions of intervention efficacy, direct human experimental evidence generally carries greater evidentiary weight.

The standard hierarchy is:

Tier 1 - High-quality systematic reviews and meta-analyses of relevant controlled human trials. Tier 2 - Randomised controlled human trials. Tier 3 - Controlled non-randomised human intervention studies. Tier 4 - Prospective observational human studies. Tier 5 - Retrospective and cross-sectional human evidence. Tier 6 - Human mechanistic, pharmacodynamic or pharmacokinetic evidence. Tier 7 - Animal evidence. Tier 8 - In-vitro and cellular evidence. Tier 9 - Expert opinion, narrative review, theoretical reasoning and unsupported mechanistic extrapolation.

Lower-tier evidence may provide important biological context.

It should not be rhetorically upgraded into higher-tier evidence.

Every Evidence Dossier must disclose how the literature was identified.

At minimum, the record must contain:

• search date;
• databases searched;
• principal search terms;
• inclusion criteria;
• exclusion criteria;
• date range where applicable;
• language restrictions where applicable;
• trial registries searched;
• citation-chasing method.

For major Evidence Dossiers, MindHeaven should normally search at least:

• PubMed/MEDLINE;
• Cochrane resources relevant to controlled trials;
• ClinicalTrials.gov;
• WHO ICTRP or equivalent trial-registry resources;
• reference lists of included papers.

Clinical trial registration is an important transparency signal. ICMJE recommends prospective registration of clinical trials in a public registry at or before first participant consent, and WHO maintains the ICTRP to improve public access to clinical-trial registration information.

Where access and relevance justify it, additional bibliographic databases may be used.

Every Evidence Dossier must contain a deliberate attempt to find information opposing the working hypothesis.

The research team should combine relevant ingredient or intervention terms with concepts such as:

• no effect,
• null result,
• failed,
• ineffective,
• not significant,
• adverse event,
• adverse effect,
• safety,
• contradictory,
• replication,
• withdrawal,
• retraction.

The objective is not to manufacture artificial balance.

The objective is to reduce confirmation bias.

A review that searches only for confirming evidence does not meet MHES requirements.

Studies are included because they help answer the defined question.

They are not selected because their conclusions are favourable.

A study cannot be excluded simply because:

• it reports a null result;
• it contradicts another study;
• it weakens a MindHeaven formulation rationale;
• it was funded by a competitor;
• it was industry funded.

Reasons for exclusion should be methodological or scope-related and, in major reports, documented.

Preprints may contain valuable emerging evidence.

They must be clearly labelled.

A preprint should not normally carry the same evidentiary weight as a completed peer-reviewed publication when otherwise comparable evidence exists.

However, peer review itself does not guarantee methodological quality.

MHES therefore evaluates study quality rather than using publication status as a substitute for critical appraisal.

Retracted research must not be used as positive evidentiary support.

Where a study previously cited by MindHeaven is subsequently retracted, materially corrected or subject to an expression of concern, the affected Evidence Dossier must be reviewed.

If the evidence grade or conclusion changes, the change must appear in the public Correction History.

Study design alone does not establish study reliability.

Randomised trials should, where appropriate, be examined using domains consistent with Cochrane Risk of Bias 2.

RoB 2 evaluates bias associated with different aspects of trial design, conduct and reporting rather than assuming that randomisation alone guarantees reliability.

Non-randomised intervention studies should, where appropriate, be assessed using principles consistent with ROBINS-I, whose current framework is designed to evaluate risk of bias in estimates from non-randomised intervention studies.

Systematic reviews used as major evidence anchors may be assessed using AMSTAR 2.

AMSTAR 2 uses critical domains rather than a simple numerical total score, and explicitly warns against reducing review quality to an overall points score.

For each important human intervention study, reviewers should consider:

13.1 Randomisation Was allocation genuinely random?

13.2 Allocation concealment Could investigators predict assignment?

13.3 Blinding Were participants, investigators and outcome assessors appropriately blinded?

13.4 Comparator quality Was the comparator capable of answering the actual scientific question?

13.5 Attrition How many participants were lost? Was loss balanced between groups?

13.6 Missing data Could missing outcomes materially alter the result?

13.7 Outcome specification Was the reported outcome predefined?

13.8 Selective reporting Were only favourable endpoints highlighted?

13.9 Multiplicity How many outcomes, subgroups and time points were analysed?

13.10 Statistical coherence Do effect sizes, confidence intervals, test statistics, p-values and sample sizes appear mutually consistent?

13.11 Biological and clinical relevance Is the observed effect large enough to matter?

13.12 Reproducibility Has another independent group observed a comparable effect?

MHES rejects the equation:

p < 0.05 = proven

Statistical significance is one part of interpretation.

Whenever available, MindHeaven should evaluate:

• absolute effect size;
• standardised effect size;
• confidence interval;
• sample size;
• statistical power;
• baseline imbalance;
• number of endpoints tested;
• prespecified versus exploratory analysis;
• missing data;
• model choice;
• sensitivity analyses;
• multiplicity correction.

A statistically significant but trivial effect should not automatically be described as meaningful.

A non-significant finding should not automatically be described as proof of no effect.

Primary outcomes should receive greater interpretive weight than exploratory secondary outcomes.

If the primary endpoint fails but one of numerous secondary endpoints succeeds, the study must not be represented as unambiguously positive.

Where trial registration or protocol information is available, reported outcomes should be compared with prespecified outcomes.

Prospective trial registration is therefore considered a positive transparency signal under MHES.

MHES does not impose an arbitrary participant threshold below which a study becomes invalid.

However, small samples increase uncertainty and often produce:

• wide confidence intervals;
• unstable effect estimates;
• limited subgroup reliability;
• reduced ability to identify adverse events.

A small study may generate an important hypothesis.

It should rarely carry a broad consumer claim on its own.

This is a mandatory MHES rule.

If a study investigates A + B against placebo and finds an effect, MindHeaven must not conclude that A produced the effect or that B produced the effect.

The valid conclusion is:

The tested combination produced the observed result under the tested conditions.

Ingredient-specific attribution requires an appropriate design capable of isolating the ingredient.

The word synergy has a high evidentiary threshold.

Finding that A + B > placebo does not demonstrate synergy.

A meaningful synergy claim normally requires evidence capable of distinguishing the combination from its individual components.

Ideally, A, B, A + B and placebo should be compared within an appropriate design.

Without such evidence, MindHeaven may describe a combination as a formulation rationale.

It must not present synergy as demonstrated fact.

When a compound is commonly combined with an active ingredient such as caffeine, placebo alone may not answer the most relevant question.

For example:

L-theanine + caffeine vs placebo

can show whether the combination differs from placebo.

It cannot establish that adding L-theanine improves caffeine itself.

The relevant comparison for that claim is:

L-theanine + caffeine vs caffeine alone.

MHES must distinguish these questions.

Statements such as:

• crosses the blood-brain barrier;
• increases neurotransmitter X;
• affects receptor Y;
• increases biomarker Z;

may establish biological activity.

They do not automatically establish better memory, better concentration, reduced anxiety, enhanced intelligence or improved performance.

MHES therefore separates mechanistic evidence from outcome evidence.

A biomarker is not automatically a meaningful human outcome.

Improving a laboratory marker does not necessarily demonstrate that the user thinks better, performs better, feels better, remains healthier or experiences a clinically meaningful benefit.

Biomarker evidence must therefore be identified as surrogate evidence where appropriate.

Findings from clinical population to healthy population require caution.

Findings from deficient population to nutritionally sufficient population require caution.

Findings from older adults to young adults require caution.

Findings from sleep-deprived subjects to rested subjects require caution.

MHES explicitly rates indirectness when evidence is extrapolated beyond the studied population.

An acute study demonstrates an acute result.

It must not automatically be converted into a long-term claim.

Likewise, a chronic supplementation trial does not necessarily establish an immediate performance effect.

MHES therefore distinguishes:

• acute effects;
• short-term effects;
• chronic effects;
• persistence after discontinuation.

Evidence concerning one form of an ingredient must not automatically be transferred to another.

Examples include:

• L-tyrosine versus N-acetyl-L-tyrosine;
• magnesium citrate versus magnesium chloride;
• choline versus citicoline;
• different botanical extracts with different standardisation.

Where equivalence is uncertain, MHES explicitly states:

Form-specific evidence is insufficient.

MHES uses six primary evidence grades.

Grades apply to specific claims, not entire substances.

MHES A - STRONG HUMAN EVIDENCE

The claim is supported by a convincing body of direct human evidence.

Typical characteristics:

• multiple relevant controlled human studies;
• at least some high-quality evidence;
• independent replication;
• broadly consistent direction of findings;
• acceptable risk of bias;
• clinically or practically meaningful outcome;
• reasonable precision.

Public label: Strong Human Evidence

MHES B - MODERATE HUMAN EVIDENCE

The claim is reasonably supported, but meaningful uncertainty remains.

Possible limitations include:

• modest number of trials;
• limited replication;
• moderate sample sizes;
• some heterogeneity;
• limited population range;
• moderate risk of bias.

Public label: Moderate Human Evidence

MHES C - PRELIMINARY HUMAN EVIDENCE

A credible human signal exists, but the evidence is not sufficient for a confident conclusion.

Typical characteristics:

• one or several small studies;
• inconsistent findings;
• exploratory outcomes;
• limited active controls;
• high uncertainty;
• lack of independent replication.

Public label: Preliminary Human Evidence

MHES D - INDIRECT / MECHANISTIC EVIDENCE

The rationale is supported primarily by mechanistic human studies, pharmacokinetics, biomarkers, animal research or cell research, without adequate direct human outcome evidence for the claim.

Public label: Mechanistically Plausible - Human Outcome Evidence Insufficient

MHES E - INSUFFICIENT EVIDENCE

Available evidence is inadequate to support or meaningfully evaluate the claim.

This may reflect:

• absence of relevant studies;
• extremely poor study quality;
• inappropriate population;
• inappropriate outcome;
• insufficient data.

Public label: Evidence Insufficient

MHES X - CLAIM NOT SUPPORTED / CONTRADICTED

Current evidence materially fails to support the specific claim or provides credible evidence against it.

Public label: Claim Not Supported

or, where justified:

Evidence Contradicts the Claim

MHES X does not automatically mean that the ingredient is ineffective for every purpose.

It applies only to the evaluated claim.

MHES does not use rules such as:

Three studies = Grade A.

A collection of five weak trials does not automatically outweigh one large, rigorous study.

Evidence grading considers:

• risk of bias;
• consistency;
• precision;
• directness;
• replication;
• study design;
• outcome relevance;
• publication bias;
• magnitude of effect.

This concept is broadly compatible with the principle used by GRADE that certainty is assessed for individual outcomes and can be reduced by limitations such as inconsistency, indirectness or imprecision.

In addition to the grade, every major claim receives a plain-language verdict.

Available verdicts include:

• Supported
• Probably Supported
• Promising but Unproven
• Not Established
• Insufficient Evidence
• Not Supported
• Contradicted

The wording must reflect the grade and evidence context.

Safety receives an independent assessment.

Possible public statuses are:

• Safety evidence reasonably established within studied use
• Safety evidence limited
• Long-term safety uncertain
• Population-specific caution
• Material safety signal identified
• Evidence insufficient for safety assessment

Safety review should consider where relevant:

• adverse events;
• serious adverse events;
• withdrawal rates;
• tolerability;
• dose;
• duration;
• drug interactions;
• pregnancy;
• lactation;
• children;
• older adults;
• liver disease;
• kidney disease;
• cardiovascular conditions;
• concurrent stimulants;
• relevant regulatory warnings.

A statement that an ingredient is "safe" must never be broader than the evidence justifies.

Every Evidence Dossier evaluating an ingredient used in a MindHeaven product must include a prominent disclosure.

Recommended wording:

Commercial Disclosure MindHeaven currently sells, develops or may develop products containing the substance discussed in this review. This creates a potential commercial conflict of interest. Under the MindHeaven Evidence Standard, this requires explicit disclosure, inclusion of contradictory and negative evidence, and separation between evidence assessment and formulation rationale.

The disclosure must appear near the beginning of the document.

Not hidden in the footer.

Funding sources for important studies must be recorded when known.

Studies should be labelled, where appropriate, as:

• publicly funded;
• academically funded;
• industry funded;
• mixed funding;
• funding unclear.

Industry funding does not automatically invalidate a study.

Likewise, academic funding does not automatically guarantee reliability.

Study methodology remains the primary focus.

A scientifically plausible or even well-supported claim is not automatically a legally authorised marketing claim.

MindHeaven must explicitly separate:

Scientific assessment

from:

Regulatory authorisation.

Within the European Union, nutrition and health claims are governed by Regulation (EC) No 1924/2006, and the European Commission maintains a register containing authorised and non-authorised health claims and applicable conditions. EFSA evaluates scientific substantiation for health-claim applications.

Therefore an Evidence Dossier may conclude:

Scientific evidence: Moderate

while separately stating:

EU authorised health claim: None identified.

These statements are not contradictory.

Evidence does not automatically dictate product formulation.

A formulation decision may also require consideration of:

• safety;
• dose feasibility;
• stability;
• solubility;
• interactions;
• delivery system;
• sensory characteristics;
• regulatory status;
• ingredient compatibility;
• cumulative exposure.

Accordingly, MHES distinguishes evidence conclusion from MindHeaven formulation decision.

ADOPT Evidence and safety rationale justify inclusion.

CONDITIONAL ADOPT Ingredient may be included, but important uncertainty must be disclosed.

HOLD Potentially interesting but evidence is not mature enough.

REJECT Evidence, safety, feasibility or regulatory concerns do not justify inclusion.

REMOVE New information materially weakens the original rationale.

MindHeaven maintains a public record of ingredients considered but rejected.

Each entry should include:

• ingredient;
• date evaluated;
• original rationale;
• evidence level;
• important safety considerations;
• reason for rejection;
• evidence that could change the decision.

A rejected ingredient is not permanently condemned.

The decision may be reconsidered if the evidence changes.

For every major MindHeaven formulation, the public rationale should allow the reader to trace:

1. Ingredient→
2. Claim considered→
3. Evidence Dossier→
4. MHES grade→
5. Dose evidence→
6. Safety assessment→
7. Formulation rationale→
8. Final decision

The objective is to create a visible provenance chain between scientific evidence and product composition.

Major Evidence Dossiers should undergo independent review where practicable.

Reviewers should have relevant expertise in fields such as:

• neuroscience;
• pharmacology;
• nutrition;
• medicine;
• biostatistics;
• epidemiology;
• research methodology.

The reviewer must be permitted to disagree with MindHeaven.

A reviewer must not be asked to endorse a product as a condition of scientific review.

Each scientific reviewer should disclose relevant:

• financial relationships;
• consulting relationships;
• employment;
• research funding;
• intellectual-property interests;
• competitive relationships;
• other relevant professional interests.

Potential conflicts should be visible where material.

Scientific disagreement is permitted.

Where a reviewer materially disagrees with a MindHeaven conclusion and consensus cannot legitimately be reached, the disagreement should not necessarily be removed.

Where appropriate, the document may state:

Reviewer Perspective: The independent reviewer considers the evidence less certain than the final MHES assessment because...

Transparent disagreement may itself be valuable information.

Every major MindHeaven evidence publication should identify:

• Research and analysis
• Lead editor
• Scientific reviewer
• Statistical reviewer, where applicable
• Literature search date
• Publication date
• Last review date
• Document version
• Commercial disclosure

Authorship and contribution roles should be meaningful rather than ceremonial.

ICMJE distinguishes substantive authorship/contributor responsibilities and recommends transparent disclosure of relationships that may influence the work.

MindHeaven may use artificial intelligence as a research-support tool.

Permitted uses may include:

• literature discovery support;
• deduplication assistance;
• table preparation;
• consistency checking;
• terminology normalisation;
• statistical cross-checking;
• translation support;
• identification of potential contradictory evidence;
• structured extraction assistance.

However:

AI must not be treated as the evidentiary source.

Every substantive scientific statement must ultimately be traceable to an identifiable source.

AI-generated citations must never be accepted without verification.

A DOI, PMID, registry record or publication must be opened and verified before inclusion.

AI must not independently determine the final MHES grade without human review.

For important references, MindHeaven should verify:

• article title;
• authors;
• journal;
• publication year;
• DOI and/or PMID;
• study design;
• sample size;
• intervention;
• comparator;
• outcome;
• reported conclusion.

Where only an abstract is available, this limitation should be recorded if the missing information could materially affect interpretation.

Major Evidence Dossiers should contain, where relevant, the following fields:

Study citation - Required: Yes Year - Required: Yes Study design - Required: Yes Population - Required: Yes Sample size - Required: Yes Intervention - Required: Yes Dose - Required: Yes Comparator - Required: Yes Duration - Required: Yes Primary outcome - Required: Where available Main effect - Required: Yes Effect size - Required: Where available Confidence interval - Required: Where available Adverse events - Required: Where available Funding - Required: Where available Trial registration - Required: Where available Risk of bias - Required: Yes for major evidence Key limitation - Required: Yes

Every Evidence Dossier must contain a short section titled:

WHAT WE KNOW

It should contain only conclusions that the current evidence reasonably supports.

No promotional language.

No speculation.

Every Evidence Dossier must contain:

WHAT WE DON'T KNOW

This section should identify:

• unanswered questions;
• missing populations;
• missing comparator groups;
• insufficient duration;
• uncertain dose;
• lack of replication;
• safety gaps;
• mechanistic assumptions.

Uncertainty is not a weakness to hide.

It is part of the evidence.

Every Evidence Dossier must include a concise summary designed to answer:

What can a careful reader reasonably conclude from the evidence today?

The Bottom Line should normally contain:

1. the strongest supported conclusion; 2. the principal uncertainty; 3. the most important limitation; 4. the MHES grade.

Example:

Human evidence suggests that the tested combination may improve selected attention outcomes under acute conditions. However, the evidence base remains relatively small and does not establish broad cognitive enhancement or long-term benefit. MHES Grade: B - Moderate Human Evidence.

Every major Dossier should contain:

WHAT WOULD CHANGE OUR CONCLUSION?

Examples:

• a large preregistered replication;
• an appropriate active-comparator trial;
• longer safety follow-up;
• direct evidence in healthy adults;
• independent replication;
• publication of an ongoing registered trial.

This converts uncertainty into a visible research agenda.

Each Evidence Dossier receives a permanent identifier.

Example:

MH-EVD-001

and version:

v1.0

Subsequent changes become:

v1.1 - minor update v1.2 - additional evidence v2.0 - conclusion materially changed

Previous major versions should remain accessible where technically practical.

MindHeaven does not silently rewrite material scientific errors.

Material corrections must disclose:

• What was stated
• What was wrong
• What has been changed
• Why it matters
• Whether the MHES grade changed
• Date of correction

Minor typographical changes do not require a formal correction notice.

Priority evidence relating directly to active MindHeaven formulations should normally be rechecked at least every six months.

Broader evidence dossiers should normally be reviewed annually.

Earlier reassessment should occur when:

• a major RCT is published;
• a new systematic review appears;
• a safety warning emerges;
• a study is retracted;
• a regulatory decision changes;
• credible contradictory evidence appears.

Pages awaiting reassessment may display:

Evidence review due.

MindHeaven invites scrutiny.

Every major Evidence Dossier should provide a mechanism for researchers, clinicians and readers to identify:

• missed evidence;
• methodological errors;
• incorrect interpretation;
• citation errors;
• undisclosed conflicts;
• statistical inconsistencies.

Recommended wording:

Challenge This Review If you are an author of research cited here, a researcher working in this field, or a reader who believes we have missed important evidence or misinterpreted a study, we want to know. Scientific conclusions improve when they are challenged.

Citing or profiling a scientist does not imply endorsement.

Where MindHeaven features researchers whose work is discussed, the following principle applies:

Inclusion in the MindHeaven Research Network, citation in the MindHeaven Library or discussion of a researcher's work does not imply endorsement of MindHeaven, its products or MindHeaven's interpretation of the evidence unless explicitly stated by that researcher.

Evidence may be simplified for social media.

It may not be altered.

A social-media post must not communicate greater certainty than the underlying Evidence Dossier.

Therefore:

MHES C must not become: "Scientifically proven."

A study involving 20 participants must not become: "Science confirms..."

A mouse study must not become: "This improves the human brain."

MindHeaven headlines may be provocative.

Their substantive meaning must remain accurate.

Acceptable: "The Study Everyone Quotes - And What It Actually Found"

Acceptable: "We Wanted This Ingredient to Work"

Unacceptable: "Scientists Prove L-Theanine Makes You Smarter"

when the underlying evidence does not establish that claim.

The following practices violate MHES:

• citation cherry-picking;
• hiding negative trials;
• presenting animal evidence as human efficacy;
• presenting mechanistic evidence as outcome evidence;
• using a combination study to prove an individual ingredient effect;
• claiming synergy without an appropriate comparator;
• transferring evidence between materially different ingredient forms without disclosure;
• presenting statistical significance as proof of practical relevance;
• presenting correlation as causation;
• presenting an uncontrolled study as proof of efficacy;
• omitting relevant commercial conflicts;
• inventing or failing to verify citations;
• describing a non-systematic search as a systematic review;
• silently deleting material corrections;
• changing an evidence grade for commercial reasons.

Scientific conclusions must not be determined by:

• inventory levels;
• sales performance;
• advertising campaigns;
• affiliate relationships;
• influencer partnerships;
• customer demand;
• product-launch schedules.

Commercial teams may request research questions.

They may not dictate research conclusions.

Where the founder, product team or commercial leadership has a direct interest in the outcome, independent scientific review should be prioritised.

Every public Evidence Dossier should prominently display:

• Claim
• MHES Grade
• Plain-language verdict
• Population evaluated
• Human evidence available: Yes / No
• Primary evidence type
• Last literature search
• Last reviewed
• Commercial relationship
• Safety evidence status

This information should be visible without requiring the reader to search through footnotes.

Every major dossier should follow approximately this order:

1. Claim 2. MHES grade 3. Bottom Line 4. Commercial Disclosure 5. What We Know 6. What We Don't Know 7. Why the Question Matters 8. Search Method 9. Human Evidence 10. Negative and Contradictory Evidence 11. Mechanistic Evidence 12. Risk of Bias 13. Statistical Interpretation 14. Dose 15. Population Applicability 16. Safety 17. Regulatory Position 18. MindHeaven Interpretation 19. Formulation Relevance 20. What Would Change Our Conclusion? 21. Evidence Table 22. References 23. Reviewers 24. Version History 25. Challenge This Review

Evidence supporting an ingredient does not establish that a commercial product contains the declared amount or meets appropriate manufacturing specifications.

Scientific evidence and manufacturing verification are separate layers of trust.

MindHeaven therefore aims to connect MHES with:

MindHeaven Batch Transparency

including, where appropriate:

• batch number;
• certificate of analysis;
• microbiological testing;
• relevant contaminant testing;
• ingredient verification;
• date of manufacture;
• laboratory identification;
• release status.

The ultimate evidence chain should become:

1. Scientific evidence→
2. Ingredient decision→
3. Formulation→
4. Manufacture→
5. Laboratory verification→
6. Specific batch

The complete MHES model is:

1. QUESTION→
2. LITERATURE→
3. CRITICAL APPRAISAL→
4. NEGATIVE-EVIDENCE SEARCH→
5. RISK OF BIAS→
6. EVIDENCE SYNTHESIS→
7. MHES GRADE→
8. SAFETY ASSESSMENT→
9. REGULATORY CHECK→
10. FORMULATION DECISION→
11. PRODUCT VALIDATION→
12. PUBLIC TRANSPARENCY→
13. CONTINUOUS REASSESSMENT

The success of MHES is not measured by how often its reviews conclude that an ingredient works.

A functioning evidence system should sometimes conclude:

Strong evidence.

Sometimes:

Promising.

Sometimes:

We don't know.

And sometimes:

We were wrong.

If every MindHeaven Evidence Dossier supports every MindHeaven commercial interest, MHES has failed its purpose.

MindHeaven commits to the following:

We will distinguish evidence from hypothesis.

We will distinguish human outcomes from mechanisms.

We will search for evidence that contradicts us.

We will publish meaningful limitations.

We will disclose commercial conflicts.

We will not turn animal evidence into human claims.

We will not call an association causation.

We will not call statistical significance certainty.

We will not call a combination effect an ingredient effect without evidence.

We will not call a review systematic unless its methodology justifies the term.

We will correct material errors visibly.

We will change our conclusions when better evidence requires it.

We will prefer uncertainty to false certainty.

We will prefer an inconvenient result to a convenient claim.

Science does not require absolute certainty.

It requires intellectual discipline about the limits of what is known.

The supplement industry has no shortage of confident claims.

What it lacks is an equally visible architecture for uncertainty.

MindHeaven intends to build that architecture.

Our responsibility is not to make every ingredient sound convincing. Our responsibility is to make the boundary between evidence and belief visible.

MINDHEAVEN Evidence before formulation. MHES v1.0

PRISMA 2020 - reporting framework for systematic reviews, including the 27-item checklist and review flow documentation. https://www.prisma-statement.org/prisma-2020-statement

GRADE Working Group - framework for evaluating certainty of evidence for individual important outcomes, using categories including high, moderate, low and very low certainty. https://www.gradeworkinggroup.org/

Cochrane Risk of Bias 2 - structured assessment of risk of bias in randomised trials. https://www.cochrane.org/learn/courses-and-resources/cochrane-methodology/risk-bias/about-risk-bias-2-rob-2

ROBINS-I - framework for assessment of risk of bias in non-randomised intervention studies. https://methods.cochrane.org/bias/risk-bias-non-randomized-studies-interventions

AMSTAR 2 - critical appraisal framework for systematic reviews of randomised and non-randomised healthcare intervention studies. https://www.amstar.ca/

CONSORT / SPIRIT-CONSORT - standards intended to improve complete and transparent reporting of randomised trials and trial protocols. https://www.consort-statement.org/

ICMJE Recommendations - principles covering authorship, disclosure of conflicts and prospective clinical-trial registration. https://www.icmje.org/recommendations/browse/roles-and-responsibilities/author-responsibilities--conflicts-of-interest.html

EFSA / European Commission - European scientific assessment and legal framework governing nutrition and health claims and the EU Register of authorised and non-authorised health claims. https://www.efsa.europa.eu/en/topics/topic/health-claims https://food.ec.europa.eu/food-safety/labelling-and-nutrition/nutrition-and-health-claims_en

END OF DOCUMENT