These records govern the material claims retained after the first repository audit. Approval means approved for the qualified wording shown here. It does not mean peer review, institutional approval, or validation of the CAI.
EXT-KOSMYNA-2025-001
External Preprint
In a 54-person preprint, the LLM-assisted group showed the weakest EEG connectivity of three tool conditions and struggled to quote its own work. The study does not establish long-term causal harm or population-wide effects.
- Public location
- / · HeroBrain, DebtReadout, CasesBand
- Confidence
- moderate; causal wording not allowed
- Versions
- External preprint v2, 2025-12-31 · Not applicable
- Review
- Approved 2026-07-17; review by 2026-10-17
Known limitations
Preprint, small and task-specific sample, only 18 participants completed session 4, and the design does not establish population-wide or long-term causal effects.
EXT-B2Q-2026-001
External Preprint
Brain2Qwerty v2 reports 61% mean and 78% best-participant word accuracy on briefly memorized typed sentences decoded from MEG recordings.
- Public location
- / · CasesBand
- Confidence
- moderate; causal wording not allowed
- Versions
- External v2 project report · Brain2Qwerty v2
- Review
- Approved 2026-07-17; review by 2026-10-17
Known limitations
Healthy volunteers typed memorized sentences in a costly MEG system; the v2 dataset is embargoed until journal publication and the result is not ready for routine clinical use.
EXT-WILLETT-2023-001
External Peer-Reviewed Evidence
An invasive speech neuroprosthesis decoded attempted speech at 62 words per minute with a 23.8% word error rate in one participant with ALS.
- Public location
- / · CasesBand
- Confidence
- high; causal wording not allowed
- Versions
- Published article · External neuroprosthesis model
- Review
- Approved 2026-07-17; review by 2027-01-17
Known limitations
Invasive system and single-participant result; it does not establish general performance across patients or non-invasive settings.
PROD-CAI-DEMO-001
Product Demonstration
The browser-only assessment demo selects from 16 illustrative items using a 2PL information function and reports an unnormed estimate with an uncertainty band. It does not produce a validated CAI score.
- Public location
- /assessment · CaiAssessment
- Confidence
- high; causal wording not allowed
- Versions
- cai-demo-2pl-v1 · cai-demo-model-v1
- Review
- Approved 2026-07-17; review by 2026-10-17
Known limitations
Illustrative fixed item parameters, no norming sample, no field calibration, no validated construct score, and no longitudinal persistence.
PROD-TRAJECTORY-DEMO-001
Product Demonstration
The displayed trajectories are synthetic demonstration scenarios used to explain measurement uncertainty. They are not participant observations, pilot results, norms, or evidence of change.
- Public location
- / · InstrumentExplorer
- Confidence
- high; causal wording not allowed
- Versions
- trajectory-ui-v1 · No fitted model
- Review
- Approved 2026-07-17; review by 2026-10-17
Known limitations
Values were selected to illustrate interface states and were not generated by a simulation study or observed in people.
HYP-COGNITIVE-AUTONOMY-001
Theoretical Hypothesis
Cognitive autonomy is a proposed multidimensional construct, not an established diagnosis, intelligence score, or measure of human worth. Its reliability, validity, fairness, and sensitivity to change have not yet been empirically established.
- Public location
- /evidence · EvidencePage
- Confidence
- high; causal wording not allowed
- Versions
- Proposed construct specification v0 · No validated model
- Review
- Approved 2026-07-17; review by 2026-10-17
Known limitations
Construct boundaries, factor structure, discriminant validity, fairness, and longitudinal interpretation remain hypotheses awaiting simulation, expert review, and human studies.
PROD-METHOD-STATUS-001
Product Demonstration
The validation architecture describes proposed methods. Phroneme has not yet produced reliability, validity, fairness, normative, or longitudinal performance estimates, and no confirmatory study is preregistered.
- Public location
- /method · MethodPage
- Confidence
- high; causal wording not allowed
- Versions
- Proposed validation architecture v0 · No validated model
- Review
- Approved 2026-07-17; review by 2026-10-17
Known limitations
The page is a technical design proposal. Implementation, expert review, ethics review, preregistration, data collection, and analysis have not occurred.
PROD-MF90-SESSION-001
Product Demonstration
The public Phroneme 90 plan builder creates a session-only wellness plan preview. It does not create an account, store a health record, diagnose a condition, produce a wellness score, or establish cognitive change.
- Public location
- /phroneme-90/plan · PlanBuilder
- Confidence
- high; causal wording not allowed
- Versions
- phroneme90-plan-preview-v1 · Deterministic rules; no predictive model
- Review
- Approved 2026-07-19; review by 2026-10-19
Known limitations
The preview cannot provide durable adherence tracking, longitudinal trends, individualized medical guidance, or validated outcome estimates.
EXT-USPAG-2018-001
External Guideline
U.S. physical activity guidance recommends that adults work toward 150 to 300 minutes of moderate aerobic activity per week plus muscle-strengthening activity on two days, with activity adapted to ability and health context.
- Public location
- /phroneme-90/foundations · FoundationsPage, PlanBuilder
- Confidence
- high; causal wording not allowed
- Versions
- Second edition guideline · Not applicable
- Review
- Approved 2026-07-19; review by 2027-01-19
Known limitations
Population guidance is not an individualized exercise prescription; symptoms, injury, pregnancy, disability, and clinical conditions may change what is appropriate.
EXT-CREATINE-2024-001
External Peer-Reviewed Evidence
A 2024 meta-analysis of 16 randomized trials involving 492 adults found a possible modest memory benefit from creatine monohydrate, while overall cognition and executive function did not improve significantly and certainty was low for several outcomes.
- Public location
- /phroneme-90/supplement · SupplementPage
- Confidence
- moderate; causal wording not allowed
- Versions
- Published meta-analysis plus corrigendum · Not applicable
- Review
- Approved 2026-07-19; review by 2026-10-19
Known limitations
The review was small and heterogeneous; it does not establish IQ improvement, dementia prevention, universal cognitive benefit, individualized dosing, or safety for every person.
PROD-MF90-DATA-001
Product Demonstration
Apple Health, Fitbit, WHOOP, and blood-work uploads are not connected in the public Phroneme 90 release. They remain blocked until the required consent, authorization, encryption, retention, export, deletion, and audit controls are implemented and tested.
- Public location
- /phroneme-90/data · DataPage
- Confidence
- high; causal wording not allowed
- Versions
- phroneme90-data-boundary-v1 · Not applicable
- Review
- Approved 2026-07-19; review by 2026-10-19
Known limitations
The route documents a release boundary only; it does not implement OAuth, HealthKit, lab extraction, token storage, or durable health-data operations.