NeuroRisk separates source records, normalized entities, comparative scores and report synthesis. A score is never presented as a probability of clinical, regulatory, commercial or investment success unless a separately validated probabilistic model exists.
ClinicalTrials.gov records normalized by NCT ID, sponsor, status, phase, condition, intervention and geography. The full curated corpus is preserved for auditability, but only records classified as in_scope or related feed analytical rankings, sponsor intelligence and enterprise reports; out_of_scope records remain traceable but are excluded from those calculations.
Broader ClinicalTrials.gov discovery coverage kept separate from the curated layer. Raw protocol JSON, source query and MeSH links are retained.
PubMed/NCBI and OpenAlex publication, author and institution metadata. Country geography is normalized conservatively and coverage is disclosed.
FDA Orange Book patent/regulatory evidence is U.S.-specific. It is not represented as worldwide patent clearance or global regulatory coverage.
Each component is log-normalized against the maximum value in the current indexed country set: lscore(v,m) = 100 × log(1+v) / log(1+m).
The current country-layer confidence score is a deterministic data-coverage indicator: base 35 points, +25 when clinical evidence exists, +25 when institution evidence exists and +15 when publication evidence exists, capped at 100. It measures presence of evidence layers; it is not calibrated predictive probability.
The public market layer ranks the top 50 evidence-bearing countries by the current Research Score, then clinical-trial volume. “Rank” therefore means comparative evidence position inside the current NeuroRisk corpus.
Every V2 report is generated from one timestamped database snapshot and carries methodology version NR-REPORT-V2.0. The automated QA requires snapshot metadata, source-linked evidence, limitations disclosure, methodology version, and consistency with current country metrics. Legacy synthetic ML confidence and the former “Estimated Strategic Market Signal” proxy are excluded.
No indexed corpus is guaranteed exhaustive of all global neuroscience research. Source coverage and update cadence vary. Entity resolution can miss or conservatively separate aliases. Country attribution in literature can be incomplete. Scores are relative to the indexed corpus and can change as evidence expands.
Research strategy, market screening, competitive intelligence, hypothesis generation and evidence discovery. Reports do not replace medical judgment, regulatory counsel, freedom-to-operate analysis, patent counsel, actuarial validation or investment due diligence.
Current source counts, geography, quality gates and registry status are published separately.