Cortex Atlas: A Consolidated Multilingual Co-Authorship Network of the Global Neurosurgery Workforce (1946-2025)

The Cortex Atlas is a curated co-authorship network of the global neurosurgery workforce, derived from OpenAlex bibliographic metadata between 1946 and 2025. It encompasses 397,254 seed authors (342,570 after identifier consolidation), 18,629,871 indexed works, and 8,484,899 unique co-author pairs derived from 25,056,414 pairwise authorship events.\n\nThis Zenodo record contains the canonical post-consolidation data layer for the project: the position-weighted edge table, per-author network metrics (degree, weighted degree, k-core, PageRank, clustering coefficient), the consolidated clinical-neurosurgeon filter (v2 term list, 37,085 authors), the seed-author table, the subspecialty-classifier outputs (29 labels, ~7.99% corpus coverage / ~53% restricted to neurosurgery-affiliated papers), the OpenAlex-indexed abstracts (66% coverage), the NetworkX graph object, the DuckDB columnar database backing the Cortex Atlas web application, and the author-merge registry (329 merge groups consolidating 700 OpenAlex identifiers).\n\nFor the analysis pipeline that produces these files (data harvest, two-sided author-identity disambiguation, subspecialty classification, network construction, per-author metric computation, web-application preprocessing), see the GitHub repository linked under Related Identifiers. Access to the GitHub repository is granted upon academic-email request via the issue-template workflow documented in the repository's CODE_AVAILABILITY.md; reviewers handling the associated Scientific Data manuscript receive read access automatically when the journal forwards the manuscript identifier to the corresponding author.\n\nThe accompanying methods paper (Alawieh, Scientific Data, in submission) describes the full pipeline in detail. A companion clinical-audience announcement in the Journal of Neurosurgery introduces the Cortex Atlas interactive web application to the neurosurgery community.