Reproduction bundles for the golgi peripheral nerve stimulation modeling platform

Reproducible golgi Study Bundles Self-contained, integrity-hashed golgi study bundles—and the full working dataset—reproducing the peripheral-nerve stimulation simulations behind Figures 4–8 of the associated manuscript. Reproducing the Results There are two ways to reproduce the results using different files in this record: Verify or quick-look a study (light): Each *.golgi.zip is a complete, replayable golgi project. You can verify it byte-for-byte or render golgi's built-in result figures without downloading the large working dataset. Regenerate the composite paper figures (full): golgi_paper_dataset.tar.gz contains the approximately 27 GB working tree used by the figure-generation scripts. Extract it at the root of a golgi checkout and run paper_figs/make_figures.py. Study Bundles Each *.golgi.zip is a complete golgi project containing: Source geometry Multi-region tetrahedral mesh Anisotropic finite-element extracellular lead fields (for each electrode montage/configuration) Fiber population with full 3D trajectories Per-contact recordings A MANIFEST.json file recording: SHA-256 hash of every file Exact golgi version Frozen dependency list Because the bundles already include the field solution and recordings, recruitment and selectivity can be reproduced without re-solving the finite-element model (FEM). The bundled mesh and fiber population also support complete re-solving and parameter re-sweeps. Importing a Study GUI File → Import Study → select the .golgi.zip file Python API golgi.Study.import_bundle("fig07_rabbit_branching.golgi.zip") Automatic download python paper_figs/fetch_bundles.py This script downloads every study bundle in this record and verifies its checksum automatically. Verifying Integrity Verify downloaded files using: shasum -a 256 -c CHECKSUMS.sha256 or replay the study: golgi replay <bundle.golgi.zip> (golgi.projects.replay.replay_study) The replay process re-hashes every file in the bundle. Every study bundle was replay-verified during packaging, confirming that every file matched its recorded hash. BUNDLES.json additionally records the file count, size, and digest for every bundle. Regenerating the Composite Figures golgi_paper_dataset.tar.gz contains the complete working dataset used by the paper figure scripts, including: paper_figs/out/data paper_figs/out/_intermediate paper_figs/out/renders results_golgi/duke_meshes From the root of a golgi checkout: tar xzf golgi_paper_dataset.tar.gz python paper_figs/make_figures.py [all | fig05 fig06 ...] Generated figures are written to: paper_figs/out/figures/{png,pdf,svg}/ Unlike the individual study bundles, make_figures.py rebuilds the complete multi-panel figures used in the manuscript. By comparison, golgi replay and golgi figure operate on individual study bundles. SHA-256 24582a9ac3d0e616cc465b99039218c457c580f0bffa43db239875d8d87d7843 Contents fig04a_dogvns_validation.golgi.zip In vivo dog cervical VNS validation (ASCENT masks → golgi pipeline; separated LivaNova-style cuff with perineurium contact impedance). Supports Figure 4 (dog VNS panel). fig04b_nrv_life_figure_data.zip NRV LIFE intrafascicular benchmark containing cached figure data only. This idealized synthetic cylinder study is not re-meshable with the current pipeline; the figure is reproduced from cached recruitment and threshold files. Supports Figure 4 (NRV LIFE panel). fig04c_bucksot_validation.golgi.zip Bucksot multifascicular cuff validation (five-fascicle nerve with circumferential and inverted electrode montages). Supports Figure 4 (Bucksot panels). fig05_swine_cervical_vagus.golgi.zip Image-derived swine cervical vagus model demonstrating multifascicular selectivity. Supports Figure 5. fig06_human_cervical_vagus.golgi.zip Image-derived human cervical vagus model demonstrating multifascicular selectivity. Supports Figure 6. fig07_rabbit_branching.golgi.zip Real 3D branching rabbit vagus reconstructed from micro-CT, demonstrating cardiac branch targeting using a multi-contact cuff. Supports Figure 7. fig08_human_scb_branching.golgi.zip Real 3D human cervical vagus reconstructed from micro-CT with multi-region epineurium and endoneurium, demonstrating branch selectivity. Supports Figure 8. supplementary_controls.zip Supplementary control analyses derived from the Figure 7 and Figure 8 study bundles, including: Matched-perineurium control Streamline and branch-detection convergence analyses Extruded versus 3D reconstruction comparison Reconstructed nerve field galleries Includes scripts, cached data, and generated figures. golgi_paper_dataset.tar.gz Complete working dataset for regenerating the manuscript figures using paper_figs/make_figures.py. Includes: Raw meshes FEM lead fields Fiber sweeps Cel-shaded renders (paper_figs/out/{data,_intermediate,renders}) Source geometry (results_golgi/duke_meshes) Extract at the root of a golgi checkout. golgi_comsol_handover.tar.gz Independent COMSOL reference dataset used by the Figure 4 finite-element cross-validation panels (a, b) — the student ran these models in COMSOL separately, and make_figures.py compares golgi's lead fields against them. Includes: COMSOL models and their meshes Evaluation points and per-contact geometry Exported COMSOL results (M1–M3) The golgi-vs-COMSOL comparison scripts (comsol_handover/) Extract at the root of a golgi checkout, alongside golgi_paper_dataset.tar.gz. Required for Figure 4. The other figures do not need it. Code and Documentation GitHub repository: https://github.com/CellularSyntax/golgi See the wiki pages: Reproducing the Paper Reproducible Study Bundles