Connectomic Surgical Risk Atlas (CSRA), Version 1: Computational Network-Vulnerability Atlas and Reproducibility Resources

Connectomic Surgical Risk Atlas (CSRA), Version 1 Computational Network-Vulnerability Atlas and Reproducibility Resource Overview The Connectomic Surgical Risk Atlas (CSRA-V1) is a versioned computational neuroscience resource developed to investigate how focal structural brain injury affects network organization, connectivity, and conditional network criticality. The framework distinguishes four complementary dimensions: direct anatomical involvement, structural disconnection, conditional network vulnerability, and potential structural reserve. CSRA-V1 provides a reproducible reference framework for studying lesion-induced network alterations and developing future patient-specific neurosurgical risk models. It is a computational research resource, not a clinically validated predictor of postoperative neurological deficits. Atlas Construction CSRA-V1 uses an HCP1065-derived normative structural connectome with 360 HCP-MMP cortical parcels, 4,928 supported structural edges, and 24 pathway-family summaries. The virtual-lesion analysis comprises 54 synthetic tumor configurations and 270 physically relocated control masks. Network trapping points (NTPs) characterize graph-dependent changes in network criticality after simulated injury. The resource includes frozen aggregate atlas tables, secondary graph-topology analyses, synthetic lesion masks, spatial-control results, computational scripts, quality-control records, and file-integrity checksums. Principal Scientific Findings The analyses demonstrate that conditional network criticality is computationally reproducible and associated with baseline structural network topology. However, recurrent network-critical locations did not demonstrate preferential anatomical enrichment under the available physically relocated and tissue-matched spatial controls. In the independent FSL FAST tissue-matching analysis, the broader matching criterion retained 16 of 18 anatomical sites. The observed original-location recurrence count was 4,094 pairs, compared with a matched-control mean of 4,239.94 pairs (conditional upper-tail proportion 0.742). These findings distinguish computational network vulnerability from anatomically specific surgical risk. They do not establish the probability of neurological deficits following surgical resection. Reproducibility and Data Availability The archived release includes computational outputs, supporting scripts, synthetic lesion configurations, spatial-control analyses, documentation, and SHA-256 integrity records. The aggregate tables can be checked using the included "verify_dataset.py" script. Re-execution of the R9 tissue-matching analysis requires independently obtained anatomical reference images and the dependencies documented in the archive. The release does not include individual patient data, raw HCP1065 tractography, original third-party anatomical templates, or external FSL FAST tissue-probability images. Complete upstream reconstruction and clinical model refitting are outside the scope of this curated dataset. Scientific Applications CSRA-V1 supports research into lesion-induced network reorganization, structural network vulnerability, computational validation, and the development of prospective neurosurgical prediction models. Future development will focus on patient-specific imaging, intraoperative mapping, longitudinal neurological and cognitive outcomes, and independent clinical validation. CSRA Online is proposed as a future collaborative, versioned research infrastructure. It is not presented as an already deployed clinical platform. Source Attribution and Acknowledgments The normative structural connectivity resource is derived from the HCP1065 population-averaged tractography atlas developed by Fang-Cheng Yeh and collaborators. Principal references: Yeh F-C. Population-based tract-to-region connectome of the human brain and its hierarchical topology. Nature Communications. 2022;13:4933. https://doi.org/10.1038/s41467-022-32595-4 Yeh F-C, et al. Population-averaged atlas of the macroscale human structural connectome and its network topology. NeuroImage. 2018;178:57–68. https://doi.org/10.1016/j.neuroimage.2018.05.027 The HCP1065 atlas uses data from the WU-Minn Human Connectome Project, led by David Van Essen and Kamil Ugurbil, supported by NIH Blueprint funding (1U54MH091657) and the McDonnell Center for Systems Neuroscience at Washington University. Original source: https://brain.labsolver.org/hcp_trk_atlas.html WU-Minn HCP data-use terms: https://www.humanconnectome.org/study/hcp-young-adult/document/wu-minn-hcp-consortium-open-access-data-use-terms FSL FAST reference images used in the independent tissue-matching analysis were obtained from the MNITemplate research repository: https://github.com/Jfortin1/MNITemplate The archive includes a separate source-attribution and rights document. Reuse and redistribution must respect the applicable original source licenses, attribution requirements, and data-use terms. Authors Hussameldin Metwali, MD; Amir Samii, MD International Neuroscience Institute, Hannover, Germany. Version: CSRA-V1 Research areas: Neurosurgery; Connectomics; Structural Brain Networks; Computational Neuroimaging; Brain Mapping; Reproducible Research.