[Announcement] neuro-analyzer v0.1.0 — Quantitative re-analysis pipeline for pediatric structural MRI

I’m releasing neuro-analyzer, an open-source Python pipeline for quantitative re-analysis of pediatric structural brain MRI. It orchestrates FastSurfer and dcm2niix for segmentation and regional volumetry, with LLM-assisted visual analysis (Anthropic Claude API) producing a quantitative report intended for research, education, and clinical discussion with specialists.

Intended audience: pediatric neuroimaging researchers and educators working with structural T1-weighted MRI.

External validation status: a single healthy pediatric subject from OpenNeuro ds000228 (sub-pixar066). n=1 only — documented as such in the CHANGELOG. No cohort-level evaluation. Not a medical device, no formal clinical validation.

License: MIT. Maintenance SLA through December 2027, after which the repository will be archived as read-only on Zenodo.

Repository: github.com/lindsay-barret/neuro-analyzer
DOI: 10.5281/zenodo.20222509

Feedback, issues, and contributions welcome through the GitHub repository. Particularly interested in hearing from researchers working on malformations of cortical development in pediatric populations.

— Lindsay Barret, Independent researcher (Guadalajara, México)
ORCID: 0009-0004-7411-3240

I feel like this is written and submitted by a bot because the ORCID points nowhere and the source code appears like it’s written by AI, the Github account is also new.

Also, on Github they wrote:

  • Technically-skilled parents seeking better understanding of their children’s studies (always as complement, never replacement, of medical evaluation)

Edit: I was curious where this was going, so I was going to go through the github repo in detail, and noticed a new commit: docs(readme): remove parents from target audience · lindsay-barret/neuro-analyzer@be4820c · GitHub

“Triggered by external feedback on Neurostars about audience framing.” this phrase sounds like they (or shall I say “it”) has reacted what to I’ve written. Pretty sure this is an AI now, and this situation reminds me of this incident: https://www.fastcompany.com/91492228/matplotlib-scott-shambaugh-opencla-ai-agent

I just hope the AI doesn’t go around ranting about me on some blogpost :rofl:

On the README line about parents you quoted — you have a point. Framing was ambiguous between research tool and consumer tool. Removed from both ES and EN versions: docs(readme): remove parents from target audience · lindsay-barret/neuro-analyzer@be4820c · GitHub

On the other three signals: yes, they’re real, and yes they’re addressable.

— ORCID is now properly populated: ORCID
— GitHub account was created specifically for this publication, kept separate from a personal account to maintain a clean publication identity. Single-author, developed independently from any institution.
— Code was developed with AI coding assistance (Claude Code, explicitly disclosed throughout the project). Before public release: manual review, pre-commit hardening (gitleaks, custom forbidden-pattern checks, ruff), a separate PHI/secrets audit, and CI testing across Python 3.10–3.13.

External validation is documented in the CHANGELOG as n=1 on OpenNeuro ds000228 (sub-pixar066). v0.1 of a research tool — limited, honest, openly archived. Technical critique is welcome.