Hyper-Brain 0.3.0

Display changes

Verification

Screenshot provenance

All 26 images in website/public/gallery are actual local application captures. They use synthetic data, not patient records or disease findings. The local application UI is currently Chinese; the introduction website and README are bilingual.

examples/prepare_gallery.py generates 240 observations with numpy.random.default_rng(20260910 + n_rois): seven independent Gaussian latent series plus Gaussian ROI noise, then Pearson correlation. AAL 116 uses seven explicitly specified illustrative hyperedges (including two IDs with identical membership). AAL 90 and Schaefer 200/400 use FC-profile kNN with k=3. Ordinary graphs use density 0.04. All fixtures pass through public upload, analysis and atlas-binding endpoints.

frontend/scripts/capture-gallery.cjs saves the images and manifest.json. The manifest records the atlas version and space, matrix checksum, configuration, exact selected members, layer counts and image checksum. ROI coordinates come from the selected atlas. Display geometry does not establish a biological network or an anatomical fiber trajectory.

Rendering was checked in local Chromium with software WebGL. The captured frame durations are observations on this machine, not a cross-device performance benchmark. A dense 400-region view deliberately retains the full structure; label crowding remains visible rather than being hidden by automatic deletion.

Documentation design

The English and Chinese READMEs use a concise introduction, installation near the top, a real workbench image, a side-by-side style grid, an expandable gallery and a compact capability table. The September 10, 2026 Trending pages for i-have-adhd and teamai-cli informed information hierarchy only. No ranking, usage, endorsement or clinical-performance claims were transferred to this project.