Hybrid Neurovascular Model for Ischemia-Constrained Computation

Target journal: JMLR
This repo contains the paper and code for a stress-sensitive vascular model
coupled to pretrained, frozen neural networks. Neuron outputs are modulated
as whole units by local perfusion $\widehat{S}(v,t)$ (vessel size enters through $\widehat{S}$).
No Hebbian/plasticity terms are used in this article.

Layout

• paper/ — LaTeX source (NN_stroke_article_14_8_2_JMLR.tex, NN_stroke.bib)
• src/ — vascular graph + stress-driven dynamics; coupling gate g(S)
• configs/, scripts/, results/ — experiments and summaries

Reproduce paper PDF

The repo builds the PDF via GitHub Actions; artifacts appear in the Actions tab.

Build status