Two-stage 3D ResNet pipeline detects neonatal brain pathologies from MRI with 0.919 AUC
Dr_Alex_Crimi · x · 2026-10-30
A paper in Frontiers in Neuroimaging presents a hierarchical two-stage deep learning pipeline mimicking clinical reasoning on paired T1/T2 neonatal brain MRI to distinguish myelination delay, HIE, and hydrocephalus.
- Stage 1 (brain-age gap screening): a 3D ResNet-18 regressor predicts brain age with R² = 0.941; a gap threshold flags myelination delay typical of preterm infants.
- Stage 2 (structural pathology): a second 3D ResNet-18 performs 3-class classification (normal / HIE / hydrocephalus), achieving AUC 0.919 ± 0.024.
- Dataset: 566 normal, 56 HIE, 38 hydrocephalus, 27 preterm cases plus combined preterm overlaps.
- Paper, models, and code are publicly released.
Related event: Two-stage 3D ResNet screens neonatal brain disorders from MRI(2 posts)→
More from Research
- Gaussian head on flow policies: 0.02% of data, similar success, up to 2x faster inference — francoisfleuret · 2026-10-30
- Northwestern's Hullman: Verification is not knowledge — AI science needs humans who can answer for results — JessicaHullman · 2026-10-30
- Humansand launches &Tutor, a tutor model trained by teaching a human-like learner model — techNmak · 2026-10-30
- CATT benchmark shows attribution methods fail to trace model answers to training data — cindy_x_wu · 2026-10-30
- Launch: personalized cancer vaccine platform to manufacture 100+ mRNA vaccines — jajoosam · 2026-10-30
- Benchmark sweep shows greedy decoding tanks coding accuracy: 70% at temp 0 vs 82.7% at temp 1.0 — Brief-Tap-6616 · 2026-10-30