AI Paper Atlas

Research map / Probabilistic and scientific modeling / Molecules and proteins

Explainable Brain Age Prediction using coVariance Neural Networks

Saurabh Sihag, Gonzalo Mateos, Corey McMillan, Alejandro Ribeiro

NeurIPS 2023 · Probabilistic and scientific modeling › Molecules and proteins · approach: Statistical analysis of ML practice · 6 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Use VNNs trained on healthy-control chronological age instead of randomly initialized VNNs (ablation)VNN brain age · OASIS-3 regional residuals AD+ vs HCrobustness of anatomical interpretabilityImproved
Use randomly initialized VNNs instead of trained ones (ablation)VNN brain age · OASIS-3 correlation of regional residuals with CDRconsistency of correlationHurt
Estimate anatomical covariance from HC group only instead of combined HC and AD+ (ablation)VNN brain age · OASIS-3 Delta-Age HC vs AD+group difference in Delta-AgeMixed
Estimate covariance matrix from a variable composition of HC and AD+ individuals (ablation)VNN brain age · OASIS-3 regional profile for Delta-Agestability of regional profileNo clear effect
Evaluate on independent ADNI-1 dataset with VNNs trained on OASIS-3VNN brain age · ADNI-1Delta-Age by cohort: 0 \pm 3.92 -> 3 \pm 5.74 -> 4.49\pm 5.34Improved
Compare Delta-Age of AD+ against HCVNN brain age · OASIS-3Delta-Age (years): 0\pm 2.83 -> 3.54\pm 4.49Improved
Higher dementia severity (CDR sum of boxes > 4 vs <= 4)VNN brain age · OASIS-3 AD+mean Delta-Age (years): 2.42 -> 6.04Improved
Evaluate on complete dataset rather than held-out test setVNN chronological age prediction · OASIS-3 HCMAE (years): 5.82 \pm 0.13 -> 5.44\pm 0.18No clear effect

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy