AI Paper Atlas

Research map / Vision-language and multimodal / Human and embodied modeling

Only Brains Align with Brains: Cross-Region Patterns Expose Limits of Normative Models

Larissa Höfling, Matthias Tangemann, Lotta Piefke, Susanne Keller, Matthias Bethge, Katrin Franke

ICLR 2026 · Vision-language and multimodal › Human and embodied modeling · approach: fMRI visual decoding

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
evaluate diverse vision models with standard RSA/LP alignment scores47 pretrained vision models · BOLD Moments, visual ROIsmodel-brain alignmentNo clear effect
use linear predictivity instead of RSA as the alignment measurebest-ranked models · BOLD Moments, visual ROIsalignment relative to brain-brain referenceMixed
add alignment pattern similarity requirement to the alignment score criterionmodel-ROI candidate selection · BOLD Moments, visual ROIsnumber of candidate brain-aligned modelsImproved
compare model-derived alignment patterns with brain-derived patternsall evaluated models · BOLD Moments, visual ROIsalignment pattern similarityHurt
compare equivalently aligned models by their cross-region alignment patternstop RSA models · BOLD Moments, visual ROIsalignment patternImproved
compare fMRI-derived alignment patterns across subjectsbrain-brain alignment patterns · BOLD Moments, visual ROIsalignment pattern similarity across subjectsImproved
compare fMRI-derived patterns to structural connectivity patternsfMRI-derived alignment patterns · HCP structural connectivity vs BOLD Momentsalignment pattern similarityMixed
normalize scores to a lower estimate of intersubject consistency (ablation)alignment scores · BOLD Moments, visual ROIsdiscriminative powerImproved
train on varied Taskonomy tasks instead of recognition or video objectivesResNet-50 Taskonomy models · BOLD Moments, visual ROIsRSA alignment: 0.06 vs 0.22 (V1, class-object vs V-JEPA 2 giant-384)Hurt

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