Language-aligned models and structured scene descriptions reveal sensitivity to compositional scene structure in the high-level visual cortex
Karim Rajaei, Arian Afshar, Radoslaw Martin Cichy, Hamid Soltanian-zadeh A
The paper shows that structured scene descriptions, not just object co-occurrence, predict high-level visual cortex responses better, and language-aligned models capture this structure.
bioRxiv · Jul 28, 2026 · PreprintNeuroscience