Geometric constraints in the development of primate extrastriate visual cortex

Abstract

Sensory systems are organized into topographic maps that shape information flow and computation across cortical circuits. Although the mechanisms establishing primary sensory maps are well characterized, how higher-order maps emerge across neocortex is unclear. Because the probability and strength of cortical connections fall off steeply with distance along a folded surface, the geometry of the cortex may be a major factor shaping the organization of higher-order maps. To test this in a well-characterized sensory system, we develop a growth model embedded in the folded surface geometry of the macaque visual cortex, using fMRI-defined V1 retinotopy as the sole functional anchor. Cortical organization emerges through algorithmic growth from primary visual cortex, governed by distance-dependent activity correlations and competition among developing projections. Without imposing areal boundaries, map orientations, or predefined topographic layouts, this process generates multiple retinotopic maps with systematic mirror reversals and smooth gradients that reflect key structural features of fMRI-derived extrastriate maps. Parameters estimated on a population template generalize across individual macaques, while individual cortical geometry accounts for fine-scale map variation around a common retinotopic scaffold. These results suggest that conserved growth rules acting on folded cortical surfaces produce stereotyped higher-order retinotopic organization under minimal explicit specification.