Brain Connectomes Navigability in Hyperbolic Space

Abstract

This study evaluates navigability in brain connectomes embedded in 2D hyperbolic geometry using greedy routing—a decentralized navigation protocol. Applying this framework to four empirical connectomes—spanning species, scales, and node types—we assess navigability via success rate, topological stretch, and two geometrical stretch variants. We introduce a novel geometrical stretch definition based on accumulated hyperbolic distance along minimal-cost paths, demonstrating improved consistency with topological stretch and potentially better reflection of the underlying geometry. Our results confirm hyperbolic geometry’s effectiveness as a latent space for efficient brain information transmission, providing further perspective on network science, complex systems, and brain connectivity.