Statistical Geometry and Information Dynamics on Hyperspherical Surfaces

TL;DR

This paper explores the statistical geometry of random chords on n-dimensional spheres, revealing critical thresholds, geometric-informational dualities, and dimension-dependent properties affecting high-dimensional inference.

Contribution

It provides explicit formulas for chord length distribution, Fisher information, and characteristic functions across dimensions, uncovering new geometric and informational phenomena in high-dimensional spheres.

Findings

Critical threshold at dimension 19 for geometric concentration

Fisher information minimized at dimension 7, where sphere volume is maximized

Characteristic function differs between even and odd dimensions, reflecting harmonic structure

Abstract

We study the statistical geometry of random chords on n-dimensional spheres by deriving explicit analytical expressions for the chord length distribution and its associated structural properties. A critical threshold emerges at dimension 19, marking the transition from curvature-dominated variability to high-dimensional concentration, where interpoint distances become nearly deterministic and probabilistic diversity collapses into geometric uniformity. We further derive a closed-form expression for the Fisher information, showing that it is inversely proportional to the square of the radius and varies non-monotonically with dimension. Notably, it attains a minimum at dimension 7, coinciding with the dimension at which the volume of the unit sphere is maximized. This reflects a unique regime of maximal spatial diffuseness and minimal inferential sensitivity, arising from the interplay between weakening curvature and still-latent concentration. The alignment between volumetric and statistical extrema reveals a deeper duality between geometry and information. We also analyze the characteristic function, which exhibits a dichotomy: in even dimensions, it takes rational-exponential form, while in odd dimensions it involves Bessel and Struve functions. This distinction reflects differences in harmonic structure and boundary regularity across dimensions. Together, these findings show how curvature and dimension jointly regulate statistical efficiency on hyperspherical domains, with implications for geometric inference and high-dimensional learning.