Information-Geometric Selection in T-Cell Receptor Cascades: How Noise Shapes Ligand Discrimination

Abstract

We model T cell recognition as Bayesian inference using a universal, unifying framework for existing T-cell phenotypic mechanisms. Antigen discrimination peaks when stochastic encounter durations match internal receptor kinetics. Irreversible commitment steps act as informational filters, maximising mutual information and explaining how physical time constraints shape immune receptor architectures.