Epidemic Spreading

Epidemic spreading is the dynamics by which contagion — whether biological, informational, or behavioral — propagates through a network of connected agents. The study of epidemic dynamics treats the network as a transmission infrastructure: the topology of connections determines not just whether an epidemic occurs, but its speed, its final size, and its vulnerability to intervention. This is network science at its most consequential: the mathematics of contagion is the mathematics of revolution, rumor, disease, and innovation, all governed by the same structural equations.

The foundational model is the SIR framework (Susceptible-Infected-Recovered), but the network-theoretic extension replaces the homogeneous-mixing assumption with a topology-dependent threshold. For a disease with transmission probability β and recovery rate γ, the epidemic threshold is determined by the spectral radius of the network's adjacency matrix. When β/γ exceeds the inverse of the largest eigenvalue, the epidemic spreads; when it falls below, the infection dies out. This is why the Alon-Boppana bound matters for epidemic dynamics: it constrains the spectral properties of sparse networks, which in turn constrains their epidemic vulnerability.

The same mathematics applies to misinformation cascades, financial contagion, and social norm adoption. The difference between a biological epidemic and an information epidemic is not the dynamics but the state variables: infections become beliefs, recovery becomes skepticism, and immunity becomes epistemic vigilance. The network does not care what it is transmitting.

Epidemic spreading is not a pathology of networks. It is a property of networks — and the only way to prevent it is to change the topology.,