Source-linked AI summary
Epidemics on Interconnected Networks
M. Dickison, S. Havlin, H. E. Stanley
TL;DR
The paper asks how epidemics behave when populations form interconnected but locally distinct networks. Using an SIR process on coupled network models, it identifies strongly and weakly coupled regimes. Strong coupling produces system-wide simultaneous epidemics, whereas weak coupling permits a mixed phase with disease confined to one network.
Problem
The paper asks under what conditions an epidemic spreads only within sub-networks and under what conditions it spreads across an entire interconnected system.
Method
The authors study the susceptible-infected-recovered (SIR) process on interconnected network systems and classify their coupling regimes.
Results
Strongly coupled systems produce epidemics across the entire interacting system, while weakly coupled systems can exhibit a mixed phase with an epidemic in only one network.
Takeaways & Limitations
The coupling regime determines whether an outbreak in one community signals epidemic activity across neighboring networks or remains largely localized.
Abstract
from arXiv · showhide
Populations are seldom completely isolated from their environment. Individuals in a particular geographic or social region may be considered a distinct network due to strong local ties, but will also interact with individuals in other networks. We study the susceptible-infected-recovered (SIR) process on interconnected network systems, and find two distinct regimes. In strongly-coupled network systems, epidemics occur simultaneously across the entire system at a critical infection strength $β_c$, below which the disease does not spread. In contrast, in weakly-coupled network systems, a mixed phase exists below $β_c$ of the coupled network system, where an epidemic occurs in one network but does not spread to the coupled network. We derive an expression for the network and disease parameters that allow this mixed phase and verify it numerically. Public health implications of communities comprising these two classes of network systems are also mentioned.
I. INTRODUCTION
The paper examines epidemics in interconnected networks, asking when disease remains confined to one component and when it spreads across the entire system. It defines strongly and weakly coupled regimes and identifies a mixed phase possible in weakly coupled systems.
- Motivation: Interconnected networks represent communities or systems with strong local ties plus links connecting them to other networks.Examples include social communities, transportation networks, and human-animal disease transmission systems.
- Research question: The paper asks which network and interconnection conditions produce epidemics confined to one sub-network or spread across the entire system.The authors emphasize that identifying these conditions matters for understanding and managing epidemic processes.
- Contribution: The authors define strongly and weakly coupled network regimes and determine the interaction strength separating them.The regimes depend on the parameters of the individual networks and their interconnections.
- Main result: In strongly coupled systems, all networks are simultaneously disease free or part of an epidemic.An outbreak in one neighboring community is therefore cause for immediate concern in the other.
- Main result: In weakly coupled systems, a mixed phase can occur in which disease is epidemic in one network but absent from the others despite interconnections.This distinction motivates different public-health coordination needs across neighboring communities.
II. MODEL
The model constructs interconnected networks with specified degree distributions and studies disease spread using the susceptible-infected-recovered process. Epidemic thresholds are analyzed through transmissibility and percolation, allowing distinct epidemic phases across coupled networks.
- Network construction: The model considers two interconnected networks of equal size and can be extended to arbitrary numbers and sizes of networks.Networks A and B receive separate intranetwork degree distributions before interconnections are added.
- Network construction: Random internetwork connections generate uncorrelated systems with specified inter- and intra-network degree distributions.The presented results use random Poissonian degree distributions, although the construction method permits arbitrary distributions.
- Disease model: The SIR process assigns susceptible, infected, and recovered states, with infection probability β and recovery after time t_r.The process begins with one infected node while all others are susceptible.
- Threshold analysis: For a single network, the epidemic threshold follows (κ−1)T_β = 1, where T_β = 1−(1−β)^t_r is transmissibility.The mean number of secondary infections is N_I = (κ−1)T_β.
- Coupled-system phases: Interconnected systems can be epidemic in both networks, disease free in both, or in a mixed phase with activity in only one network.The phase boundaries are controlled by κ_A, κ_B, and κ_T, calculated for individual and fully coupled network structures.
III. STRONGLY-COUPLED NETWORK SYSTEMS
Strongly coupled systems are defined by sufficient internetwork connectivity for epidemics to emerge simultaneously across networks A and B. In this regime, internetwork links make the system behave like a single network and permit spread at lower infection strength than either network alone.
- Coupling criterion: For fixed intranetwork parameters, the critical interaction strength ⟨kAB⟩c separates strongly coupled systems from weakly coupled systems.The boundary is defined by whether κT exceeds or falls below κB.
- Epidemic onset: In strongly coupled systems, epidemics emerge simultaneously on networks A and B.The coupled-system threshold βc(κT) governs emergence on the giant component formed by the entire interconnected network.
- Epidemic onset: The disease spreads across the interconnected system as a single network because internetwork connections activate an epidemic before either network spreads independently.This regime therefore excludes the mixed phase in which only one network experiences an epidemic.
- Numerical verification: The largest infected cluster confined to one network decreases relative to the all-links cluster, indicating system-wide rather than local epidemic spread.This ratio is used to compare clusters formed with intranetwork links alone against clusters formed with all links.
- Epidemic threshold: Internetwork connections lower βc, allowing less virulent diseases to spread than would spread on either network alone.The coupled network’s critical value is smaller than both isolated-network thresholds βc(κA) and βc(κB).
- Special case: Identical intranetwork degree distributions always produce strongly coupled systems, whose phase diagram resembles that of a single network.For ⟨kA⟩=⟨kB⟩, the critical interaction strength is ⟨kAB⟩c = 0.
IV. WEAKLY-COUPLED NETWORK SYSTEMS
Weakly coupled networks can support a mixed phase in which an epidemic is established in the more strongly connected network while remaining localized in the other. Increasing infection strength or internetwork connectivity eventually produces an epidemic across both networks, while the mixed phase shrinks toward the strongly coupled regime.
- Phase behavior: The mixed phase is governed by separate thresholds, with βc(κB) < βc(κT) < βc(κA).The coupled-system threshold βc(κT) lies between the isolated-network thresholds when κB > κT and network B is more strongly connected.
- Network asymmetry: Interconnections initially affect epidemic spreading on the weaker network, while the stronger network’s epidemic remains localized there.Adding links cannot decrease epidemic spread; in the mixed regime, network A plays no role in spreading on network B.
- Phase behavior: A mixed phase occurs when network B becomes epidemic while network A contains only small infected clusters.For βc(κB) < β < βc(κT), network B has a finite infected fraction while network A does not sustain an epidemic.
- Coupling transition: The mixed phase becomes small and difficult to identify as networks approach the strongly coupled regime.The phase diagram contains disease-free, mixed, epidemic, and weak-to-strong coupling transition regions.
- Critical behavior: At βc(κB), network B shows critical survival scaling, whereas network A receives only infrequent, non-epidemic infections in the weakly coupled case.Network B follows the expected t^-1 survival decay, but network A’s survival curve cannot directly identify its epidemic threshold.
- Critical behavior: A non-zero survival probability gap identifies the mixed phase, and the gap vanishes upon entering the epidemic phase.The gap is defined as the minimum relative difference between the networks’ survival probabilities.
V. CONCLUSIONS
The paper distinguishes strongly and weakly coupled interconnected network systems and characterizes how epidemics spread across them. Strong coupling produces system-wide epidemics, whereas weak coupling permits a mixed phase in which spreading remains confined to the more intraconnected network; these boundaries were analyzed analytically and numerically.
- The paper introduces strongly coupled and weakly coupled regimes for interconnected network systems.
- In strongly coupled systems, epidemics occur across the entire interacting network system, with interconnections enhancing epidemic spreading.
- In weakly coupled systems, a mixed phase exists in which epidemics do not always occur across the full interconnected system.
- In the mixed phase, interconnections affect epidemic spreading across the less intraconnected network.
- The boundaries and behavior of the mixed phase were demonstrated analytically and numerically.
- Identifying which communities form strongly or weakly coupled systems could inform public policy and highlight epidemic dangers from increased human–animal interaction.