Modeling Coupled Epidemic-Information Dynamics via Reaction-Diffusion Processes on Multiplex Networks with Media and Mobility Effects

Guangyuan Mei, Yao Cai, Su-Su Zhang, Ying Huang, Chuang Liu, Xiu-Xiu ZhanIEEE Trans. Network Science & Engineering 2026 · DOI

TL;DR

This paper introduces a novel model that simultaneously simulates how infectious diseases spread and how information about them disseminates across complex social networks. By integrating factors like mass media, higher-order social interactions (beyond just pairs of people), and population movement, the study offers a more realistic picture of epidemic dynamics and suggests new strategies for containment.

What Problem Does This Paper Address?

Understanding and controlling infectious disease outbreaks is a critical challenge. While many models exist, they often simplify real-world complexities. For instance, they might focus only on direct person-to-person contact, neglecting how information spreads through groups or via media. Similarly, they might overlook how people move between communities, which is a major driver of disease spread. This paper tackles the problem of creating a more comprehensive model that captures the intricate interplay between disease transmission and information diffusion, considering these crucial real-world factors:

  • Information Diffusion: How awareness about a disease spreads, influenced by both direct contacts and mass media.
  • Disease Transmission: How the actual infection spreads through a population.
  • Higher-Order Interactions: Recognizing that people interact in groups (e.g., families, workplaces, social circles) and not just in pairs.
  • Population Mobility: How people moving between different locations (cities, regions) facilitates or hinders disease spread.
  • Mass Media Influence: The role of news, social media campaigns, and other broadcast information in shaping public awareness and behavior.

By integrating these elements, the authors aim to provide a more accurate representation of epidemic dynamics and identify more effective intervention strategies.

Methods

The researchers developed a coupled disease-information spreading model built upon a multiplex network framework. This means they used two interconnected layers to represent different aspects of the system:

  1. Virtual Information Layer: This layer models information diffusion using a simplicial complex network. Unlike traditional networks that only consider pairwise connections (edges), simplicial complexes can represent higher-order structures (like triangles, tetrahedrons, etc.), allowing for the modeling of group interactions. A dedicated media node is also included to represent the influence of mass media.
  2. Physical Contact Layer: This layer models disease transmission using a metapopulation network. Here, the population is divided into several subpopulations, and connections between them represent transportation routes, allowing for the simulation of individual mobility. A reaction-diffusion process is used to model disease spread within and between these subpopulations.

To analyze the model, they employed two main approaches:

  • Microscopic Markov Chain Approach (MMCA): This is a theoretical framework used to derive the dynamical equations and calculate epidemic thresholds. It models the state of each individual as a Markov process, considering their current state and the states of their neighbors.
  • Monte Carlo (MC) Simulations: These are used to validate the MMCA results and explore the model's behavior under various parameter settings. The simulations involve running the model many times to obtain statistically robust results.

The model considers four distinct states for individuals: Unaware-Susceptible (US), Unaware-Infected (UI), Aware-Susceptible (AS), and Aware-Infected (AI). The spreading process is divided into four stages at each time step: information diffusion, individual movement, disease transmission, and individual return.

Datasets

Not discussed in the paper. The study uses simulated networks and parameters rather than real-world datasets.

Key Insights and Ideas

  • Synergistic Effects: Mass media and higher-order social interactions (simplices) both significantly boost public awareness, which in turn helps to suppress disease transmission. This highlights the dual benefit of strengthening media campaigns and leveraging group communication.
  • Diminishing Returns of Higher-Order Interactions: While higher-order interactions are beneficial, their impact on suppressing disease transmission tends to weaken as the "order" of the simplex increases. This suggests that while groups are important, the benefit of very large, complex groups might not scale linearly.
  • Mobility's Nuanced Role: Increased connectivity within a metapopulation network (more pathways between communities) can actually reduce overall disease prevalence. This counterintuitive finding suggests that better-connected systems might distribute the disease more evenly, preventing localized outbreaks from becoming severe.
  • Strategic Migration: Guiding individuals to move towards less connected or more isolated subpopulations can be an effective strategy to mitigate epidemic spread, especially when combined with other interventions.
  • Media as a Threshold Shifter: The epidemic threshold (the point at which an epidemic can take hold) increases with media broadcasting strength, indicating that effective media campaigns can make it harder for diseases to spread widely.

Key Results

  • MMCA Validation: The theoretical MMCA predictions showed strong agreement with the results from extensive Monte Carlo simulations, validating the proposed model.
  • Information Suppresses Disease: The study found that increased public awareness (driven by media and group interactions) significantly reduces disease prevalence. For example, higher-order structures suppressed disease spread by at least 5% compared to a media-only scenario.
  • Media Enhances Awareness: Mass media broadcasting effectively increases the proportion of aware individuals, leading to a decline in infected individuals and raising the epidemic threshold.
  • Connectivity and Disease: Increasing the number of connections between subpopulations (higher 'm' in the BA model) led to a reduction in the proportion of infected individuals, suggesting that better-connected networks can lead to more balanced population distributions and less severe outbreaks.
  • Heterogeneity's Complex Impact: Subpopulation heterogeneity (differences in population size and connectivity) has a nonlinear impact on the epidemic threshold. In some cases, extreme heterogeneity can be beneficial for control, while in others, it can exacerbate spread.

Limitations and Future Work

The authors acknowledge several limitations:

  • Information Accuracy: The model assumes all disseminated information is accurate, neglecting the impact of misinformation or conflicting messages.
  • Simplified Disease Model: The use of a simple SIS (Susceptible-Infected-Susceptible) model might not fully capture the complexity of diseases with multiple stages or varying infectiousness (e.g., HIV/AIDS).
  • Homogeneous Subpopulations: The model assumes fully connected graphs within subpopulations, overlooking inherent heterogeneity in real social networks.
  • Cross-Layer Independence: Connections in the virtual information layer are assumed to be independent of physical subpopulation affiliations, which might not hold true in reality.
  • Static Networks: Both physical and virtual network topologies are treated as static, whereas temporal network dynamics could play a significant role.

Future work could explore the impact of misinformation, incorporate more complex disease models, account for heterogeneity within subpopulations, and investigate temporal network dynamics.

Summary updated

BibTeX
@article{mei2026modeling,
  author = {Guangyuan Mei and Yao Cai and Su-Su Zhang and Ying Huang and Chuang Liu and Xiu-Xiu Zhan},
  title = {{Modeling Coupled Epidemic-Information Dynamics via Reaction-Diffusion Processes on Multiplex Networks with Media and Mobility Effects}},
  journal = {IEEE Trans. Network Science \& Engineering},
  year = {2026},
  doi = {10.1109/tnse.2025.3632506},
}
Figure