Figure 1

Schematic diagram portraying how social networks with different association preferences are exposed to disease outbreak: Dynamics of a constantly shifting social network with different association preferences subjected to disease propagation. (A) Three initially identical networks (but with different association preferences) are created; subsequent association shifts causes them to diverge into networks with different structures over time. (B) At $t=t^{*}=50$, a source of infection $v_i$ is chosen at random; $v_i$ changes to state $\mathit{I}$ and the spreading process begins. The matrix $M^{50}$ is generated. Individuals $v_j$ with arcs connecting them to $v_i$ in any of the networks become exposed, or not, in all of the networks with the appropriate arcs depending on $M_{i,j}^{50}$ and/or $M_{j,i}^{50}$. Networks continue to shift according to individual preference. (C) At $t=51$, the matrix $M^{51}$ is generated. Individuals in state $\mathit{E}$ changes to state $\mathit{I}$ and the spreading process continues. All of the networks continue to shift. (D) Disease propagation and association shifts continue; infection numbers recorded at each time $t$. (E) Cumulative numbers of infection are recorded for each network.Reuse & Permissions

Figure 2

Network characteristics, degree distribution and network centrality measures of different network types after convergence: A representative sample of stable network characteristics after convergence in the three network types (betweenness, closeness, and degree in panels A, B, and C, respectively). After converging to stable structures, the three networks showed varying levels of agreement to each other within each of the measures: degree distributions (top of each panel), network centrality measures (middle of each panel) and overall network contact structures (bottom of each panel). The size of the nodes within each network (bottom of each panel) represents their relative individual centrality according to the metric of association for their population. Note that network centrality can only be compared within measure across networks (i.e., betweenness to betweenness measure of two different networks, betweenness centrality for one network cannot be compared to any of the closeness or degree centrality measures).Reuse & Permissions

Figure 3

Network degree centrality over time in increasingly large networks. As network size increased, so did the time required for the network to reach stability. Disease was introduced into the system (and, in the static case, the network was frozen) only after the network had achieved at least 90% of the centrality measure at stability; indicated for each curve by the barbell.Reuse & Permissions