Epidemic spreading in temporal multilayer networks coupling with individual behavioral changes

With the recognition of travel infection, understanding the spatiotemporal spread of epidemics driven by human mobility has received considerable attention in multilayer networks. However, current studies neglect the time-varying feature of social interactions and the impact of individual behavioral changes. To address this problem, first, a multilayer activity-driven network with attractiveness is proposed to capture spatiotemporal epidemic spreading at the same timescale with the evolution of social interactions. Second, three key individual behavioral changes are introduced into the model, including travel restriction, self-isolation for infected individuals, and self-protection for healthy individuals, by associating with mobility rate, activity, and attractiveness of infected individuals, respectively. Third, a non-Markovian Susceptible-Infected-Recovered spreading dynamic is established by utilizing the quenched mean-field theory and the analytical expression of the epidemic threshold is derived. Finally, extensive experiments are conducted to analyze how travel infection and individual behavioral changes affect spatiotemporal epidemic spreading. We find that the critical factor driving the role of travel infection is the intensive travel interactions, which are significantly influenced by travel strength (i.e., contact coefficient, the proportion of travelers, and mobility rate) and the social attributes of travelers (i.e., activity and attractiveness). The travel duration also shows a complexly dynamic effect on disease transmission due to travel infection and recovery. Among the interventions, self-isolation and self-protection are the most effective for epidemic control and exhibit symmetrical effects when applied equally. Our work provides a theoretical fundamental to control spatiotemporal epidemic spreading.