Phys. Rev. E 97, 042306 (2018)

Every winter we hear forecasts on how bad the upcoming flu season is going to be. Seasonal peaks are but one example of cyclic epidemics, a phenomenon that is shared by several diseases — some of which have periods of months or even years. Seasonality can’t explain these oscillations — but perhaps our reaction can.

From avoiding public transport to quarantines, people respond to outbreaks by changing habits. Mathematically, these changes can be captured by adaptive networks, where nodes redirect their links from infected neighbours towards healthy ones at a certain rate. However, epidemic models on these networks haven’t displayed the expected periodicity so far.

Now, Neil Sherborne and colleagues demonstrate that an adaptive model with a delay between cutting and reconnecting links does the trick, leading to oscillations between states with high and low disease prevalence. Besides making sense behaviourally, this model provides a starting point for studying infections and adaptive behaviour.