Infection-induced behavioural changes reduce connectivity and the potential for disease spread in wild mice contact networks

Abstract

Infection may modify the behaviour of the host and of its conspecifics in a group, potentially altering social connectivity. Because many infectious diseases are transmitted through social contact, social connectivity changes can impact transmission dynamics. Previous approaches to understanding disease transmission dynamics in wild populations were limited in their ability to disentangle different factors that determine the outcome of disease outbreaks. Here we ask how social connectivity is affected by infection and how this relationship impacts disease transmission dynamics. We experimentally manipulated disease status of wild house mice using an immune challenge and monitored social interactions within this free-living population before and after manipulation using automated tracking. The immune-challenged animals showed reduced connectivity to their social groups, which happened as a function of their own behaviour, rather than through conspecific avoidance. We incorporated these disease-induced changes of social connectivity among individuals into models of disease outbreaks over the empirically-derived networks. The models revealed that changes in host behaviour frequently resulted in the disease being contained to very few animals, as opposed to becoming widespread. Our results highlight the importance of considering the role that behavioural alterations during infection can have on social dynamics when evaluating the potential for disease outbreaks.

Disease implications of animal social network structure: A synthesis across social systems

“…Second, we assume that social contacts do not covary with pathogen characteristics and remain unaltered after an infection is introduced into a population. Infection has in fact been shown to alter the social connectivity of hosts (Croft et al., ; Lopes, Block, & König, ) and recent theoretical work has demonstrated that negative correlations between transmissibility and contact rate can diminish the impact of connectivity (White, Forester, & Craft, 2017). Future species‐specific studies can take advantage of host‐specific experimental manipulations, where possible, to gain in‐depth insight towards host behaviour—infection feedback (Ezenwa, Ghai, McKay, & Williams, ; Silk, Croft, Delahay, Hodgson, Boots, et al., ).…”

Covariation between the physiological and behavioral components of pathogen transmission: host heterogeneity determines epidemic outcomes

“…These findings were consistent across the two different formulations of final transmission probability that were simulated. While intuitive, these results are important because reduction of activity and contact rate because of infection are welldocumented (Croft et al 2011, Welicky and Sikkel 2015, Lopes et al 2016), but less commonly incorporated into disease models.…”

“…In natural systems, it has been demonstrated that these interrelated facets of transmission can vary widely between individuals. In fact, empirical studies suggest that unequal contact rates are the rule rather than the exception (Craft and Caillaud 2011), that contact rates can vary with infection-induced behavioral changes (Croft et al 2011), and that these changes are likely non-uniform across individuals (Lopes et al 2016). Innate and plastic heterogeneity in susceptibility to infection has been documented for several species (Dwyer et al 1997, Beldomenico and Begon 2010, Gibson et al 2016, and variability in infectiousness has also been observed, particularly when concomitant infections are present (Cattadori et al 2007).…”

The application of statistical network models in disease research

“…). Furthermore, these changes in behaviour have been shown to alter contact network structure, with implications for transmission (Tunc & Shaw ; Lopes, Block & König ). Therefore, accounting for the co‐dynamics of network structure and infection is key to improving our understanding of disease spread and control in many systems (fig.…”

“…; Tunc & Shaw ); however, there has been relatively little use of empirical data to explore this topic (but see Rohani, Zhong & King ; Reynolds et al . ; Lopes, Block & König ). Using empirical data to test hypotheses about the relationship between sociality and disease (e.g.…”