# Partitioning Social and Spatial Drivers of Infection Risk

## TL;DR

This study shows how social and spatial behaviors influence infection risk in a rodent species, highlighting the importance of considering these factors in disease transmission models.

## Contribution

The paper introduces a novel approach combining capture-mark-recapture data and proximity loggers to study infection risk in wild rodents.

## Key findings

- Exposure to Morogoro virus is strongly linked to spatial overlap with infected individuals.
- Direct contact occurs with only a few individuals despite broad spatial overlap.
- Spatial structuring significantly influences infection risk in the study population.

## Abstract

The rate at which individuals contact each other is central to the transmission of diseases through populations. Most simple models assume homogenous mixing, with all individuals equally likely to contact each other within a population, where contact can either scale with density (the more individuals the more likely contact will occur) or scale independently of density (where individuals maintain contacts regardless of density). However, there is growing evidence that contact rates are non‐homogenous, and both spatial and social structuring are likely to play an important role in producing and maintaining heterogeneous contact behaviour. Additionally, assuming homogenous mixing is a deliberate simplification, yet it can undermine a model's predictive power when spatial or social structure is important, as is often the case for many wildlife diseases. Here, we investigated the relationship between measures of social and spatial behaviour in a non‐territorial rodent species (
Mastomys natalensis
) which exhibits significant seasonal fluctuations in density and exposure to a mammarenavirus, Morogoro virus (MORV), using an extensive and unique capture‐mark‐recapture dataset. We followed this up with a study to investigate the extent to which spatial overlap may correlate meaningfully with contact rates using miniaturised proximity loggers developed in house. Exposure to MORV was strongly associated with the proportion of home range overlap with other exposed individuals, and negatively associated with the proportion of home range overlap with conspecifics regardless of exposure status. Including spatial autocorrelation suggested that consistent spatial structuring across the study area also played an important role in determining exposure to MORV. Finally, our proximity logger experiment demonstrated that home range overlap may overestimate direct contact behaviour in
M. natalensis
as individuals showed a high degree of spatial overlap, but most contacts were associated with a small number of individuals.

Contact rates drive disease spread, yet are often oversimplified as uniform across individuals. Using data on
Mastomys natalensis
, we show that exposure to Morogoro virus is linked to spatial overlap with infected conspecifics, but not overall overlap, and that direct contact occurs with only a few individuals despite broad home range overlap—highlighting the importance of spatial and social structure in predicting transmission.