Risk Cues Can Push Transmission in Opposite Directions
Host animals do not respond to disease risk with perfect information. In this modeling study, the way individuals judged danger at a fine spatial scale determined whether their avoidance behavior reduced infection or, in some cases, helped sustain it. The central finding was stark: behavioral responses to perceived risk can lower prevalence, but the wrong proxy can also make transmission worse.
The analysis focused on a simple but biologically realistic setting in which hosts travel between home and away locations, adjusting movement when they perceive danger. Those responses took two forms. Some individuals reduced the overall frequency of trips, an absolute risk response. Others kept traveling but shifted toward less risky destinations, a relative risk response. Both strategies are intuitively defensive. Only one consistently improved disease outcomes.
Why the Choice of Proxy Matters
The models tested three cues that animals might use to gauge danger: local prevalence, local host density, and the local transmission coefficient, described here as “place.” Each cue captures something about risk, but none is guaranteed to match the true driver of transmission in every setting. That mismatch is the biological fault line running through the analysis. If transmission is density dependent, density may be informative. If transmission depends more on encounter probability per contact, prevalence or place-based cues may tell a different story. When hosts act on a proxy that does not align with the actual transmission process, the defensive move can land in the wrong direction.
This is where spatial scale becomes decisive. Real prevalence estimates are often aggregated over broad areas, while actual transmission unfolds through close-range contact at local sites. A cue that looks sensible in a coarse map can fail at the location where infection is actually passed along. In that setting, behavior based on incomplete spatial information does not simply become less efficient; it can become counterproductive.
Absolute Avoidance Reduced Prevalence More Reliably
Across the modeled scenarios, absolute responses generally performed better than relative ones. Cutting back on travel lowered prevalence more effectively than selectively rerouting trips. The reason is straightforward in epidemiological terms: fewer departures mean fewer opportunities for exposure, fewer risky encounters, and less movement of infection between home and away settings. Reducing travel also suppresses the mixing that allows local outbreaks to seed new locations.
A one-sentence result matters here: some risk-averse behaviors increased transmission instead of suppressing it.
That outcome emerged when the proxy driving movement was poorly matched to the mode of transmission. In those cases, hosts either relied on information that had little bearing on the actual hazard or ignored the information that did matter. A density cue in a frequency-dependent system can mislead movement decisions just as a prevalence cue can misfire when local contact structure, not headcount, governs spread. Behavioral caution, in other words, is only protective when it is wired to the correct ecological signal.
Fine-Scale Movement Can Reshape Spatial Epidemics
The “simple trips” framework captures a common pattern in animal and human mobility: repeated movement between a familiar base and other locations. Even this stripped-down architecture was enough to show how spatial structure changes epidemic behavior. When travel decisions were modified by perceived risk, the spatial distribution of infection shifted as well. Local movement choices altered who mixed with whom, where transmission was concentrated, and how quickly infection circulated between sites.
This matters because mobility is not just a bystander variable in disease systems. It is part of the transmission machinery. Changing trip frequency changes contact opportunities. Changing destination choice changes the composition of exposure. Changing both at once can either choke off spread or channel it into the remaining high-risk routes. The models make clear that spatial disease dynamics depend on how hosts interpret their environment, not only on how pathogens move through it.
Mismatch Between Risk Signal and Transmission Mode
The most consequential insight comes from the conditions under which behavioral responses became detrimental. Those scenarios arose when the risk proxy and the transmission mode were mismatched. If infection spread in a density-dependent way, but hosts adjusted behavior using a cue that did not track density well, the response could fail to reduce exposure. If the system was governed by frequency-dependent transmission, yet movement decisions followed an irrelevant local signal, the supposed avoidance behavior could preserve or even amplify risk.
This is a biologically important warning for any system in which animals learn from partial information. Hosts do not have direct access to the force of infection. They infer danger from clues in the environment, from the crowding around them, or from the risk attached to a place itself. When those clues align with actual transmission, movement can become an efficient self-protective strategy. When they do not, the same behavioral machinery may preserve contact networks that keep the pathogen moving.
Implications for Spatial Ecology and Disease Control
The modeling framework has immediate implications for how risk-averse behavior is interpreted in field studies. Observing movement reduction does not automatically mean transmission will fall. Observing site selection toward apparently safer locations does not guarantee that infection pressure will decline. The biological effect depends on whether the chosen cue is a valid proxy for the underlying transmission process. A public health or wildlife management strategy that assumes all avoidance is beneficial would miss this asymmetry.
There is also a practical message for surveillance. Broad prevalence estimates may be too blunt to capture the local conditions that matter at the contact scale. Fine-scale spatial data on host density, movement, and place-based transmission conditions are likely to be more informative when trying to predict whether behavior will dampen or intensify outbreaks. The same logic applies to conservation and veterinary systems, where animals often respond to disturbance, crowding, or habitat quality long before pathogens are directly observed.
The model does not claim that behavior is inherently harmful or helpful. It shows something sharper. Behavioral defense works when the risk cue matches the mechanism of spread, and it fails when the cue is misaligned with the biology of transmission. Imperfect information is not a minor nuisance in these systems; it can determine whether avoidance lowers prevalence or feeds the epidemic through the remaining routes of contact.