Parasites have a large impact on a host’s life-history evolution, individual behavior, and population dynamics.  Recent models have suggested migratory recovery as a strategy to reduce disease prevalence, and experimental work has shown that diseased fish prefer water temperatures that facilitate recovery and lower transmission rates.  However, current models ignore the behavioral coupling between social spread of disease, and social influence to movement.  We form a model which couples disease transmission with a social influence to movement decisions, and perform simulations to ask how environmental factors and transmission rates influence transient group movement and disease prevalence.  The model represents movement between two habitats.  The 'breeding' habitat has higher disease transmission rates, but is preferred by healthy individuals because it allows for breeding.  The 'recovery' habitat is preferred by diseased individuals because it has low transmission and high recovery rates, but does not allow breeding.  An individual's preferred location depends on its disease state and on the motion of others around them via a social network. The model demonstrates that social influence to movement can dramatically change the prevalence of disease as well as the movement dynamics of a group.  Current work seeks to test model predictions using experiments with stickleback fish and common species of ectoparasites.

Introduction

Living in groups has both pros and cons.  It is normally advantageous in terms of enhanced abilities for predator detection and for finding food.  However, an important negative consequence is that the continuous close proximity to group mates can facilitate the spread of disease.  The mechanism of "migratory escape" may have evolved to mitigate spread of disease, if individuals move from a high-risk to a low-risk environment in terms of infection spread \cite{Hall_2014,Johns_2015}.  For instance, seasonal movement between high infection risk, low altitude winter grounds and low infection risk, high altitude summer grounds allows red deer, Cervus elaphus, to reduce their exposure to Ixodes ricinus ticks \cite{Qviller_2013,Mysterud_2015}. Further theoretical and empirical work has suggested that "migratory recovery", where movement between environmentally distinct habitats leads to recovery from infection, could be an additional mechanism to reduce overall disease prevalence in a population \cite{Shaw_2016,Daversa_2018}, and that the spatial distribution of hosts during transient phases influences disease spread \cite{Daversa_2017}.  Other work has shown that individual location preferences change with disease state \cite{Mohammed_2016}. For example, infected individuals may move to warmer habitats that cause an increase in their surface or core body temperature to the detriment of the parasite, a phenomenon known as “behavioral fever" \cite{COVERT_1977,Kluger_1975,Rakus_2017,REYNOLDS_1976,Satinoff_1976,Moyer_1995}. This strategy is particularly effective in ectotherms, which rely on their external environment to regulate internal temperature \cite{Rakus_2017}. Such preferences could facilitate movement decisions that aid in the recovery of infected individuals, and lower overall disease prevalence.
The mechanisms of migratory escape, migratory recovery, and behavioral fever involve a sort of "spatial escape", where individuals move to an area that facilitates recovery.  However, even though individuals may have their own location preferences, migratory movement decisions are often made in a group setting \cite{Berdahl_2018}. Group movement decisions emerge as a result of both social and non-social information use \cite{P_rez_Escudero_2011}.  Models have suggested that only a small number of leaders are needed to change the direction of the whole group \cite{Guttal_2010,Torney_2010,Pais_2014}.  When preferences are conflicting, the group may compromise or split, depending on the number of individuals and the degree of disagreement \cite{Couzin_2005,Strandburg_Peshkin_2015}
There are there two ways that movement can affect disease prevalence.  The first is moving to an area with higher recovery rates, or escaping an area with high disease transfer rates.  The second is escaping from other diseased individuals in the population, who harbor the disease and may transfer it. Group fission/fusion may a key mechanism to balance the trade-off between the benefits of staying together as a group, and the costs due to increased disease transfer.
It is not known how social influence to movement affects the coupling of disease spread and migratory movement.   For example, in what conditions will an increase in epidemic spread be caused by an over-reliance on social movement cues? Does the use of social and non-social information depend on disease state?  How do individuals balance the trade-off between the benefits of staying in a group, and the costs associated with increased disease transfer? 
To address these questions we form a model that couples disease transmission to movement decisions.  The model represents movement between two habitats:  healthy individuals prefer the  'breeding' habitat, while diseased individuals prefer the 'recovery' habitat.   An individual's movement depends both on its disease state, as well on the motion of others around them via a social network.  We model disease transfer by using a network version of SIS model, and defining transfer rates depending on distance between individuals.  We use the model to ask how social influence to movement affects disease prevalence, depending on the disease transfer and recovery rates in the environment.
We then investigate how a group may mitigate social spread of disease, by considering that healthy individuals can detect diseased individuals, and that adjust their social movement based on the observed disease state of others.  We ask what conditions [motion rules?] predict an increase in disease prevalence, versus what is necessary to lead to group splitting and fission dynamics, to negative effects of how social cues to movement affect disease state [this sentence needs rewording].  We conclude by discussing these results in the context of (sticklebacks-endo parasites), a model system for host-parasite interactions, and outline experimental tests of the model's predictions for social movement and disease transfer.
[to add:  actual model predictions, when do it].
Must add references

Results

We form a model that couples disease transmission with both non-social and social influences to movement.  In the model, individuals move between a breeding area and a recovery area (Figure \ref{505956}).