Dynamic pathogen detection and social feedback shape collective hygiene in ants

Collective action often surpasses individual efforts in efficiency. This is particularly evident in social insects, whose colony-level success depends on the coordinated actions of individual members. Cooperative disease defense, or social immunity, is a crucial aspect of this collective behavior. Social insects exhibit various hygienic behaviors, including grooming, to remove infectious particles from infected individuals. While the behaviors themselves are documented, the underlying decision-making rules and their contribution to colony-level protection remain largely unclear. This study aims to understand the individual decision-making rules that shape collective hygiene in ants, focusing on how individual ants make grooming choices and how these individual decisions contribute to efficient colony-level pathogen removal. Understanding these rules is crucial for comprehending the emergence of complex collective behaviors and their role in disease defense within social insect colonies. The study uses the garden ant *Lasius neglectus* and the fungal pathogen *Metarhizium brunneum* as a model system to directly test the hypotheses on ant grooming behavior.

Previous research has highlighted the importance of collective action in various biological systems, from interacting genes and neurons to cells in tissues and animal collectives. Social insects, like ants, bees, and termites, exemplify the emergence of self-organized collective behavior, exhibiting remarkable coordination in tasks such as food collection, nest construction, and defense. Studies have described the hygienic repertoire of social insects, including grooming, but a quantitative understanding of the decision rules governing these behaviors and their contribution to colony-wide disease protection remained largely unexplored. While some studies have investigated the impact of social interactions on disease transmission, the individual-level decision-making processes underlying collective hygiene behaviors require further investigation. This study builds upon this existing knowledge by focusing on quantifying the decision rules guiding individual ant grooming choices and modelling their contribution to collective pathogen removal at the colony level.

To investigate individual decision-making in ant hygiene, the researchers designed experiments where ants could choose between grooming two individuals with different pathogen loads. Groups of six *Lasius neglectus* ants were used, with two ants treated with high (F) or low (f) doses of *Metarhizium brunneum* spores, or a non-pathogenic control (C). The behavior of all ants was recorded at high temporal resolution (1/5 s). The spore load of each ant was quantified before and after a 90-minute observation period using droplet digital PCR (ddPCR), targeting genetically encoded fluorescent labels (GFP and RFP) on the spores. This allowed for tracking spore removal and transmission. The researchers analyzed self-grooming, allogrooming (grooming others), and the use of formic acid (an antimicrobial agent). Probabilistic modeling was employed to infer the decision rules guiding individual ant grooming choices. The models considered various factors, including the perceived pathogen load on other ants and the recent grooming history of the individual ant. Model selection was performed using cross-validation to identify the best-fitting model. A separate experiment examined the effect of choice restriction on pathogen removal efficiency by comparing groups where ants could freely choose whom to groom versus groups where grooming targets were predetermined. The statistical analysis of data was performed using R and MATLAB.

The study's key findings demonstrate that ant grooming decisions are dynamically adjusted based on perceived pathogen threat and social feedback. Ants preferentially groomed individuals with higher spore loads, but suppressed their own grooming activity after receiving grooming themselves. This negative feedback loop likely prevents the most infected individuals from spreading pathogens while focusing hygienic efforts on those most in need. The probabilistic models identified that ants make their grooming decisions based on a simple rule: the perceived spore load encountered on others recently and the grooming received recently. This rule doesn't require global knowledge of the colony's spore load, suggesting efficiency in large colonies. The results show that the ants' ability to make informed grooming choices is crucial for efficient colony-level pathogen removal. Preventing ants from freely choosing whom to groom significantly reduced overall pathogen removal, underscoring the importance of individual decision-making in collective hygiene. The analysis revealed a clear bias towards grooming the ant with the currently highest spore load, even if this wasn't the ant with the initially highest load, illustrating the dynamic and responsive nature of ant grooming behavior. The model accurately predicted ant grooming behaviors across the entire experiment. Furthermore, the study showed that the efficiency of pathogen removal increases with the difference in spore load between treated ants.

The findings address the research question by identifying simple, yet effective, individual decision rules underlying collective hygiene in ants. The results highlight the importance of dynamic information processing and social feedback in shaping individual ant behavior and ultimately colony-level disease resistance. The simple, local rules identified do not rely on global colony information, suggesting scalability and robustness. The study's significance lies in its mechanistic explanation of how collective hygiene emerges from individual-level decisions in a complex social system. The identified rules, involving both excitatory (perceived pathogen load) and inhibitory (social feedback) factors, provide a robust and efficient strategy for disease control. The findings contribute to a broader understanding of collective behavior and social immunity in social insects and provide valuable insights for future research on other complex biological systems.

This study reveals that simple individual-level decision rules, based on dynamically updated information about pathogen load and social feedback, are sufficient to generate efficient collective hygiene in ant colonies. The preferential grooming of highly infected individuals, coupled with self-grooming suppression after being groomed, contributes to effective colony-wide pathogen removal. Future research could explore the neurological mechanisms underpinning these decision rules and investigate the generality of these findings across different ant species and social insect groups. Further studies could also examine the role of other environmental factors and potential trade-offs associated with the observed strategies.

The study focused on a specific ant species (*Lasius neglectus*) and a particular fungal pathogen (*Metarhizium brunneum*). The generalizability of the findings to other ant species and pathogens needs to be explored in future research. The experimental setup, while providing detailed behavioral data, might not fully replicate the complexity of natural ant colonies. The model assumes perfect detection of spore loads, which could be an oversimplification of the ants’ sensory capabilities. Finally, the study primarily focuses on grooming behavior and might not fully capture the complete spectrum of ant hygienic strategies.