Precise tactile stimulation of worker ants by a robotic manipulator reveals that individual responses are density- and context-dependent

Ant workers are often specialized in specific tasks, and it is well-established that the main task an ant performs in the colony can be used to predict its sensitivity and responses to task-associated stimuli. An often-overlooked aspect of ants' task specialization is that individuals often switch tasks throughout the day and are not always engaged in functional tasks. Furthermore, the tasks individuals engage in are often correlated with other context-specific factors, such as worker density, which can independently influence individuals' behaviour. Given this intra-individual variation in task engagement and its correlation with density, it is currently unknown how these two factors interact to modulate ants' sensitivity and responses to stimuli.To address this question, we built a robotic manipulation system that allowed us to teleoperate a dummy inside ant colonies and to provide simulated antennations to ants when performing different tasks in areas with different worker densities. We coupled this manipulation system with a custom-built automated tracking system (FORT) that allowed us to track individual identities and locations as well as to record the ants' responses to the dummy stimulation.We found independent effects of task and worker density on ants' responsiveness and alarm towards the dummy. Ants were less likely to respond and be alarmed by the dummy when stimulated in areas with high worker density. Responsiveness but not alarm was further influenced by the task being performed, with ants doing broodcare being the least responsive. Our results suggest that ants' behaviour is density-dependent and that ants experience a process of habituation to tactile stimulation. Additionally, ants' responsiveness is modulated by the task they are performing at a given time, showing that sensitivity to stimuli is context-dependent.Our robotic set-up constitutes a valuable tool to systematically investigate social insect behaviour under unprecedented experimental control to unravel the individual-level behavioural rules that underpin the organization of social insect colonies. The integrated system presented here opens new research avenues to empirically investigate the effects of more complex stimuli on social insect behaviour and has the potential to significantly further our understanding of decentralized collective systems.