Task-dependent vocal adjustments to optimize biosonar-based information acquisition

Abstract Animals need to acquire adequate and sufficient information to guide movements, yet information acquisition and processing is costly. Animals thus face a trade-off between gathering too little and too much information and, accordingly, actively adapt sensory input through motor control. Echolocating animals provide the unique opportunity to study the dynamics of adaptive sensing in naturally behaving animals, since every change in the outgoing echolocation signal directly affects information acquisition and the perception of the dynamic acoustic scene. Here we investigated the flexibility with which bats dynamically adapt information acquisition depending on a task. We recorded the echolocation signals of wild-caught Western barbastelle bats ( Barbastella barbastellus ) while flying through an opening, drinking on the wing, landing on a wall, and capturing prey. We show that the echolocation signal sequences during target approach differed in a task-dependent manner; bats started target approach earlier and increased information update rate more when the task became increasingly difficult, and bats also adjusted dynamics of call duration shortening and peak frequency shifts accordingly. These task-specific differences existed from the onset of object approach, implying that bats plan their sensory-motor program for object approach exclusively based on information received from search call echoes. We provide insights into how echolocating animals deal with the constraints they face when sequentially sampling the world through sound by adjusting acoustic information flow from slow to extremely fast in a highly dynamic manner. Our results further highlight the paramount importance of high behavioural flexibility for acquiring information. Summary statement Having the right information for a specific job is crucial. Echolocating bats flexibly and independently adjust different call parameters to match the sensory-motor challenges of four different tasks.