Acoustic noise is a main source of disturbance for animals that use sounds for navigation (echolocation) and social communication. Noise can mask sensory responses to relevant stimuli and thus affect the ability to negotiate the environment. To avoid noise, animals rely on a variety of behaviors such as increasing call level and modifying the pitch and timing of vocalizations. Noise avoidance is a typical example of active sensing in which animals rely on sensory representations of the noise to adjust the physical attributes of their calls. The aim of this project is to investigate the role of the cortico-subcortical network composed of frontal cortex, auditory cortex and inferior colliculus in noise avoidance in bats. Bats are a well-established animal model for research in hearing and vocal behavior, and they thus provide a strong theoretical framework for studying noise avoidance circuits. Our guiding hypothesis is that noise avoidance involves an interplay between ascending and descending neural pathways that connect cortical and subcortical structures forming direct and indirect sensing loops. We propose a tandem project in which an experimentalist (Hechavarria, GU Frankfurt am Main) and a computational group (Triesch, FIAS, Frankfurt am Main) will team-up to record, model, and perturb the frontal cortex-auditory cortex-inferior colliculus network. In our experiments, bats will echolocate in noisy and quiet environments while they are swung in a pendulum. We will: i) record from cortical and subcortical areas simultaneously in behaving bats, ii) perturb neural activity by deactivating brain areas with focus on the frontal cortex, and iii) develop models with feedforward and feedback connectivity to understand how cortico-midbrain loops participate in the active efficient coding of information in noisy environments.