Servicenavigation

Project Details

Back

Projekt Print View

Degrees of relevance? The neuronal representation of sound sources in primary auditory cortex during active localization

Subject Area Cognitive, Systems and Behavioural Neurobiology
Term from 2019 to 2024
Project identifier Deutsche Forschungsgemeinschaft (DFG) - Project number 418090239
 
Final Report Year 2024

Final Report Abstract

How do we hear out individual talkers from a crowd or a car in heavy traffic? In everyday life, multiple sound sources are often simultaneously active, yet our brain unravels this mix of sounds into distinct information streams according to the sources of origin. Yet a functional understanding of the neuronal mechanisms of this phenomenon, which is termed auditory scene analysis (ASA), remains elusive. Most studies on ASA to date focused on the instant processing of simple sounds in passive listening settings. In realistic ASA, however, identification of a particular sound source critically depends on its goal-specific relevance. Moreover, a fundamental element of sensing & perception is voluntary self-motion to gather information about the resulting changes in sensory inputs. This interdependence of self-motion and the modulation of sensory information is fundamental to neural processing and perception. To reflect these critical components in a lab environment, in this project we developed a behavioural paradigm named “SIT” (Sensory Island Task). In SIT, animals explore an open-field arena to find a sensory target relying solely on changes in the presented stimulus, which is controlled by closed-loop position tracking in real-time. Within a few sessions, animals learn to search for a particular area (“island”), which triggers the presentation of the target stimulus (the island location is randomized across trials). SIT allows testing psychometric discrimination and identification during self-motion across sensory modalities and species. We also established a version of SIT to study human sensory processing. As a main goal of this project, we combined SIT with concurrent chronic multi-electrode brain recordings to study the neuronal representations. To this end, we applied SIT to study sound localization and source identification in freely behaving and engaged animals. An arena was equipped with two diametrically opposed loudspeakers and the target cue was a change in the allocentric location of the active loudspeaker. Expert animals were implanted in primary auditory cortex (A1) with custom-made tetrodes. Chronic recordings of action potentials from multiple neurons in A1 during active localization and self-motion revealed previously unreported spatial representations. Specifically, the task-specific identity of the sound-source (rewarded or un-rewarded) altered the spatial preference of A1 neurons with high temporal dynamics. Neural network decoding demonstrated that the diverse activity patterns of the recorded neuronal ensembles provide spatio-temporally co-existent information about both the egocentric location and the identity of sensory objects. Our findings suggest that the task-specific relevance of individual sources (such as talkers on a party) generate specific population patterns of A1 activity that hold context-dependent information about multiple aspects of an auditory scene. This insight is crucial for our understanding of active sensing, such as tracking a sound source while moving. Our findings also raise fundamental questions about the underlying circuits and mechanisms, as well as what additional scene aspects might be encoded in A1 during active listening. Multiple follow-up projects, partially funded by the DFG, are ongoing to answer these questions with the goal of refining our understanding of sensory object identification during ASA.

Publications

 
 

Additional Information

© 2026 DFG Disclaimer / Copyright / Privacy Policy

Textvergrößerung und Kontrastanpassung