CRCNS US-France Research Proposal: Neural computations of adaptive temporal integration in auditory cortex

People are remarkably skilled at perceiving and understanding complex sounds such as speech and music. Yet, the brain mechanisms that allow people to perceive complex sounds remain poorly understood. A key challenge faced by the brain is that sounds are organized across a vast range of timescales from milliseconds to seconds and minutes. To derive information from sound, the brain must have mechanisms for flexibly integrating information across these diverse timescales. Understanding these mechanisms is critical to understanding how people understand speech, perceive music, and navigate the world using sound. In this project, researchers aim to use computational models to understand how ferret cortex processes multiple types of sound at multiple timescales. The researchers also plan to give lectures, conduct outreach, and mentor undergraduates publicize the results and foster the unique skills needed for this type of cutting-edge research. The proposed project investigates possible mechanisms the brain might use to flexibly integration information. A key question the researchers investigate is whether the brain can change the time window it integrates information over depending on the duration of sound “structures” such as words in speech. They aim to test whether the brain can change its integration window over short and long timescales, and whether this flexibility depends on the types of sounds that one hears and their behavioral relevance. They intend to study these questions using neural recordings from the auditory cortex, a new paradigm for studying temporal integration in the brain, and computational models derived from deep neural networks. This project is funded jointly by the Cognitive Neuroscience Program in the Social, Behavioral, and Economic Sciences Directorate and the Neural Systems Cluster in the Biological Sciences Directorate. A companion project is being funded by the French National Research Agency (ANR). This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.