Computing Through Interactions: Linking cortical population dynamics to information flow and behavioral representations

Computing Through Interactions: Linking cortical population dynamics to information flow and behavioral representations

Jason MacLean, University of Chicago

Abstract: Neural population activity is variable, but that variability is not necessarily noise. In recurrent cortical networks, shared fluctuations can reflect structured interactions that shape how activity propagates, what information is available to downstream neurons, and how behavioral representations emerge. I will describe work from our laboratory examining this structure across models, sensory cortex, and motor behavior. These studies show how recurrent interactions organize population variability, how covariance structure can carry stimulus-dependent information and influence downstream responses, and how network organization is refined during learning and natural behavior. Together, this work identifies structured variability and interaction dynamics as key features linking cortical population activity to computation.

Biography: Jason MacLean is Professor of Neurobiology at the University of Chicago. His laboratory studies how recurrent interactions among neurons give rise to population dynamics, information flow, and behavior. By combining experiments, theory, and quantitative analysis across sensory and motor cortex, his group asks how structured variability in neural populations reflects computation, how those interactions are reorganized by learning, and how collective dynamics are expressed during natural behavior.