CRCNS Research Proposal: Linking correlated variability to flexible behavior

The brain depends on the coordinated activity of groups of neurons to perceive the world, make decisions, and guide behavior. Although individual neural responses are highly variable, this variability is often shared across groups of neurons, which can strongly influence how information is represented and used. This project tests the idea that the brain actively reorganizes shared neuronal variability to support different behaviors. If successful, this research would change our understanding of neuronal variability from a source of error to a fundamental mechanism that enables flexible brain function. These discoveries have the potential to advance basic knowledge of how the brain processes information, learns new tasks, and selects actions, while also providing new principles for designing artificial neural networks that better capture the computational strategies used by biological brains. To test this hypothesis, the researchers combine large-scale recordings of neuronal populations, causal perturbations of visual cortex, and biologically grounded computational models. The project tests how patterns of shared neuronal variability change as animals perform different tasks, and whether those changes improve the transmission of task-relevant information and ultimately behavior. The researchers plan to identify the pattern of population activity that best captures shared neuronal variability, referred to as the correlated variability axis, and determine how it relates to sensory signals, task demands, and action planning. The project tests three predictions: (1) perturbations that shift neuronal activity along this axis produce the largest effects on behavior, (2) task-relevant signals become preferentially aligned with this axis during task performance, and (3) shared neuronal variability is more closely linked to planned actions than to perceptual judgments. 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.