Project Summary/Abstract: Project 4, Multi-Region Interactions So far, our U19 program has largely focused on how individual brain regions support working memory and decision-making. While this work has identified a large network of regions that contribute to decision-making, we now need to understand how these regions interact with each other. What information is sent between which regions, and when? How does such interaction depend on task demands? This project’s goal is to address these gaps by combining cutting-edge recording methods, recordings during cell-type- specific perturbations, new computational methods, and a panel of decision-making tasks. We propose to collect two datasets of multi-region, large-scale, cellular-resolution recordings (Aim 1). In the first, we will record simultaneously from a large network of regions, using next-generation Neuropixels NXT technology. In the second, we will image simultaneously from three regions with our mesoscope while obtaining cell-type-specific information through Core 4, Neuroanatomy. For both datasets, we will apply new methods that we have developed to estimate time-dependent information flow across regions. In addition to this broad, large-scale survey, we will perform two more targeted experiments. In the basal ganglia, we will address the puzzle of how perturbation of neural activity that is not choice selective can nevertheless lead to choice biases, by combining cell-type-specific perturbations with downstream recordings (Aim 2). We will also address the broad and recently prominent question of the cerebellum’s role in cognition (Van Overwalle 2020; Wagner 2020; Stoodley 2021). By manipulating cerebellar populations while recording in thalamus and cortex, we will test the hypothesis that, by analogy to the cerebellum’s role in refining and stabilizing motor trajectories, this region also serves to refine and stabilize neural trajectories in frontal association areas (Aim 3). Taken together, we expect that these experimental and computational approaches will provide the foundation for a multi-region description of how decisions are formed, which will provide important constraints for our multi-region mechanistic model in Project 5. This work will substantially advance three priority areas of the BRAIN Initiative: the brain in action, demonstrating causality, and identifying fundamental principles.