Project Summary / Abstract Recent breakthroughs in neural recording technologies suggest the possibility of understanding the collective dynamics of large-scale brain circuits. However, investigations into these circuits are typically limited to short recording sessions, and overlook the possibility that dynamics can change over longer timescales due to differences in arousal, cognitive state, learning, and low-level biochemical turnover. Identifying which aspects of network behavior are sensitive to these factors, and which are persistent, would produce deeper and more contextualized understandings of many different neural systems. This goal poses severe data analytic challenges. While hundreds of neurons can be recorded over long time periods, we lack established statistical methods that track changes to the high-dimensional structure of network interactions over time. I will work with Dr. Scott Linderman, an expert in neural time series analysis, to overcome this challenge. I will collaborate with two premier experimental labs (Dr. Lisa Giocomo and Dr. Krishna Shenoy) to study circuit-level plasticity across different species (rodents and nonhuman primates) and behavioral tasks (navigation and motor learning). How these circuits reconfigure themselves over multiple hours, days, and weeks is poorly understood. This work will yield early scientific results in this regard, and develop general-purpose statistical tools for the neuroscience community.