Summary To measure the detailed function of neural networks in vivo, 2-photon population calcium imaging is widely used, but has important limitations including low temporal resolution and poor single-spike detection. These limit its ability to measure physiologically relevant activity patterns, particularly in cerebral cortex. A powerful alternative is voltage imaging with genetically encoded voltage indicators (GEVIs). Modern GEVIs have high sensitivity and reduced bleaching, detect spikes with 1-2 ms time resolution, work under 2-photon (2p) conditions, and report both subthreshold and spike signals. In this project, we optimize in vivo 2p imaging methods for two modern GEVIs in somatosensory cortex (S1) of awake, behaving mice. We use conventional 2p resonance-galvo imaging in small fields, and free-space angular chirp-enhanced delay (FACED) 2p imaging in larger fields (>100 neurons), both of which detect spikes at 1-2 ms resolution. Aim 1 optimizes these methods for pyramidal cells and major interneuron types, and quantitatively calibrates optical spike detection. Aim 2 applies these methods to study the dynamics and interactions of sensory and cognitive signals in specific cell types in S1. Layer 2/3 pyramidal neurons mix rapid whisker touch signals (<10 ms resolution) with slower cognitive signals (e.g., for decision or expectation). We will measure rapid sensory-evoked spike and subthreshold dynamics in two functionally distinct, spatially intermixed pyramidal cell classes, in order to test whether these represent distinct networks. We also study a cognitive signal, the response to unexpected deviant (oddball) sensory stimuli, which is thought to be a long-latency, top-down signal. We study how this cognitive signal interacts with rapid sensory signals in pyramidal networks, and how it recruits inhibitory interneurons. This long-latency deviant response corresponds to the mismatch negativity (MMN) signal, which is a widely used EEG biomarker for schizophrenia, and our results may reveal novel circuit mechanisms for this signal. This project brings together the Feldman lab, with expertise in neural coding and circuit function in S1 cortex, and the Ji lab, with expertise in in vivo 2-photon imaging method development and GEVI imaging. This early-stage project is based on an ongoing collaboration in which we have developed reliable methods for 2-photon voltage imaging of whisker-evoked activity in S1 in vivo, using ASAP4.6-Kv, a high-sensitivity GEVI developed by Michael Lin. Overall, this project will establish optimized methods for 2p GEVI imaging in vivo, and use them to probe first-level questions about sensory and cognitive dynamics in S1 networks. In the long run, 2p GEVI imaging promises to reveal network activity at millisecond time scales, revolutionizing our understanding of cortical circuits in health and disease.