Optimization of spatial and temporal sampling resolution for optophysiology in large-scale two-photon calcium imaging

Large-scale neural circuit analysis requires multiphoton calcium imaging systems capable of monitoring thousands of neurons, creating fundamental trade-offs between spatial coverage, temporal resolution, and data quality. While reduced sampling can dramatically increase acquisition speed or the total observable neural population, the effects of spatiotemporal downsampling on cell detection and spike extraction performance have not been systematically characterized. This study systematically evaluates cell detection and spike inference performance across varying spatial and temporal sampling (frame rate in Hz) on mouse cortical data. We demonstrate that spatial sampling pitch ≤3 µm is the critical determinant of detection accuracy while temporal sampling shows greater tolerance, with reliable cell detection maintained at ≥10 Hz, and these relationships are consistent across cortical depths and regions. Additionally, smaller cells are more sensitive to temporal sampling on detection accuracy than larger cells. These results establish quantitative guidelines for optimizing high-throughput microscopes for large-scale calcium imaging experiments.