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

Area of Science:

• Neuroscience
• Optical Imaging
• Computational Biology

Background:

• Large-scale neural circuit analysis relies on multiphoton calcium imaging.
• System design involves trade-offs between spatial coverage, temporal resolution, and data quality.
• The impact of spatiotemporal downsampling on neural data analysis remains unclear.

Purpose of the Study:

• To systematically evaluate cell detection and spike inference performance under varying spatiotemporal sampling conditions.
• To establish quantitative guidelines for optimizing high-throughput microscopy for neural circuit analysis.
• To understand the influence of sampling parameters on the accuracy of calcium imaging data.

Main Methods:

• Analysis of mouse cortical data using multiphoton calcium imaging.
• Systematic variation of spatial sampling pitch (µm) and temporal sampling (frame rate in Hz).
• Evaluation of cell detection and spike inference accuracy across different sampling parameters.

Main Results:

• Spatial sampling pitch ≤3 µm is critical for accurate cell detection.
• Reliable cell detection is maintained at temporal sampling rates of ≥10 Hz.
• Smaller neurons are more sensitive to temporal sampling rates than larger neurons.
• These findings are consistent across cortical depths and regions.

Conclusions:

• Spatial resolution is the primary factor for accurate cell detection in calcium imaging.
• Temporal sampling rates of 10 Hz are sufficient for reliable cell detection, allowing for faster acquisition or larger populations.
• Quantitative guidelines are provided for optimizing microscope design for large-scale neural circuit analysis.