Discussing the Ophys QC Pipeline

[Dedalus9]

I am starting a dedicated discussion thread to discuss the quality control pipeline for the ophys data.

My GitHub repository can be found here: https://github.com/Dedalus9/openscope-ophys-qc.git. In it, I include notebooks that show how to load the data, what's included in the data, and some preliminary work for analyzing the impact of motion artifacts on data quality.

[Dedalus9]

Ophys QC Analysis.pptx

These were my takeaways from our meeting on 6-2-26:
Qualitative Comparisons

• Generate figures of a few ROIs that were heavily impacted by motion alongside examples with minimal motion impact for qualitative comparison
• Include example frames showing how somas and dendrites appear in field of view

Metric Descriptions and Concerns

• Every metric in the pipeline should include a plain-language description of what it measures and why it matters
• Include lateral (x/y) drift and axial (z) drift as explicit, reportable metrics
• Soma/Dendrite probability may be problematic metrics: the boundary between a small soma and a large dendrite process may not be clean.
• Investigate relationship between animal behavior and motion-induced ROI degradation.

New Figures to Add

• Histogram of the average number of frames any individual ROI is in frame across a session. This captures how often cells drift in and out of the focal plane.
• Distribution of SNR across all ROIs computed across a set of sessions.

Validation Metrics: Open Questions

• We need to establish what our SNR metric should be. For instance, candidate SNR metrics include representational quality and noise level of pyramidal neurons relative to background.
• We need to define and agree on SNR metrics.