matchings
commented
11 months ago This was discussed endlessly and no solution was ever put into place. Also, we only use mesoscopes now for ophys production, so the need for different thresholds for different microscopes is no longer relevant.
Metric / Plot Impacted
The threshold is what should be re-evaluated and adjusted. It is based on assumptions from the beginning of the brain observatory project that may no longer be accurate.
Update Motivation We currently have a blanket threshold of 10um for failure that was developed rather arbitrarily based on data from the Nikon and Scientifica microscopes. However, based on the fact that this threshold was fairly arbitrary and that we have expanded our imaging modalities to include multi-plane imaging on both the deepscope and the mesoscope we should adjust this threshold and potentially set microscope specific thresholds.
The justification for microscope specific thresholds being that each microscope or microscope type has a specific point spread function in z that will impact the ability to resolve a signal in z.
Requested Feature/Solution A working group should be formed to investigate and to determine the appropriate formula/calculation to set the zdrift threshold based on the principals of imaging and microscope setup.
A common formula/calculation should be created that will be applied to each microscope/imaging modality such that the thresholds themselves may vary, but the framework by which they are calculated/set is the same.
Scope This would be assessed on an imaging modality/microscope type level, which likely intersects with specific projects but is not actually based on the project.
Additional context & analysis I propose the following people be involved in the working group: @ledochowitsch, @matchings, @DowntonCrabby @nataliaorlova