search

PennLINC / xcp_d

Post-processing of fMRIPrep, NiBabies, and HCP outputs
https://xcp-d.readthedocs.io
BSD 3-Clause "New" or "Revised" License
78 stars 26 forks source link

Adding "missingness metric" to the quality report #774

Open andreifoldes opened 1 year ago

andreifoldes commented 1 year ago

Summary

I think it would be useful to add a measure reporting how many ROI have "all zero" timeseries - I guess for each parcellation.

Additional details

It important to know whether there are going to be NaNs in the correlation matrix because a lot of downstream analysis software is anxious about these things.

tsalo commented 1 year ago

This is a good idea. We've been discussing something similar in light of #757.

tsalo commented 1 year ago

@andreifoldes How would you want to see the coverage/missingness represented?

andreifoldes commented 1 year ago

I guess it would make sense for it to be either in the Summary section or (above/under) the Correlation Heatmaps from Four Atlases section?

If it was in the summary the neatest would be to have the count of NaNs for each xcp_d parcellation in the form of a dropdown or a table... or maybe better only present it if there are NaNs and as more like a warning?