Hi Afni experts,
I am looking to run ICA analysis on data previously processed with afni. I know I will have to make some modifications regarding how to analyze and needed guidance on one particular aspect regarding motion scrubbing.

Previously I would run through afni_proc.py and get a file of combined TRs to delete from extreme outliers and motion that I would incorporate at the end using 3dDeconvolve with my seed time series. With ICA I would not be using a seed obvioulsy but is it still valid to delete these timepoints for this type of analysis? Furthermore, would I be able to take my processed file(despike, ricor, slice time,motion correction, B0, anaticor, smooth - no bandpassing for ICA) and run through 3dDeconvolve without any stims in order to delete these outliers with the following command (if this is not correct please let me know what the command would be) or is there a better way to do this scrubbing (ie incorporate somehow in @Anaticor or another tool to censored timepoint info without running GLM)?

3dDeconvolve -input $rest_file \
-polort $pol \
-censor $censor_file \
-tout -rout -fitts fit_${sub} \
-bucket $sub.anaticor.smooth.censor

The bucket file is usually what I would use for seed based as it is the output comparison of a seed and the voxel timeseries, however, I am not sure if that is the appropriate output in this case to feed into melodic or another tool (if afni has for ICA).

Thanks,
Ajay