Hi Harris,

I used *confounds.tsv file from fMRIPrep for this, which provides 'FramewiseDisplacement' and 'stdDVARS' values for every data point. I created empty vector of zeros and then I identified time points with FD > 0.5 mm and stdDVARS > +/- 3 SD and set them to ones. Next, I loaded this vector into CONN as confound regressor.

I think that running ART in CONN would also work, but as I know ART looks for for SPM's 6 motion parameters *.txt file so you should also create such file by subsetting *confounds.tsv table and saving it without column names in a .txt file.

Hope this help!

Best,
Karolina