Dear Aisling,

Unfortunately it seems these QA covariates are only generated when ART is ran via CONN, it is not possible to get them if you use ART beforehand (as I do). However these covariates are only useful for QA, they are not impacting the analysis.

If you really want to get them, an alternative is to input as dataset 0 the original fmri files of your subjects, and re-run ART (and only ART) in the CONN preprocessing, this will regenerate the 1st-level motion covariates and correctly generate the 2nd-level QA covariates.

Also an alternative you might consider for future experiments is to run the preprocessing directly from CONN, as you did initially, but per subject (in the CONN preprocessing window, at the bottom there is a drop-down box to select whether the preprocessing applies to all subject or just one). In this case, it is possible to input a different number of slices for each subject, and I expect you should also get the 2nd-level QA covariates this way.

Hope this helps,
Stephen