[Freesurfer] obtaining residuals after regressing out nuisance parameters

Dear Freesurfer experts, I am running some functional connectivity analyses using FSFAST in Freesurfer 5.1. I would like to compare seed time courses (which I later use as predictors) obtained after regular pre-processing with seed time courses obtained after regressing out several nuisance parameters. I am a little bit unsure how to use the output from selxavg3-sess for the latter. After defining a single subject analysis with mkanalysis-sess with -notask and only the nuisanance regressors (-nuisreg), I currently run selxavg3-sess with -no-con-ok and -run-wise. Can I then take the time courses from the respective beta.nii per run? Or am I completely off here? Thanks! Caspar