How to perform nuisance regression in fsl?


I am trying to do nuisance regression of white matter, csf and linear trend for rs-fmri data using fsl after ICA-AROMA. After reading some posts in fslarchive I think the first step would be extracting white matter / csf signals from segmented images.

I used SPM to segment images into c1, c2, c3 etc. Do I need to extract these signals from the subject space or standard space? And how can I do that specifically? A step-to-step guide of the subsequent steps would also be much appreciated.

Thank you very much!