ReHo on residuals questions (DPARSF)



I'm trying to calculate ReHo on residuals images obtained after GLM analysis ("simulate" resting-state). 

Basically I just want to regress signal from WM and CSF, Bandpass filter, detrend and finally calculate ReHo. I have 2 questions

1) Nuisance regressors (WM, CSF) are supposedly calculated from SPM a priori maps. However, DPARSF is apprently looking for masks created based on functional images. How do I create such template without going through all the classic preprocessing steps? (I'm dealing with residuals images so I dont want to do slice timing, realign, etc...)

2) It is better to calculate ReHo on un-smoothed data. However, residuals are computed eith images already smoothed (GLM analysis) . Is it still ok to calculate ReHo on those images?


Thanks a lot.

1. WM and CSF masks come from aprioris or segmentation. They have to be resliced to the resolution of funcitonal images. Give any of fucntional images in standard space to reslice them are OK.

2. If you performed smooth beforehand, the ReHo values would be enhanced.