Hello everyone,

It was so useful to read questions posted by others in the forum. Now I am going to answer my question by myself,

First, we can use the first level beta nii file, as well as other mat files once you know how to import the Z scores for second level analysis in spm fMRI，these values are already fisher z transformed. I prefer to use second level analysis in spm fmri, because I am more familiar with it (please see the reference posted elesever https://www.nitrc.org/forum/forum.php?th...);

Second, we do not need to regress out all the second level covariates, but we need to do qulity ckeck by view the FD value and global signal change. Using ART's intermediate settings, the threshold values are 0.9 and 5 respectively. (see reference https://www.nitrc.org/forum/forum.php?thread_id=11887&forum_id=1144)

Other helpful notes learnd from the forum is that we do not need to regress the global signal out cause Conn uses CompCor, instead of global signal regression, to address potential subject-movement and physiological confounding effects without the risk of artificially introducing anticorrelations into your functional connectivity estimates.

Hope it helps,

Hua
Originally posted by Hua Yu: