I am interested in looking at brainstem, basal ganglia and prefrontal activation.

I am conducting indirect normalisation. I first coregistered the subject's EPI mean image to their T1 anatomical. I then conducted segmentation, yielding c1,c2,c3,c4,c5 output files. 

I am wondering if I should include c1 (GM) or c2 (WM) segmented files in the normalisation phase? When I view c2, there appears to be more brainstem, specifically periaqueductal gray (which is our top interest).

I have already conducted the experiment using c2 output. We get nice results for periaqueductal gray, putamen and PFC. Should I re-do the experiment using c1?