> To first do an FWE correction for voxels within an ROI and then to extract the betas for
> visualization purposes. Does it make sense to do it that way?
Sure, that's legitimate. You can use, for example, 3dROIstats, to extract the effect estimates for that purpose.
Gang
