Pete,

Hey! The intensity normalization causes each individual fMRI volume to have the same mean. I don't think FSL generally recommends it.
Feat does not do a global signal regression. A global regression would calculate the timecourse across the whole brain mask, and
regress it away. There is not a Feat option for this, but you could do it after the pre-processing as many people do. fsl_glm would
work for this.
By default, Feat normalizes the 4D mean to 10,000 (it does this just because BOLD units are arbitrary). It comes out to not be exactly 10,000 but that is only because a mask is
applied afterwards, so it usually is slightly below 10,000 for all subjects after preprocessing.

Chris

On Thu, May 17, 2012 at 3:28 PM, Peter Fried <[log in to unmask]> wrote:

Hi FSL Group,

Does the "Intensity Normalization" setting in Feat Pre-stats accomplish the same thing as Global Signal Regression (by incorporating the mean time course of all voxels as a nuisance regressor in the GLM)? My assumption is that it does not, but I am not sure how they are related and which step is preferable. Thanks.

Cheers,
Pete Fried

Boston University School of Medicine