Hello,
Using SPM8 I am performing a 1 sample T test of a group of FDG PET scans in patients of different ages (one scan per subject). Age is the only convariate of interest and I am interested in voxels that show an association with age.
In regard to the one sample T test design...
1. I am using Global normalisation > Normalisation > Proportional, to account for global differences between scans. Is this proportional to whole brain MEAN voxel intensity or another parameter?
2. I am using Relative threshold masking (0.8) to mask out extra-cerebral voxels. Is this masking procedure done before or after the global normalisation step?
Thanks in advance for anybody's help.
Cheers,
Kevin London
|