Dear All,
I noticed that using PET model: Population main effect: 2 cond’s, 1
scan/cond (paired t-test) with or without global normalization, gives
different smoothness values in the statistics table. The smoothness FWHM
is much higher using analysis without global normalization.
Smoothness FWHM = 11.3 13.0 14.8 mm = 5.7 6.5 7.4 voxels (with global
norm.)
Smoothness FWHM = 21.8 25.0 28.6 mm = 10.9 12.5 14.3 voxels (w/out global
norm.)
I was using the same PET model and scans (realigned, normalized and
smoothed with Gaussian kernel 12x12x12 mm, voxel size = 2 2 2 mm) in both
analyses.
To match the smoothness for both analyses, I had to decrease Gaussian
kernel to 6x6x6 mm for the analysis without global normalization and got:
smoothness FWHM = 10.5 11.8 14.0 mm = 5.2 5.9 7.0 voxels.
Can someone explain why there is such a discrepancy in the smoothness FWHM
in analysis with or without global normalization?
Thank you in advance,
Beata
|