Dear Mike,
> We are trying to understand fully SPM's threshold maksing
> procedure and how it affects regions prone to susceptibility
> artifact. As we understand it now, SPM initially removes all voxels
> whose T2* intensity (on the smoothed images) is below 1/8 of the
> global mean. The mean of the remaining voxels (the upper 7/8ths) is
> then used as a second threshold such that voxels whose T2* intensity
> fall below the mean are removed from the subsequent statistical
> analyses (and labeled NaN on contrast images). Are we close?
Yes, that's pretty close. The only thing I'd also mention is that if you
use proportional scaling, SPM scales scans by its estimated mean
intracerebral intensity before making a decision which voxels make it to
the analysis.
Also note that it's entirely up to the user to change the SPM default
behaviour. If you want to control what voxels SPM removes from the
analysis or which one it should keep, you can switch off the implicit
masking and use your own mask or no mask at all. have a look at
http://www.jiscmail.ac.uk/cgi-bin/wa.exe?A2=ind0006&L=spm&P=R13693
Stefan
--
Stefan Kiebel
Functional Imaging Laboratory
Wellcome Dept. of Imaging Neuroscience
12 Queen Square
WC1N 3BG London, UK
Tel.: +44-(0)20-7833-7478
FAX : -7813-1420
email: [log in to unmask]
|