Hello,
extracting the GM volume values at defined voxels is definitely a good
step. It may then be easiest to do partial correlations with e. g.
correction for age effects and also the other regressors of no interest of
your GLM (as. e. g. total GM) using a statistical program outside matlab.
(This does not mean that there are no matlab functions that do partial
correlations for you... however I do not know them).
Hope this helps,
Philipp
|