Hello SPM-list,
we would like to feed maps gained from fMRI-ICA data to a second level
voxelwise test in SPM5, however, face the problem that from some subjects
only one
such maps could be gained whereas from others more than one (<=3) are
available.
As we would like to exploit the power of these multiple epochs, we our
questions are:
(1) Is a 2nd-level t-test of such a stack of images against zero valid at
all? What other method
would be suited to test the hypothesis that these maps in "average" are
significantly
distinct from zero?
(2) In the SPM5 factorial analysis model setup, sphericity can be defined
as non-violated or
violated - would this setting (violated in this case) account for the
described problem,
i. e. partly "dependency" of groups, in an AN(C)OVA model?
Any opinions or recommendations are appreciated,
thanks a lot
Philipp
Max Planck Institute of Psychiatry
NMR Research Group
Kraepelinstr. 2-10
80804 Munich
Mail: [log in to unmask]
Phone: 0049-89-30622-413
|