Dear SPM community,
I'm a relatively new user and have a 2 part question regarding the
significance level in an SPM analysis. We ran a group comparison in
which group 1 consisted of 10 controls and group 2 consisted of 19
patients, with AnCova global normalization. As I understand (or maybe
misunderstand) the literature, the SPM{Z} map contains the conjoint
test analysis on voxels which survived the uncorrected F test indicated
on the SPM{Z} map.
Part 1 of my question:
Neither of the 2 clusters on the SPM{Z} is significant at the set,
cluster, or voxel level. (In fact they have p>0.4 for all categories).
Why are such clusters displayed?
Part 2 of my question:
As it happens, the 2 clusters that did show up on the SPM{Z} are in very
close agreement with those reported by other workers who independently
tested a similar set of subjects. Can anyone suggest a more
appropriate way to interpret this data ?
Thanks in advance for any insights,
Mark Slifstein
Brain Imaging Division
Department of Neuroscience
New York State Psychiatric Institute
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|