Dear SPM:
I have a question about how to set up an analysis looking for
interaction effects. The experiment was originally an fmri study in
which there were rest and active scans. These scans were previously
combined for each subject using spm_fmri_adjmean.m, leaving me with a
rest scan and an active scan for each subject. What I want to do is
examine if there are any areas which are activated with respect to a
particular covariate AND whose activity to that covariate is
modulated by group membership. Here's what I did
1) mean correct covariate and setup for 2 scans per subject. So if
the mean corrected covariate is X then this becomes -X(i) X(i)
-X(i+1) X(i+1) etc...
2) Set up grouping indicator covariate- there are 13 subjects
belonging to group 1 and 14 belonging to group 2. Would I set this up
as
[repmat([-1 -1],1,13) repmat([1 1],1,14)] ? I did not mean correct
this covariate.
3) I then multiplied the grouping covariate by the mean corrected
covariate. The resulting interaction covariate is still mean
corrected it seems.
Then I included the interaction covariate as the covariate of
interest and the other 2 covariates (1 and 2) as covariates of no
interest. Is this correct?
Darren
Darren R. Gitelman, M.D.
-------------------------------------------------------------------------
Cognitive Neurology and Alzheimer's Disease Center
E-mail: [log in to unmask] WWW: http://www.brain.nwu.edu
Voice: (312) 908-9023 Fax: (312) 908-8789
Northwestern Univ., 320 E. Superior St., Searle 11-470, Chicago, IL 60611
-------------------------------------------------------------------------
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|