Dear SPM:
A few questions about the correlation graph (upper right)
1) Is this graph a measure of othogonality between conditions, in a sense?
2) Is there a way to maximize orthogonality between conditions? I've tried
randomization, but I want to know how one can be sure the design is
maximally orthogonal.
For example: In a particular event related study I've designed with 4
conditions (1 null, and 3 response events) I still see correlations of
.32-.35 between the different conditions (excluding the null condition).
This seems high.
3) IMHO the graph might be more useful if you displayed it using the
following command imagesc(abs(corrcoef(sX)),[0 1]) (in
spm_fmri_design_show). Since ultimately the degree of correlation matters,
more than the sign??, this allows one to better view no correlation (0) vs.
high correlation.
thanks,
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
-------------------------------------------------------------------------
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|