List,
Please reference this previous Email for more details:
http://www.mailbase.ac.uk/lists/spm/2000-07/0301.html
Briefly, Stefan answered a question regarding a "multi-subject covariates
only" design, where there was only one scan per subject. The problem was
that there were "unestimable block effects" as the number of basis functions
equalled the number of scans plus one.
I had exactly the same problem...
The solution given was to use a "single subject covariate only" and treat
each scan as coming from the same subject.
My question: In this type of analysis, does one just ignore the individual
subject's variability? Is there no way to control for this, short of getting
more scans per subject?
Regards,
James
--
James Patterson
Louisiana State University
Biomedical Research Institute
PET Imaging Center
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|