Hello,
When trying to do a single-subject vs. group qdec study with covariates, I always get errors, such as:
ERROR: matrix is ill-conditioned or badly scaled, condno = 1e+08
Possible problem with experimental design:
Check for duplicate entries and/or lack of range of
continuous variables within a class.
This error seems quite sensible to me, and I understand why the GLM is unable to deal with it, but is there any way I can "force" it to use the covariates on the single subject, such as is possible in SPM5? Doing uncontrolled analyses are fine (I wouldn't trust the controlled data very much anyway) but it would, in theory anyway, eliminate the hassle of coming up with a well-matched control group for these kinds of analyses.
Any help would be appreciated; thanks as always.
Victor Laluz : Imaging Coordinator : UCSF Memory and Aging Center