I have two sets of identical blocks within each fMRI run, so conditions {1,2,3,4,5,1,2,3,4,5}, and there are three separate runs of the two blocks.
If I run the analysis treating each with the contrasts set up with this numbering scheme (e.g. 3 vs 1 has two occurrences) , I get the following output from /usr/local/freesurfer/dev/fsfast/toolbox/fast_selxavg3.m
... ntptot = 687, nX = 136, DOF = 551 Saving X matrix to /autofs/cluster/birn/users/helmer/fMRI_improv/fs-fast/subj008/bold/improv.sm05.lh/Xtmp.mat XCond = 2926.01 (normalized)
However, if I want to treat the two sets of blocks separately, i.e., {1,2,3,4,5,6,7,8,9,10} so I can look at contrasts such as 3 vs 1 as well as 8 vs 6 then fast_selxavg3.m gives:
ntptot = 687, nX = 146, DOF = 541 Saving X matrix to /autofs/cluster/birn/users/helmer/fMRI_improv/fs-fast/subj008_10block/bold/improv.10blocks.sm05.lh/Xtmp.mat XCond = 4.33238e+08 (normalized) ERROR: design is ill-conditioned
1) why does nX increase by 10 in the 2nd case? 2) why is fast_selxavg3.m unhappy and XCond so large? Can it not handle 10 separate blocks?
The entire log file in the 2nd case is attached.