Matrix condition errors for block design

Hi, Qianqian-

The outputs here look like they came from processing with TENT(), but your provided afni_proc.py command above in the current thread has BLOCK(). So, it would first be good to match your afni_proc.py command with the output, to help address what is happening.

In this earlier thread with Gang and Rick, it looks like you moved your afni_proc.py processing to use TENT() (and Gang did mention there might still be potential for a collinearity issue). Is it possible that the outputs shown here came from the results of that processing?

Some things that might help clarify here:

  • At the bottom of the utilized proc script, there is a comment containing the afni_proc.py command that created it. In the posted proc script above, it wasn't included, but I do see "BLOCK()" used in 3dDeconvolve.
  • in the output stats file label snapshot, having so many Coef+Tstat pairs with the same prefix but then just changing "#0", "#1", "#2", "#3", etc. is a hallmark of running TENTS(), not "BLOCK".
  • afni_proc.py will not overwrite preexisting results directories or proc scripts---those have to be removed before rerunning a processing. So, when rerunning, it is good to use another name or move preexisting things, so there is no confusion.

--pt