Hi Anita,
Since the required GOFORIT level is based on the number of warnings
that 3dDeconvolve spits out, it can vary. One "problem", like an
all-zero regressor, might result in many warnings.
The only point of the option is for you to acknowledge that you
understand the warnings and that you consider them to be okay (e.g.
the all-zero regressor is intentional, since this subject made no
mistakes on stimulus A). So you can use a larger number than is
necessary to be sure 3dDeconvolve runs.
But play close attention. If you accidentally use the same stim
times file for 2 different stimuli, it will hurt.
For efficiency, you could keep a large GOFORIT and study the stored
output of the processing script after it is done. If the warnings
suggest a problem, fix it and re-run. Otherwise, you think the
warnings are okay for that subject.
One way to test is to re-run 3dDeoncolve without any unnessary
output options (and change the names of those that are necessary).
If an all-zero regressor should be the only problem for example,
remove that regressor from the command and see if it runs without
any warnings.
Just please, be sure. It is worth the time. :)
- rick
