Hi Andrew,

Given that 3dDeconvolve showed such strong complaints and
that 1d_tool.py did not (those correlation values are big,
but warning-big, not error-big :), it seems that some subset
of your regressors of interest is not linearly independent
(though individual pairs may be okay).

What does timing_tool.py say about rest? The program might
not work on AM2 formatted files, I'm not sure... You would
have to fill in XXXX (which might be a list if the durations
vary).

timing_tool.py -multi_timing stimuli/*.1D \
-run_len 298 -multi_stim_dur XXXX \
-multi_show_isi_stats

Also, do you have the libraries for xmat_tool.py to work?
That allows you to 'browse' subsets of regresors, say.

If you don't have the python libraries for xmat_tool.py,
feel free to mail me your xmat file, and I will take a
look.

- rick