Hello Michael:
I don't know. I just now processed a 3d+time dataset using 3dDeconvolve with
-polort 3, and the results look okay.
Are you saying that the sub-brick labeled 'Base t^3 t-st' is identically zero?
Have you looked at this same dataset using the Deconvolution plugin, with Base
set to Cubic? How do the plugin values compare with the bucket dataset values?
They should be the same, up to round-off error.
Doug Ward