One possibility is that the 7th regressor is causing some
multicollinearity problem that 3dDeconvolve is choking on
in some way.

Another is that you are at the edge of memory for your
computer, and some allocations are failing and causing
trouble. How much memory do you have (including swap)?

You might start by using 1dplot on all of the regressors,
and looking for whether the 7th matches another (it is
difficult to tell if it is a sum of others). Another
test might be to remove some (or all) of those motion
regressors, and re-run the command.

Does the -xout option actually output the X matrix?

- rick