yes, i've tried these options today...and they give me the same thing.
I've isolated the 3 columns and just fed that in.
I've also figured out the orientation of the coordinates when messing with the 3dUndump command as well. (right anterior--like Rob Cox suggested, it is different) Using this command I get a very similar looking spot, but in different locations as well. (even different than those locations from 3dSurf2Vol). Mostly now they are far outside the brain.
At least this tells us the coordinates are right in relation to each node, but they are all off by the same amount or not matching up with the gridparent the way we want it too....i think. so there's hope!
I'll keep plugging away.
-j9