Thanks! I ran the script as you suggested. The result was a 3d+time dataset along with the .dx , .dy , and .psi output files. Each of these files appears to be in two columns. The first column appears to be indicating the volume/acquisition and the second column looks as though it indicates the alignment (in mm) of that particular slice (out of 32).

When I run the standard 3dvolreg on the 2d-aligned time series, the resulting translocation output text files indicate minimal additional alignment in that particular (sagittal) plane.

I have one additional question though. I routinely use the -dfile output of 3dvolreg to create six motion vectors that I include in my 3dDeconvolve model. This output file now only has miniscule residual corrections in the sagittal plane, and thus may not adequately control for head nodding in the regression model.

Can I substitute the output from -rprefix (e.g. "2dcorrection.oldrms"), specifically the second column output as the motion-regressor for the motion in the sagittal plane?

Thanks for any advice,

Jim B