Hi Rick;

rick reynolds wrote:
> How did you create this dataset? Was this with an AFNI program, or
> in some other way? If you are interested in uploading it so I can
> see, please see the upload link at the top of the main Message
> Board page.

I have uploaded welling_axial.tar.gz, containing warped_axial+orig.HEAD and
.BRIK; these are the files we've been talking about. The files were created by
the FIASCO program "pghtoafni", which I wrote, so it's all my fault. They do
load into AFNI correctly, and show the correct orientation within AFNI (as far
as I can tell). We interconvert between Pittsburgh MRI and AFNI files pretty
freely, using this program and the analogous AFNI-to-PghMRI reader.

> It's nice to have an example to illustrate the issues that you are
> talking about. But your axial+orig dataset looks very strange to
> me. The orientation appears to be RAS, yet the A to P coordinates
> go from +99 and desend with a DELTA of -0.9375. That does not seem
> right, as that should correspond to a dataset in RPS orientation,
> not RAS.


FIASCO uses a physical coordinate system where X is R-L, Y is P-A, and Z is S-I (which makes a right-handed coordinate system). The layout of the voxel data in the raster files is such that anterior rows come before posterior rows, though- this is a problem which you must also have encountered. This issue, plus my reading of the docs for the AFNI .HEAD file, plus a bunch of experimentation to get things to read into AFNI correctly, led me to the combination of ORIENT_SPECIFIC and DELTA values which appear in the file. I'd be happy to change them, if you can help me to understand the way to describe data in that order in the .BRIK file. The point is that the subject's nose comes earlier in the file than the back of their head, but the coordinate system in which the ORIGIN and the DELTA are specified must presumably remain right-handed.

Thanks,
-Joel