Dear Mr. Ashburner, dear SPM list.
Last year (Thu, 28 Oct 2010) there was an email conversation mentioning problems with the normalization parameter files written by an SPM "New Segment" toolbox (as it was seemingly called then).
I am having trouble using the normalization parameter files (and inverse parameter files) created using the "normal", integrated segmentation approach, with the versions SPM 8.3684 and 8.4290, getting practically the same error message as the person did back then.
Running 'Normalise: Write'
Failed 'Normalise: Write'
Reference to non-existent field 'VG'.
In file "/afs/cbs.mpg.de/software/spm/8.3684/7.11/2.7/spm_write_sn.m" (v2534), function "get_xyzmat" at line 430.
In file "/afs/cbs.mpg.de/software/spm/8.3684/7.11/2.7/spm_write_sn.m" (v2534), function "spm_write_sn" at line 96.
In file "/afs/cbs.mpg.de/software/spm/8.3684/7.11/2.7/config/spm_run_normalise_write.m" (v2312), function "spm_run_normalise_write" at line 25.
The following modules did not run:
Failed: Normalise: Write
Last year, Mr. Ashburner answered to the person having this problem:
"...The new segmentation does generate deformation fields,
but not in a form that is easy to use via the Normalise->Write option.
You would need to use the Deformations Utility (with the y_*.nii file
that is optionally generated), which is not so straightforward.
I will eventually get around to streamlining this procedure. ..."
Could it be, that the segmentation approach in SPM 8.3684 and 8.4290 still has this same problem...?
Since I really need to use the inverse normalization parameter files, in order to bring functionally defined ROIs back into the subjects' own space, I would like to ask for help.
I would use the Deformations Utility, but if it is not very straightforward, I would be extremely glad for any help on how to start, and information on what the problem with the deformation fields is and what we need to do to solve it (and what the y_*.nii file is, and how to obtain it).
Otherwise, could someone tell me how far I would have to go back into old SPM versions, to find one where the parameters are still given in a way that the normalize->write function can handle?
Barbara Vogt, PhD student
Max Planck Research Group 'Body & Self'
Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig
Stephanstr. 1a, D-04103 Leipzig, Germany
Phone: ++49 (0)341 9940-2669
Fax: ++49 (0)341 9940-2204