Dear Lien,
When you save an object map in Analyze AVW as an image, Analyze has no
way of knowing which image the object map was based on. It therefore
gets written with default voxel sizes of 1 x 1 x 1 mm.
If you change the voxel sizes of your ex_obj_img.hdr to the values in
the MPRAGE.hdr you should be fine.
We've got ancient in-house tools to manipulate 7.5 headers; otherwise
SPM99 used to do it... I understand John's reasoning about having
removed the facility from SPM2 onwards, but there definitely are
situations like the one you describe when such a manipulation is both
needed and legitimate.
Can you for example try "coregister -> reslice only" in SPM5 and reslice
the ex_obj_img to the MPRAGE? This _might_ work...
Good luck,
Alexander
-----Original Message-----
From: SPM (Statistical Parametric Mapping) [mailto:[log in to unmask]]
On Behalf Of Lien Le
Sent: 21 April 2008 19:25
To: [log in to unmask]
Subject: [SPM] Reading/Changing Header Files in SPM5
Hello,
I am using the Analyze 8.1 Region of Interest (ROI) module to define
regions and store them as object maps. I then converted the object maps
into an Analyze 7.5 image and header format because I want to load this
ROI
into SPM5 to use as a mask for statistical analyses.
However, when I load this ROI image (along with the MPRAGE volume that
Analyze 8.1 generated from the original DICOM files and on which the
tracings were done) into SPM5 it gets displayed incorrectly. The ROI
and
the MPRAGE image that I see in SPM5 are the mirror images of the same
ROI
and MPRAGE image that I saw in Analyze 8.1. The ROI is also much bigger
in
SPM5 than it ought to be. In SPM5, if I overlay the ROI onto the MPRAGE
volume (originally generated from Analyze 8.1), this ROI mask is no
where
near the original structure that it was traced from (e.g. the
hippocampal
ROI is not where the hippocampus is on the MPRAGE).
When I displayed this ROI (and overlaid it onto the MPRAGE volume image
generated from Analyze 8.1) using FSL view, everything looks correct
(e.g.
FSL does not display the mirror image of what I saw in Analyze, the ROI
is
not bigger in size, the hippocampal ROI lies directly over the actual
hippocampus in the MPRAGE). This suggests that SPM5 is reading the
header
information of the files generated from Analyze 8.1 incorrectly. Can
anyone let me know how to read and change the header files in SPM5?
Also,
is there a particular piece of information within the header file that I
can modify to correct for these problems?
Thank you,
Lien Le
Junior Specialist
Center for Mind and Brain
Neurocognitive Development Lab
(530) 747-3808
[log in to unmask]
|