On Nov 4, 2011, at 10:46 PM, yaping wang <ypwang.npu@gmail.com mailto:ypwang.npu@gmail.com> wrote:
Hi, Dear Dr Fischl, Sorry to interrupt you. I have a simple question to ask you. It's about the orientation of the input image for FreeSurfer. The version I used now is version 4.5.0.
First I referred to the "README" in ${FreeSurfer_Dir}/subjects, but I didn't found the directory named "ernie". So I referred the following images for the input image orientation: ${FreeSurfer_Dir}/subjects/bert/mri/orig/001.mgz (002.mgz and 003.mgz) and ${FreeSurfer_Dir}/subjects/bert/mri/rawavg.mgz ${FreeSurfer_Dir}/subjects/bert/mri/orig.mgz
I show them in "FreeSurfer_sample.JPG", the first row is the 001.mgz (the orientations of 002.mgz and 003.mgz are same with 001.mgz, so I just show 001.mgz), the second row is the the rawavg.mgz and the orig.mgz. From http://surfer.nmr.mgh.harvard.edu/fswiki/ReconAllDevTable, I think 001.mgz (002.mgz and 003.mgz) is the input image, and the rawavg.mgz and orig.mgz are the following two output images.
I test out data with "RAS" orientation as input image, it also gets the following two output image, and the orig.mgz has the same orientation as the sample image of Freesurfer. I show them in "ourTestData.JPG". Same with the previous figure, the first row is our input image, and the second row is the following two output images. But the origin is so different. So I have the following questions:
- What's the correct orientation FreeSurfer request? Is it correct for the orientation of the input image of our test
data? I have found that it can applied to "RAS" and "LAS" direction for "mri_watershed". So the input image can be any orientation, is it correct?
FreeSurfer does not require any particular orientation. The orientation information needs to be correct. Inside of FreeSurfer, it will automatically be converted to LIA.
- Will the "mri_robust_template" or "mri_add_xform_to_header" change
the orientation of the input image to the orientation FreeSurfer request?
No.
- What's the meaning of the origin shown in MRIcro? It comes from
the template or ...? If the "orig.mgz" of our test data has reoriented to the orientation of FreeeSurfer, why the origin is different the "orig.mgz" of FreeSurfer sample data? And why it's changing for our test data (from the input data--> rawavg.mgz---> orig.mgz, it's always changing)?
I don't know what MRIcro is doing. The origin usually means the xyz of the first voxel. FS changes the orientation to LIA (and voxel size to 1mm) because this is assumed in all subsequent processing.
doug
Thank you so much!
Best. Yaping <FreeSurfer_sample.JPG> <ourTestData.JPG>