Hello,
on behalf of a collegue who has not yet registered for the freesurfer mailing list I want to send you the questions below.
Thanks for your help
Claus
-----------------------------------------------------------------------
Hi,
we are currently trying to spatially transform functional data into the anatomical 256*256*256 1mm isovoxel COR-anatomy. We already have a register.dat and therefore a rotation-translation matrix. Our functional data are in a 128*128*24 bshorts file (1.406*1.406*2mm voxels). Slice orientation is approximately orthogonal to the calcarine sulcus Our problem is that the functional data do not fall on the correct location in the anatomical volume.
Our approach so far is: We built a datacube from the anatomical COR* files and oriented it with respect to the RAS system so that: R corresponds to -X (the inverted first coordinate direction in our data cube) A corresponds to Y S corresponds to Z
From the functional data (in bshort format) we reconstruct a similar volume and
orient it the same way. We assume the origin of the two coordinate system (the anatomical and the functional) in the center voxel of the two volumes
Whe now calculate [xAnat, yAnat, zAnat,1]'=[4*4-matrix from register.dat]*[xFunctional, yFunctional, zFunctional,1]' The x*, y* and z*s are in millimeters.
An example: we choose the voxel in the center of gravity of the functional volume [0,0,0,1]' and multiply it at the right of the matrix in register.dat. Shouldn't this give us the location of that voxel in the anatomical volume ?
Our qusestions: Can the register.dat be used to transform the functional data onto the COR-anatomy ? If yes how ? How does the coordinate system look like in both volumes especially the functional? The *.bhdr file related to the functional bshorts contains RAS-coordinates of the corners of the (first ?) slice. What is the corresponding coordinate system ? Scanner, COR-anatomy ...?
We already checked the wiki-site that describes the coordinate systems and transforms but this did not solve our problem.
Thank you very much for your help.
PS. The 4*4 transformation matrix we are using:
1.081372e+00 5.667235e-02 0.000000e+00 8.327621e+00 3.764737e-02 -7.183544e-01 -6.946586e-01 -6.509724e+01 3.635561e-02 -6.937064e-01 7.193404e-01 -4.349084e+01 0.000000e+00 0.000000e+00 0.000000e+00 1.000000e+00
Well, if the volumes do not line up, then you're register.dat is wrong. Use tkregister2 to register the original functinoal to the anatomical (use the --regheader the first time), then use mri_vol2vol to resample the functional into anat space (make sure to use the --invxfm flag with vol2vol).
doug
Claus Tempelmann wrote:
Hello,
on behalf of a collegue who has not yet registered for the freesurfer mailing list I want to send you the questions below.
Thanks for your help
Claus
Hi,
we are currently trying to spatially transform functional data into the anatomical 256*256*256 1mm isovoxel COR-anatomy. We already have a register.dat and therefore a rotation-translation matrix. Our functional data are in a 128*128*24 bshorts file (1.406*1.406*2mm voxels). Slice orientation is approximately orthogonal to the calcarine sulcus Our problem is that the functional data do not fall on the correct location in the anatomical volume.
Our approach so far is: We built a datacube from the anatomical COR* files and oriented it with respect to the RAS system so that: R corresponds to -X (the inverted first coordinate direction in our data cube) A corresponds to Y S corresponds to Z
From the functional data (in bshort format) we reconstruct a similar volume and
orient it the same way. We assume the origin of the two coordinate system (the anatomical and the functional) in the center voxel of the two volumes
Whe now calculate [xAnat, yAnat, zAnat,1]'=[4*4-matrix from register.dat]*[xFunctional, yFunctional, zFunctional,1]' The x*, y* and z*s are in millimeters.
An example: we choose the voxel in the center of gravity of the functional volume [0,0,0,1]' and multiply it at the right of the matrix in register.dat. Shouldn't this give us the location of that voxel in the anatomical volume ?
Our qusestions: Can the register.dat be used to transform the functional data onto the COR-anatomy ? If yes how ? How does the coordinate system look like in both volumes especially the functional? The *.bhdr file related to the functional bshorts contains RAS-coordinates of the corners of the (first ?) slice. What is the corresponding coordinate system ? Scanner, COR-anatomy ...?
We already checked the wiki-site that describes the coordinate systems and transforms but this did not solve our problem.
Thank you very much for your help.
PS. The 4*4 transformation matrix we are using:
1.081372e+00 5.667235e-02 0.000000e+00 8.327621e+00 3.764737e-02 -7.183544e-01 -6.946586e-01 -6.509724e+01 3.635561e-02 -6.937064e-01 7.193404e-01 -4.349084e+01 0.000000e+00 0.000000e+00 0.000000e+00 1.000000e+00
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
freesurfer@nmr.mgh.harvard.edu
-
Claus Tempelmann -
Doug Greve