Hello,
I have a question about the effects of FreeSurfer's resampling algorithms. The way our pipeline goes is that we convert our DICOM images into .nii.gz format using FSL tools. One of the commands we use is FSLCPGEOM. I was never sure why, but our script would always break without it. Looking into it in more detail now, I see that it is replacing the voxel size information in the Nifti header with the voxel size in the template that FSLCPGEOM is copying from. This is causing all of our data to show isovolumetric 1x1x1 information in the header, even if the voxels really are non-isovolumetric. I have been having a difficult time combining FreeSurfer results across our different studies and I am wondering if this might be the reason why. All of our studies have voxels acquired near 1x1x1 isovolumetric, usually no more than .3mm away in any direction. What would the effect be of pushing data through the FreeSurfer reconstruction pipeline that is not really 1x1x1 isovolumetric?
Thank you, Jeff Sadino
One more thing that might be useful: for my current dataset, everything was processed in FreeSurfer 4.3.1.
Thank you, Jeff
On Mon, Mar 7, 2011 at 12:41 PM, Jeff Sadino jsadino.queens@gmail.comwrote:
Hello,
I have a question about the effects of FreeSurfer's resampling algorithms. The way our pipeline goes is that we convert our DICOM images into .nii.gz format using FSL tools. One of the commands we use is FSLCPGEOM. I was never sure why, but our script would always break without it. Looking into it in more detail now, I see that it is replacing the voxel size information in the Nifti header with the voxel size in the template that FSLCPGEOM is copying from. This is causing all of our data to show isovolumetric 1x1x1 information in the header, even if the voxels really are non-isovolumetric. I have been having a difficult time combining FreeSurfer results across our different studies and I am wondering if this might be the reason why. All of our studies have voxels acquired near 1x1x1 isovolumetric, usually no more than .3mm away in any direction. What would the effect be of pushing data through the FreeSurfer reconstruction pipeline that is not really 1x1x1 isovolumetric?
Thank you, Jeff Sadino
Well, .3mm is still 30%, which means that your brains will show up 30% smaller/larger. This could mess up any number of things. Unfortunately, the right answer is to fix your data then re-run everything. Try using FreeSurfer's mri_convert to convert from dicom. You can also just pass the dicom directly to recon-all with the --i flag.
doug
Jeff Sadino wrote:
One more thing that might be useful: for my current dataset, everything was processed in FreeSurfer 4.3.1.
Thank you, Jeff
On Mon, Mar 7, 2011 at 12:41 PM, Jeff Sadino <jsadino.queens@gmail.com mailto:jsadino.queens@gmail.com> wrote:
Hello, I have a question about the effects of FreeSurfer's resampling algorithms. The way our pipeline goes is that we convert our DICOM images into .nii.gz format using FSL tools. One of the commands we use is FSLCPGEOM. I was never sure why, but our script would always break without it. Looking into it in more detail now, I see that it is replacing the voxel size information in the Nifti header with the voxel size in the template that FSLCPGEOM is copying from. This is causing all of our data to show isovolumetric 1x1x1 information in the header, even if the voxels really are non-isovolumetric. I have been having a difficult time combining FreeSurfer results across our different studies and I am wondering if this might be the reason why. All of our studies have voxels acquired near 1x1x1 isovolumetric, usually no more than .3mm away in any direction. What would the effect be of pushing data through the FreeSurfer reconstruction pipeline that is not really 1x1x1 isovolumetric? Thank you, Jeff Sadino
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
freesurfer@nmr.mgh.harvard.edu
-
Douglas N Greve -
Jeff Sadino