Hi Doug,
Thanks for getting back to me. I'm right around the vicinity of my desired results, but had just three more questions.
1) When running the mri_vol2surf command and specifying --reg, is there any way I can pass in a registration file generated by FSL's flirt command (i.e. flirt -in functional.nii -ref structurual.nii.gz ... -omat functional.mat, where functional.mat is the registration file from my functional to my structural) Currently, I'm using a register.dat file generated by bbregister, but I'd rather use the flirt outputted regustration file because it's being used in other parts of my pipeline. Is there perhaps a way to convert from the flirt output to a register.dat file? I thought using bbregister with init-fsl might let me specify a flirt registration file, but this was not the case.
2) To the best of my understanding, when used on functional data and outputting to .w files (and subsequently viewing the .w file in ascii form), the mri_vol2surf command generates a surface file where each vertex has attached to it a weight corresponding to the functional activity. However, how does this work when I use a 4D functional file as input? I'd expect the same sort of output, but instead each vertex would have a series of weights (instead of just a single weight) corresponding to the functional activity at each timepoint. Even when I output to a .mgh file while using a 4D volume with mri_vol2surf and overlay the resulting .mgh file in tksurfer, it still seems like there is only one "weight" for each vertex. Am I missing something?
3) Is there a way to overlay the resulting .mgh (or .w or .asc) from mri_vol2surf in freeview?
Thanks again for all of your help.
Regards, Luis
----------------------------------------------------------------------
Message: 1 Date: Tue, 26 Jun 2012 12:54:20 -0400 From: Douglas N Greve greve@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] mri_vol2surf and Spherical Surfaces To: freesurfer@nmr.mgh.harvard.edu Message-ID: 4FE9E93C.3090206@nmr.mgh.harvard.edu Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Hi Luis, yes there is a more efficient way. Specify the native res functional as the --mov argument and specify the registration that you used to create the hires_functional as the --reg argument (instead of --regheader). While you can use a .w file, I would recommend instead a .mgh file (we're trying to phase out the .w); If you use mgh format, not need to spec a --out_type. That command will convert to the native subject surface. If you want to go to fsaverage you can either run mri_surf2surf after that or spec --trgsubject fsaverage doug
On 06/25/2012 03:43 PM, Luis R. Piloto wrote:
Hello Fellow Freesurfers,
I need some verification that my method for mapping functional data onto a spherical surface is sound.
What I'm currently doing is running recon-all on my high-resolution, structural volume (.nii.gz format). After that, I take a single timepoint of my functional data (that has been both registered and upsampled to the high-res structural space), and run it through mri_vol2surf with the following command: mri_vol2surf --mov highres_functional_slice00.niii.gz --hemi lh --regheader my_subject_name --out_type paint -o ./my_output_surface.w
What I'm most worried about is whether this method of registration (i.e. specifying --regheader) is sufficient to go from the upsampled/registered functional to the surface? Also is there some more efficient way of doing this? I feel like there might be a smarter way to go from the functional to the structural without having to explicitly register and upsample before using mri_vol2surf.
After this step, I simply pass my_output_surface.w to a program that reads the corresponding vertex and functional activity at that location and look up the vertex coordinates in my lh.sphere file. Any feedback would be really appreciated as I'm still really new to this software suite and was given a highly specific task to take care of. Thanks in advance!
Regards, Luis Piloto
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
On 07/02/2012 12:01 PM, Luis R. Piloto wrote:
Hi Doug,
Thanks for getting back to me. I'm right around the vicinity of my desired results, but had just three more questions.
- When running the mri_vol2surf command and specifying --reg, is there any way I can pass in a registration file generated by FSL's flirt command (i.e. flirt -in functional.nii -ref structurual.nii.gz ... -omat functional.mat, where functional.mat is the registration file from my functional to my structural) Currently, I'm using a register.dat file generated by bbregister, but I'd rather use the flirt outputted regustration file because it's being used in other parts of my pipeline. Is there perhaps a way to convert from the flirt output to a register.dat file? I thought using bbregister with init-fsl might let me specify a flirt registration file, but this was not the case.
You can convert it with something like tkregister2 --mov example_func.nii.gz --fsl fsl.mat --s subject --reg reg.fsl.dat
- To the best of my understanding, when used on functional data and outputting to .w files (and subsequently viewing the .w file in ascii form), the mri_vol2surf command generates a surface file where each vertex has attached to it a weight corresponding to the functional activity. However, how does this work when I use a 4D functional file as input? I'd expect the same sort of output, but instead each vertex would have a series of weights (instead of just a single weight) corresponding to the functional activity at each timepoint. Even when I output to a .mgh file while using a 4D volume with mri_vol2surf and overlay the resulting .mgh file in tksurfer, it still seems like there is only one "weight" for each vertex. Am I missing something?
It should output a series of weights if the input has multiple "frames". Check the input and output with mri_info. Don't use .w files, they only support one frame (and we're trying to phase them out).
- Is there a way to overlay the resulting .mgh (or .w or .asc) from mri_vol2surf in freeview?
Yes. I don't know off the top of my head. Maybe another from the FS team can chime in. doug
Thanks again for all of your help.
Regards, Luis
Message: 1 Date: Tue, 26 Jun 2012 12:54:20 -0400 From: Douglas N Grevegreve@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] mri_vol2surf and Spherical Surfaces To: freesurfer@nmr.mgh.harvard.edu Message-ID:4FE9E93C.3090206@nmr.mgh.harvard.edu Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Hi Luis, yes there is a more efficient way. Specify the native res functional as the --mov argument and specify the registration that you used to create the hires_functional as the --reg argument (instead of --regheader). While you can use a .w file, I would recommend instead a .mgh file (we're trying to phase out the .w); If you use mgh format, not need to spec a --out_type. That command will convert to the native subject surface. If you want to go to fsaverage you can either run mri_surf2surf after that or spec --trgsubject fsaverage doug
On 06/25/2012 03:43 PM, Luis R. Piloto wrote:
Hello Fellow Freesurfers,
I need some verification that my method for mapping functional data onto a spherical surface is sound.
What I'm currently doing is running recon-all on my high-resolution, structural volume (.nii.gz format). After that, I take a single timepoint of my functional data (that has been both registered and upsampled to the high-res structural space), and run it through mri_vol2surf with the following command: mri_vol2surf --mov highres_functional_slice00.niii.gz --hemi lh --regheader my_subject_name --out_type paint -o ./my_output_surface.w
What I'm most worried about is whether this method of registration (i.e. specifying --regheader) is sufficient to go from the upsampled/registered functional to the surface? Also is there some more efficient way of doing this? I feel like there might be a smarter way to go from the functional to the structural without having to explicitly register and upsample before using mri_vol2surf.
After this step, I simply pass my_output_surface.w to a program that reads the corresponding vertex and functional activity at that location and look up the vertex coordinates in my lh.sphere file. Any feedback would be really appreciated as I'm still really new to this software suite and was given a highly specific task to take care of. Thanks in advance!
Regards, Luis Piloto
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 -
Luis R. Piloto