Thanks once again for all the help on this. I am running the GLM in native space, typically so that I can get the beta values to use in an MVPA analysis. So it sounds like if you specify per-run motion-correction and registration, and don't tell it to resample to anything (e.g., the MNI brain or the surface), then there's no guarantee that the voxels will properly line up if you want to do, for instance, a single-subject GLM in the subject's native space? So in this case, per-session motion correction would be better so that all the functional volumes in the session are registered properly?
Thank you, this is very helpful! A few quick followup questions:
- If you don't specifically tell preproc-sess what space to resample
too (e.g., by using the "-surface" or "-mni305" flags), does it automatically resample to the subject's anatomical volume, then?
I don't think it will do anything if you don't specify
- If fmc.nii.gz is already registered to the subject's volume, then
how come a registration file has to be specified when visualizing GLM results using tksurfer? If fmc.nii.gz is already resampled to anatomical space, then surely any GLM contrasts should be as well, no?
fmc is not registered to the anatomical, it is registered to the template. The template is registered to the anat through the registration file. The data are still in the native functional space.
- If per-run motion correction and registration is specified in
pre-proc sess, then there will be a different functional-->structural transformation for each run, correct?
Correct
If this is the case, then when visualizing session-wide GLM results using tksurfer, which registration file should be specified when loading up the overlay, since there'll be a different registration file for each run?
In what space are you doing the GLM? Usually, the raw data are resampled onto the surface (eg, fmc.fsaverage.lh.sm05.nii.gz), and the GLM is done in this spae.so no registration is needed.
Apologies for these (probably basic) questions, I just want to make sure I am crystal-clear on what preproc-sess is doing.
Cheers, and many thanks,
JohnMark
yes and yes
On 12/07/2017 02:23 PM, Taylor, Johnmark wrote:
Thanks once again for all the help on this. I am running the GLM in native space, typically so that I can get the beta values to use in an MVPA analysis. So it sounds like if you specify per-run motion-correction and registration, and don't tell it to resample to anything (e.g., the MNI brain or the surface), then there's no guarantee that the voxels will properly line up if you want to do, for instance, a single-subject GLM in the subject's native space? So in this case, per-session motion correction would be better so that all the functional volumes in the session are registered properly?
Thank you, this is very helpful! A few quick followup questions:
- If you don't specifically tell preproc-sess what space to resample
too (e.g., by using the "-surface" or "-mni305" flags), does it automatically resample to the subject's anatomical volume, then?
I don't think it will do anything if you don't specify
- If fmc.nii.gz is already registered to the subject's volume, then
how come a registration file has to be specified when visualizing GLM results using tksurfer? If fmc.nii.gz is already resampled to anatomical space, then surely any GLM contrasts should be as well, no?
fmc is not registered to the anatomical, it is registered to the template. The template is registered to the anat through the registration file. The data are still in the native functional space.
- If per-run motion correction and registration is specified in
pre-proc sess, then there will be a different functional-->structural transformation for each run, correct?
Correct
If this is the case, then when visualizing session-wide GLM results using tksurfer, which registration file should be specified when loading up the overlay, since there'll be a different registration file for each run?
In what space are you doing the GLM? Usually, the raw data are resampled onto the surface (eg, fmc.fsaverage.lh.sm05.nii.gz), and the GLM is done in this spae.so no registration is needed.
Apologies for these (probably basic) questions, I just want to make sure I am crystal-clear on what preproc-sess is doing.
Cheers, and many thanks,
JohnMark
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 -
Taylor, Johnmark