Dear Patrick,
On Tue, 8 Jan 2019 18:15:42 +0000, Patrick <[log in to unmask]> wrote:
> Dear Dr. Gaser, Dr. Flandin,
>
>Thank you for replying.
>Could you please elaborate on how to use the atlas definitions in the normalized space? Using the read_annotations script (or using cat_io_FreeSurfer), I can get the vertices and the label information. However, that doesn't help in moving from surface ROIs and applying them to functional data which is in volume space, right? Or perhaps I am missing something here?
There are several additionally steps necessary to map the (co-registered) functional data to the surface and use surface-based atlas definitions. I would rather propose to use volume-based atlases that are based on rsfMRI parcellations:
https://github.com/ThomasYeoLab/CBIG/tree/master/stable_projects/brain_parcellation/Schaefer2018_LocalGlobal/Parcellations/MNI
These atlases are much easier to use and if your functional data are in MNI space this should be quite straightforward. In future I will also provide a pipeline for the surface-based approach in CAT12 to use surface-based atlases for functional and even DTI data. But this will take more time to develop.
>I also had a look at spm_mesh_to_grid which looks promising. However, I couldn't figure out how to specify "T" (array of data to be interpolated). Say I read in the surface data using cat_io_FreeSurfer which results in a structure having faces and verticies. I have the mat file from avg152T1. What goes in as an input for T? I am guessing it would be the label ID for the vertices but I get a dimension mismatch error at line 56.
EVen if this tool provides a mapping from surface to volume space I am not sure how reliable this will be for transforming atlas ROI definitions.
>I have just begun my foray into surface based data so probably I am missing something that is straightforward. My apologies for that.
No, this is absolutely not straightforward...
Best,
Christian
>
>
>Thank you for your help
>
>Warm RegardsPB
> On Tuesday, January 8, 2019, 2:44:59 PM GMT+5:30, Christian Gaser <[log in to unmask]> wrote:
>
> Dear Patrick,
>
>On Mon, 7 Jan 2019 16:43:29 +0000, Patrick <[log in to unmask]> wrote:
>
>>Dear Dr. Gaser/CAT Users,
>>Greetings for the new year!
>>I have two questions:
>>1. Once the surface pipeline is run during the segmentation stage, is there a way to export the subject level regions from the surface atlases? I am interested in using the surface segmentation for functional connectivity analysis. Is such a thing possible?
>No, this is not possible because of the missing inverse surface deformation (from template to native space). However, you can use the atlas definitions in normalized space, which might be a good alternative in that case:
>[vertices,label,colortable]=read_annotation('../fsaverage/label/lh.aparc.annot');
>
>>2. Are there in-built utility in CAT to export surface files to volume file? From a brief survey, it seems that the reverse is possible (i.e. a volume file can be projected as a surface file). I am wondering if the reverse is also possible?
>>For both the above, I understand that moving between surface and volume can be inaccurate and problematic. From literature review, it seems that various groups have used freesurfer parcellation for graph theory analysis. I would like to know if I could do the same using CAT surface analysis.
>While the mapping from volume to surface is relatively straightforward, the opposite way is not possible. There is no point-to-point mapping possible and the values in the volume that are not crossing the surface should be interpolated.
>
>Best,
>
>Christian
>>
>>Many thanks for your time and help
>>
>>Best Regards
>>PB
|