Dear Fredrik and Christian (x2),
I have never used CoSMoMVPA but it should be able to work with GIfTI files:
https://github.com/CoSMoMVPA/CoSMoMVPA/blob/master/mvpa/cosmo_check_external.m#L451-L455
Having SPM in the MATLAB path should be enough to get GIfTI support.
As Christian B. mentioned, current version of SPM allows to compute
approximate geodesic distances with Dijkstra's algorithm but next
version of spm_mesh_geodesic (in SPM next release) implements exact
geodesic distances. One could then write a version of spm_searchlight
that works with meshes instead of volumes.
Looking at Christian G.'s cat_io_data2mat, it seems useful to export
volume- or surface-based data as an array. I couldn't initially
understand how the geometric information was preserved (as required for
a searchlight analysis): I presume it is through the 'ind' variable?
Best regards,
Guillaume.
On 14/04/2021 20:17, Fredrik Bergström wrote:
> Dear Guillaume,
>
> thanks for your reply. Doing surface-based searchlight MVPA would also
> be an option, but e.g. CoSMoMVPA (and I believe I also tried TDT) cannot
> load .gii files, which is actually what prompted me to think about
> converting .gii to .nii instead. It would be ideal if there is some
> easy way of loading and converting the surface .gii files to 2d vectors
> that easily can be used by TDT or CoSMoMVPA, but I do not know any such
> toolbox or function.
>
> Best,
> Fredrik
>
>
> On Wed, Apr 14, 2021 at 7:18 PM Guillaume Flandin <[log in to unmask]
> <mailto:[log in to unmask]>> wrote:
>
> Dear Fredrik,
>
> Best would be to perform the external analysis directly on the surface.
> For a surface-based MVPA analysis, you have, for example, CoSMoMVPA:
> http://www.cosmomvpa.org/
> <https://eur01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.cosmomvpa.org%2F&data=04%7C01%7C%7Cc9d8a2971134412d2fea08d8ff79f368%7C1faf88fea9984c5b93c9210a11d9a5c2%7C0%7C0%7C637540246528950932%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=i1Lbz%2FHxtkaprv5NMR52f1kFgb7z2t6M%2FTSOG%2B%2BBwTE%3D&reserved=0>
> For the sake of completeness, functions to project surfaces to volume
> and vice-versa are spm_mesh_to_grid and spm_mesh_project but I get
> tut-tutted each time I mention them as they are unlikely to be the
> appropriate answer here.
>
> Best regards,
> Guillaume.
>
>
> On 14/04/2021 12:07, Fredrik Bergström wrote:
> > Dear all,
> >
> > Is there a straightforward way (for windows users) to convert surface
> > gifti files generated by CAT12 to volumetric nifti files for external
> > analysis (e.g., searchlight-based MVPA)? And possibly back again?
> >
> > Best,
> > Fredrik
>
> --
> Guillaume Flandin, PhD
> Wellcome Centre for Human Neuroimaging
> UCL Queen Square Institute of Neurology
> London WC1N 3BG
>
--
Guillaume Flandin, PhD
Wellcome Centre for Human Neuroimaging
UCL Queen Square Institute of Neurology
London WC1N 3BG
|