Back to the main page.

Bug 2115 - make a standard_fem volume conduction model

Reported 2013-04-16 14:48:00 +0200
Modified 2013-04-25 17:02:59 +0200
Product: FieldTrip
Component: forward
Version: unspecified
Hardware: PC
Operating System: Windows
Importance: P3 normal
Assigned to: Lilla Magyari
Depends on:
See also:

Lilla Magyari - 2013-04-16 14:48:25 +0200

start from download the fuzzy and discrete data store in ft/template/atlas/brainweb/standard_fuzzy.mat and standard_discrete.mat take the standard_discrete and use it on the wiki as examples/create_fem_model_from_brainweb store the standard fem model in ft/template/headmodel/stahndard_fem.mat

Lilla Magyari - 2013-04-16 14:58:13 +0200

document the standard_fem.mat model on the wiki under template/headmodel

Lilla Magyari - 2013-04-16 14:58:45 +0200

document the standard_fuzzy.mat and standard_discrete.mat on the wiki under template/atlas

Robert Oostenveld - 2013-04-18 10:49:47 +0200

I started creating the segmentations. Here for reference: mri = ft_read_mri('phantom_1.0mm_normal_crisp.mnc') mri.tissue = mri.anatomy mri = rmfield(mri, 'anatomy') mri.tissuelabel = {} mri.tissuelabel = {'CSF' 'Grey Matter' 'White Matter' 'Fat' 'Muscle/Skin' 'Skin' 'Skull' 'Glial Matter' 'Connective' }' segmentation = mri save brainweb_crisp.mat segmentation

Robert Oostenveld - 2013-04-18 20:10:09 +0200

*** Bug 1823 has been marked as a duplicate of this bug. ***

Robert Oostenveld - 2013-04-18 20:23:26 +0200

I used something like this to make the probabilistic representation. filename = { 'phantom_1.0mm_normal_csf.mnc' 'phantom_1.0mm_normal_gry.mnc' 'phantom_1.0mm_normal_wht.mnc' 'phantom_1.0mm_normal_fat.mnc' 'phantom_1.0mm_normal_m+s.mnc' 'phantom_1.0mm_normal_skn.mnc' 'phantom_1.0mm_normal_skl.mnc' 'phantom_1.0mm_normal_gli.mnc' 'phantom_1.0mm_normal_mit.mnc' 'phantom_1.0mm_normal_bck.mnc' }; tissuelabel = { 'csf' 'grey_matter' 'white_matter' 'fat' 'muscle_and_skin' 'skin' 'skull' 'glial_matter' 'connective' 'background' }; for i=1:10 mri = ft_read_mri(filename{i}); segmentation.(tissuelabel{i}) = mri.anatomy; end

Robert Oostenveld - 2013-04-18 20:24:16 +0200

(In reply to comment #5) I also renamed the tissuelabel for the crisp one for consistency with the probabilistic. All labels are in lowercase and use underscore instead of space.

Robert Oostenveld - 2013-04-18 20:32:22 +0200

roboos@mentat001> svn commit Adding atlas/brainweb Adding (bin) atlas/brainweb/brainweb_discrete.mat Adding (bin) atlas/brainweb/brainweb_fuzzy.mat Transmitting file data .. Committed revision 8056.

Robert Oostenveld - 2013-04-18 20:45:25 +0200

I have documented it at Lilla, the rest is up for you.

Lilla Magyari - 2013-04-25 15:13:12 +0200

(In reply to comment #8) Hi Robert, I have made the standard fem. I had the following problems: - the brainweb segmentation contained also the bar above the head. I circumvent this problem by segmenting the standard mri for scalp with FT. Such segmentation gives an inclusive (filled) scalpmask which do not contain the bar. I checked and the FT scalp contains and follows pretty well the contour of the brainweb skin. So, I set to 0 all voxels in the brainweb segmentation which were 0 in the FT scalp. So, I ended up with a brainweb segmentation without a bar. - I made a fem model from all tissues of the brainweb. Although it has 9 tissues instead of 5. I was not sure how to handle the extra 4 tissues (leave them out, merge with other tissues...etc.?). I did the following: I made the headmodel from all 9 tissues and I assigned the following connectivity values: csf 1.79, gray 0.33, white 0.14, skin 0.43, skull 0.01 All of these values were suggested by Johannes. glial matter: 0.14, same as white matter fat: 0.025, muscle_and_skin: 0.16 based on an article (I'll attach it). I think it would be nice first to compute a leadfield using the tutorial data and standard electrodes. And compare it to a BEM. How about that? Best, Lilla

Lilla Magyari - 2013-04-25 15:13:53 +0200

Created attachment 466 conductivities of fat and muscle

Robert Oostenveld - 2013-04-25 16:33:24 +0200

(In reply to comment #9) suggestions for improvement add vol.coordsys = 'mni'; look into papers describing the brainweb data tp determine whether connective -> skin or skin_and_muscle conductivity? glial_matter -> white or grey conductivity? change the two brainweb "atlases" so that they don''t have the bar in fieldtrip/template/atlas/brainweb/*.mat. add the coordsys=mni to the two brainweb atlases. give him a nose-job subsequently add some standard electrodes and calculate some leadfields, compare to standard_bem.

Robert Oostenveld - 2013-04-25 16:35:58 +0200

(In reply to comment #11) Please ask jan-mathijs (CC) how to use ft_volumesegment for giving him a nose-job (i.e. manual edit of some voxels).

Robert Oostenveld - 2013-04-25 17:02:59 +0200

reminder for myself: