crash in RAP-MUSIC + filter issue

[cgbenar]

hello everybody

I have a crash when I click on the execute button of the "parameters RAP MUSIC" window

Error using ==> fileparts at 17
Input must be a row vector of characters.

I traced it to the fact that StudySubject.SubjectTess is empty

I am using MATLAB 7, R2007a

??

also: I noticed the high-pass filtering is bypassed in brainstorm_filt.m

all the best,

Christian

[Sylvain]

Hi Christian,

Did you compute a head model with the volume source grid box checked ? Also when you're data is all loaded and you are ready to press the 'execute' button type in the Matlab prompt:

Users = get_user_directory;
Study = find_brainstorm_structure(Users.CurrentData.StudyF ile)


Check whether Study.Subject and Study.SubjectTess are valid files and are not empty.

From you error message, it is likely Study.SubjectTess is empty. This field should contain the filename for tessellated head tissues.

Please first check everything above and we'll figure out together how to make things work for you.

[cgbenar]

hello Sylvain

- yes there is a grid for the BEM (VolGrid_CD) but not tessellation file

Subject: [1x116 char]
Study: [1x118 char]
SubjectImage: [1x74 char]
SubjectTess: []
SubjectTriConn: []
SubjectVertConn: []
Channel: [1x111 char]
HeadModel: [1x113 char]
Data: {[1x124 char] [1x115 char]}
Results: {[1x140 char] [1x140 char]}

cheers

Christian

[Sylvain]

Right.

Would you be so kind to try out a spherical head model first just to see whether RAP would work then ?

thanks - Sylvain.

[cgbenar]

Hi Sylvain

still the same crash with a spherical model

I also had the crash on the tutorial data (with spherical model)

:)

[Sylvain]

OK Christian,

please let me know which BST version you are using. If it's not July 10 or later, please update by downloading the thing again. Then I'll be able to help you.

Cheers!

[cgbenar]

yes the version is BrainStorm v. 2.2.38 (10-Jul-2007)

[Messor_structor]

any news on this ? Having the same problem with R2007a. Tried the version i downloaded this morning and the demodata with tutorial.

[Sylvain]

Sorry but I'm currently on travel so I have little time to test run things. Could you please tell me whether Minimum-Norm imaging do work with your installation?

Best,

[cgbenar]

yes Sylvain, the minimum norm works fine
Cheers
Christian

[Sylvain]

Hi Christian,

Sorry for this too long hiatus.

Could you please forward me the contents of Users.STUDIES and Users.SUBJECTS as described above. They should point at the folders where the anatomy and meg/eeg data are on your disk, respectively.

Subject: [1x116 char]
Study: [1x118 char]



Thanks.

[Sylvain]

Me again.

FYI, I was not able to reproduce this crash. I have just retested the RAP procedure with MEG data and it worked fine. Wo we should be able to fix things together once we have secured the integrity of yoru database and data.

Cheers

[cgbenar]

hello Sylvain!!; I'm back to testing, sorry for the delay;
I tried to trace again the bug; in dataplot_cb.m, after the line

StudySubject = find_brainstorm_structure(Users.CurrentData.StudyF ile);

I get :

StudySubject.SubjectTess[]

Subject: [1x105 char]
Study: [1x107 char]
SubjectImage: []
SubjectTess: []
SubjectTriConn: []
SubjectVertConn: []
Channel: [1x100 char]
HeadModel: [1x102 char]
Data: {[1x113 char] [1x104 char]}
Results: []


To answer your question on the database, I have no User variable in the matlab workspace; within the dataplot_cb function, I have
Users =
Comment: 'Marsepil'
STUDIES: 'C:\cgb2008\projects\localisation\datasets\marsepi l'
SUBJECTS: 'C:\cgb2008\projects\localisation\subjects'
FILELIST: [1x1 struct]
CurrentData: [1x1 struct]
and in
C:\cgb2008\projects\localisation\subjects
I have a folder
marsepil12_MARSEPIL_20060928_run02-f_Subjectds

with files

CTF_Fiducials.mat
brain_10000V_tess.mat
brain_10000V_tess_CTF.mat
cortex_10000V_tess.mat
cortex_10000V_tess_CTF.mat
marsepil12_MARSEPIL_20060928_run02-f_brainstormsubject.mat
outer_skull_10000V_tess.mat
outer_skull_10000V_tess_CTF.mat
scalp_10000V_tess.mat
scalp_10000V_tess_CTF.mat
scalp_10000V_tess_CTF_tess_CTF.mat
subjectimage.mat

:)

[Sylvain]

Hi Christian,

could you please let me know about the contents of Users.CurrentData.StudyFile ?

I think the problem might be coming from the fact that brainstorm does not find the tessellation files for your subject though they seem to be sitting appropriatly in a SUBJECT folder. Do you actually see this subject when you use the Subject editor in BrainStorm? http://neuroimage.usc.edu/brainstorm/GUI_BST_DM_SubjectEditor.htm

[cgbenar]

in parametric estimation, line 381
StudySubject = find_brainstorm_structure(User.CurrentData.StudyFi le);

I have User.CurrentData.StudyFile=
marsepil20\marsepil12_MARSEPIL_20060928_run02-f.ds\BrainStorm\marsepil12_MARSEPIL_20060928_run02-f_brainstormstudy.mat


I can see this subject in the study manager as well as the tesselation files
in the marsepil12_MARSEPIL_20060928_run02-f_brainstormsubject.mat, however, the tesselation variable is empty


cheers

Christian

[cgbenar]

hello Sylvain

maybe I found a way around, I added 2 lines in rap_music_parameters to avoid looking in an empty SubjectTess (code follows); can this lead to unexpected problems?

cheers

Christian


if(~isempty(StudySubject.SubjectTess)) % TEST CGB
[PATH,NAME,EXT,VER] = fileparts(StudySubject.SubjectTess);

TessRap = fullfile(PATH,[[NAME '_rap'] EXT VER]); % possible tesselation file
if(exist(TessRap,'file')), % there is a reduce one for rap
StudySubject.SubjectTess = TessRap; % map it in
end
end; % CGB END

[Sylvain]

Hmmm

maybe for visualization at the end of the rap_music estimation, but you will realize that rapidly. It's a bug I had never encountered before, probably because all my testing subjects had tessellation files in their folders. Thanks Christian!

How does it run now?