Add robust settings for EPI skullstripping using bet from FSL, in particular when there is strong signal bias.

update job_do_segmentCAT12

solution :
[,,idx_roi_in_netwok] = intersect(n.roi, ts_data.ts_table.abbreviation, 'stable')

this leads to pb when you have several sequences with different number of slices

n_outlier is printed, adding the ratio & percentage could help interpreting (even though the plot is very clear)

job_afni_proc_multi_echo does not handle .nii.gz file

TODO :

• make sure input can be .nii.gz or .nii
• make sure output can be .nii.gz or .nii

At the moment, with an unedited cat_defaults.m file, this cmd_prepend line is not enough to activate the expertgui option in the case if cluster execution.

solution
rgb_map_num = colormap(handles.axes);

solution : isfinite
logical(mask_global_Y(isfinite(mask_global_Y)));

this happens a lot during pilote study, its annoying...

Using cellfun(‘isempty’,…) should be much faster than cellfun(@isempty,…)
Need to check that !

Proposition

Create meinfo.mat file for each subject.

Disk access is neglectable I think, except if there are 100+ subjects, which is rare

• check there is no duplicate n-dir (cannot have in .dvs file 10dir and 10dir)
• fprintf(fid,'# on Diff card set b-value to b=%d \n', max(data(i).bval)); -> use the value after tolerance check
• use the JSON and the field ImageOrientationPatientDICOM : if the values are not all integer, apply rotation to the vectors ?

the code is done, need to wrap it in a function

Hi,

I am new to Github and relatively new to Matlab and was wondering whether you could give an example of how to use par as input for this function. Do you need to construct the variable beforehand? What if you would like to use the default?