[FieldTrip] "segmentation is not star-shaped" warning

[Maris Skujevskis]

Dear Fieldtrip community,

I am working on constructing an EEG volume conduction model. When running
ft_prepare_mesh I get a warning message the meaning if which is not clear,
hence I do not know how to fix it. Any help will be appreciated. Details
below.

Code executed:

>mri = ft_read_mri(file_name_mri);

>cfg                             = [ ];
>cfg.brainsmooth        = 5;
>cfg.scalpsmooth        = 4;
>cfg.brainthreshold     = 0.5;
>cfg.scalpthreshold     = 0.1;
>segm_mri  = ft_volumesegment(cfg, mri);

>cfg                          = [ ];
>cfg.method             = 'projectmesh';
>cfg.tissue               = {'brain','skull','scalp'};
>cfg.numvertices     = [3000 2000 1000];
>surf_mesh             = ft_prepare_mesh(cfg, segm_mri);

The code executes till the end, but I get this message:
>"Warning: the segmentation is not star-shaped, please check the surface
mesh"

When inspecting the mesh visually, I find no obvious abnormalities.
The meshes that do get the warning look more or less the same as the meshes
of other subjects that do not get the warning.

What is most interesting, when I use a segmented Fieldtrip
template ('~/matlab/fieldtrip/external/spm8/templates/T1.nii') as an input
to ft_prepare_mesh, I also get the same warning.

Here are 3 related questions:
What could be another way of checking/another indicator that the surface
meshes have been constructed incorrectly, since visually they seem ok?
Could changing the smoothing and thresholding parameters improve the
segmentation? (I haven't yet found values that do not produce the warning,
though.)
Is this an 'important' warning that signals that my subsequent source
localization might not work or might work incorrectly (i.e., I should not
just ignore it) ?

Thanks in advance,
Maris
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20150819/56c664ef/attachment-0001.html>