Dear Jan,
It seems that in your masterfile the paths to the lesion masks don't include directories (i.e. lesion.nii.gz instead of subj01/lesion.nii.gz), what results in either using the same mask for all the subjects or not detecting any mask at all. That would explain the error that labels from lesion masks were not found (in different space or non-existent). Hope it solves your problem!
Regards,
Hanna
_
Hanna Nowicka
DPhil student
FMRIB Analysis Group
University of Oxford
________________________________________
From: FSL - FMRIB's Software Library [[log in to unmask]] on behalf of Ludovica Griffanti [[log in to unmask]]
Sent: 13 July 2017 11:19
To: [log in to unmask]
Subject: Re: [FSL] BIANCA error
Hi Jan,
I am not sure about the ultimate cause of the error, but here are a few things you can do/check:
1) the transformation matrix needs to be the transformation from FLAIR to MNI, not from T1 to FLAIR. It is needed to calculate the MNI coordinates of the lesions as spatial features.
2) Is your FLAIR image brain-extracted? you need one of the images to be brain extracted and give the relative position to —brainmaskfeaturenum option.
3) the training subjects are probably not enough. Since BIANCA applies a leave-one-out approach, when segmenting subject n.1 it only uses 2 subjects as training. I would suggest to start with 10 (5 at the very minimum)
4) for the number of training points I suggest to use more nonlespts (non lesion training points) than trainingspts (lesion training points) as the non-lesion tissue is more heterogeneous and would benefit for more points. I would recommend using 10000 (or at least 5000) instead 1000
Let me know how it goes,
Ludovica
—
Ludovica Griffanti, PhD
Analysis Postdoctoral Research Assistant
Wellcome Centre for Integrative Neuroimaging (WIN)
Oxford Centre for Functional MRI of the Brain (FMRIB)
Nuffield Department of Clinical Neurosciences, University of Oxford
John Radcliffe Hospital
Oxford, OX3 9DU, UK
email: [log in to unmask]
> On 13 Jul 2017, at 10:12 am, Jan R. Schuere <[log in to unmask]> wrote:
>
> Dear Franklin,
> i have figured out my first problem and now I end up with almost the same error as you have.
>
> Traceback (most recent call last):
>
> File "/apps/prod/x86_64/fsl-5.0.10/bin/bianca", line 1033 in <module> p_main()
> File "/apps/prod/x86_64/fsl-5.0.10/bin/bianca", line 1007 in p_main()
> labels = labelvals[idvals]
> IndexError: index 8933309 is out of bounds for axis 1 with size 0
>
>
> My command:
>
> bianca --singlefile=masterfile.txt --labelfeaturenum=4 --brainmaskfeaturenum=1 --querysubject=1 --trainingsnums=1,2,3 --featuresubset=1,2 --matfeaturenum=3 --trainingspts=2000 --nonlespts=1000 --selectpts=noborder -o sub2bianca
>
>
> My masterfile:
> subj01/20170322_123358t2FLAIR3dsagipatspules028a001.nii.gz subj01/T1_2_Flair.nii.gz subj01/T1_2_Flair.mat lesion.nii.gz
> subj02/20170322_110016t2FLAIR3dsagipatspules028a001.nii.gz subj02/T1_2_Flair.nii.gz subj02/T1_2_Flair.mat lesion.nii.gz
> subj03/20170306_150739t2FLAIR3dsagipatspules029a001.nii.gz subj03/T1_2_Flair.nii.gz subj03/T1_2_Flair.mat lesion.nii.gz
>
>
> Background to Masterfile:
> 1 column is my FLAR image
> 2 column is my T1 registered on FLAIR
> 3 column is my transformation matrix
> 4 column is my binary roi with lesions, marked on space in FLAIR image
|