Print

Print
Hi everybody!

A followup and more problem.

I finally manage (thank to Saad) to get my seeds_to_target_image using probtrackx in target classification mode.

What I'd like to do now, is a sub-classification of those voxels that are more connected with the frontal cortex, using as target subregions of the frontal cortex itself (the idea is really similar to the paper of Tziortzi and colleagues (2014) Connectivity-based functional analysis of dopamine release in the striatum using diffusion-weighted MRI and positron emission tomography).

I thresholded the seeds_to_frontal images at 1/4 of the paths (5000) and then binarized, in order to create the seeds point for the next step.
I entered these seeds point in probtrackx, using two different subregions in the frontal cortex as target masks.

Please note that the call that I used for this second step is identical to the one I used for the first step, with the exception of the name of the file involved.

I checked, and both the seed and the targets are in diffusion space and where they should be.

The problem is that probtrackx load the sample, start tracking, but then spent 0 seconds (as seen on the terminal) trakcing, and then pass to the next subject.
In the output directory, I find the seeds_to_target images, but they contain only 0.

Have you any idea about what's happening now ?

Thank in advance

Best

Alain



Date: Wed, 23 Jul 2014 13:07:50 +0000
From: [log in to unmask]
Subject: Re: [FSL] classification tractography
To: [log in to unmask]

Yes that’s right.

On 23 Jul 2014, at 13:56, Alain Imaging <[log in to unmask]> wrote:

Yep Saad,
that was clear, thanks!

So, just to test my understanding: let's say that I want to create some ROI in the striatum of each subject, based on the connectivity of the striatum itself with the cerebral lobe. All I need is the os2t option, in order to select those voxel in my seed that are more connected to the target I'm interested in.
Am I right ?

Alain

Date: Wed, 23 Jul 2014 12:51:59 +0000
From: [log in to unmask]
Subject: Re: [FSL] classification tractography
To: [log in to unmask]

the omatrix2 and os2t options are independent from each other.

omatrix2 gives you an output matrix that has size SEEDSxTARGET, each counts the number of streamlines from each seed voxel to each voxel in a target mask (which you must provide with the otarget2 option).

os2t produces a set of output NIFTI files called seeds_to_<target>, each of which tells you how many streamlines reached each of your target masks from each seed voxel. To use this option, you must provide a list of target masks using the targetmasks option.


I hope this is now clear?

Cheers
Saad

On 23 Jul 2014, at 13:34, Alain Imaging <[log in to unmask]> wrote:

Hey!

Thanks for the rapidity!

Maybe it is not clear to me how omatrix and target works together.
I read again and again the wiki but I'm not sure to understand.
Can you explain me ?

Thanks a lot

Alain

Date: Wed, 23 Jul 2014 12:32:14 +0000
From: [log in to unmask]
Subject: Re: [FSL] classification tractography
To: [log in to unmask]

Hi - no the problem is that you are trying to use —omatrix2 option without providing a —otarget2 mask.

Cheers
Saad

On 23 Jul 2014, at 13:27, Alain Imaging <[log in to unmask]> wrote:

Dear all,

I'm (trying to) performing a classification tractography using probtrackx2.

I have one seed mask and 4 target masks (the four lobes).

Since I have a lot of subjects, I'm trying to script a routine in order to automatize the process.

Everything works smooth on the registration side of the script, but I can't go through the tracking.

Here's my call for the tracking

probtrackx2 -s ${subject}.bedpostX/merged -m ${subject}.bedpostX/nodif_brain_mask.nii.gz -x ${subject}.bedpostX/${subject}_my_seed.nii.gz -o ${subject}_seed2target --dir=${subject}.bedpostX/seed2tar --targetmasks=${subject}.bedpostX/${subject}_targetlist.txt --os2t=${subject}_s2t --omatrix2 -V  1 


And here's what I see on the terminal 



Log directory is: /data/user/Tract_shape_high/subject.bedpostX/seed2tar
basename   /data/user/Tract_shape_high/subject.bedpostX/merged
maskfile   /data/user/Tract_shape_high/subject.bedpostX/nodif_brain_mask.nii.gz
seeds      /data/user/Tract_shape_high/subject.bedpostX/subject_striatum_in_diff.nii.gz
output     subject_seed2target
verbose    1
nparticles 5000
nsteps     2000
usef       0
rseed      12345
randfib    0
fibst      1
Running in seedmask mode
load seeds
done.
Load bedpostx samples
1_1
1_2
1_3
2_1
2_2
2_3

nfibres  : 2
nsamples : 50

Done loading samples.
Image Exception : #22 :: Failed to read volume 
terminate called after throwing an instance of 'RBD_COMMON::BaseException'
probtrack_connectivty_atlas_step1.sh: line 90: 13270 Aborted                 (core dumped) probtrackx2 -s ${dirroot}${subject}.bedpostX/merged -m ${dirroot}${subject}.bedpostX/nodif_brain_mask.nii.gz -x ${dirroot}${subject}.bedpostX/${subject}_striatum_in_diff.nii.gz -o ${subject}_seed2target --dir=${dirroot}${subject}.bedpostX/seed2tar --targetmasks=${dirroot}${subject}.bedpostX/${subject}_targetlist.txt --os2t=${dirroot}${subject}_s2t --omatrix2 -V 1
user@berz:~/script/shell$  




Here's how the file targetlist.txt looks like


/data/user/Tract_shape_high/subject.bedpostX/subject_Occipital_in_diff.nii.gz
/data/user/Tract_shape_high/subject.bedpostX/subject_Parietal_in_diff.nii.gz
/data/user/Tract_shape_high/subject.bedpostX/subject_Temporal_in_diff.nii.gz
/data/user/Tract_shape_high/subject.bedpostX/subject_Frontal_in_diff.nii.gz



Plase note that the log file in the output directory do not contain error, but only the call.


It seems to me that the problem is in the target list (or in the target images), since the seed and the samples appear to have been loaded.


Any idea or suggestion ?

I'm stucked and a little bit frustrated: I manage to use convertwarp that to me has always looked like some kind of sci-fi thing, and now I'm stucked on this !


Thank you in advance


Alain

------------------------------------------------------------------
Saad Jbabdi, PhD
MRC Career Development Fellow &
University Research Lecturer

FMRIB Centre, University of Oxford,
John Radcliffe Hospital, Oxford, OX3 9DU, UK. 
tel (+44)1865-222466  (fax 717)

------------------------------------------------------------------
Saad Jbabdi, PhD
MRC Career Development Fellow &
University Research Lecturer

FMRIB Centre, University of Oxford,
John Radcliffe Hospital, Oxford, OX3 9DU, UK. 
tel (+44)1865-222466  (fax 717)

------------------------------------------------------------------
Saad Jbabdi, PhD
MRC Career Development Fellow &
University Research Lecturer

FMRIB Centre, University of Oxford,
John Radcliffe Hospital, Oxford, OX3 9DU, UK. 
tel (+44)1865-222466  (fax 717)