Print

Print
Hi Saad

Thanks a lot for your reply, much appreciated. That's good to know. I used the matrix1 option with 82 seeds, so I don't know what you mean with inclusion/exclusion mask. This is what I used:

/Applications/fsl/fsl/bin/probtrackx2 --network -x /Volumes/DTI/PROBTRACKX/HC005/masks.txt -l --onewaycondition --omatrix1 -c 0.2 -S 2000 --steplength=0.5 -P 5000 --fibthresh=0.01 --distthresh=0.0 --sampvox=0.0 --forcedir --opd -s /Applications/fsl/fsl/subjects/preprocessed/HC/005.bedpostX/merged -m /Applications/fsl/fsl/subjects/preprocessed/HC/005.bedpostX/nodif_brain_mask --dir=/Volumes/DTI/PROBTRACKX/HC005

Is it ok in my case to normalise by the values in the waytotal file?

Thanks again for your help.
André

Von: FSL - FMRIB's Software Library [[log in to unmask]]" im Auftrag von "Saad Jbabdi [[log in to unmask]]
Gesendet: Donnerstag, 26. November 2015 13:48
An: [log in to unmask]
Betreff: Re: [FSL] normalization of connectivity matrix after probtrackx

Hi André

We generally recommend normalising by the value stored in “waytotal”.  This normalises both for the number of voxels in the seed mask and for the fact that streamlines may be rejected during sampling (if inclusion/exclusion masks are being used).  I don’t think you need to normalise by the size of the target (would you expect twice as many samples if your target was twice as big?). The most sensible thing to do is to ensure the target mask is anatomically or functionally meaningful so you can e.g. compare across subjects etc.

Cheers,
Saad


On 25 Nov 2015, at 15:21, André Schmidt <[log in to unmask]> wrote:

Dear fsl experts, 

I am struggling with the normalization of my connectivity matrices obtained from probtrackx. I know that one suggested way for this is dividing the streamlines by the waytotal numbers. However, I want to normalize by taking into account the size of the source and the target seed - is it appropriate to divide the number of streamlines between two seeds by the product of the voxel numbers of these two seeds?

Thanks a lot for your help
André

------------------------------------------------------------------
Saad Jbabdi, PhD
Associate Professor
MRC Career Development Fellow

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