JISCMail - FSL Archives

Dividing by waytotal is not going to give you valid probabilities in this case because the waytotal combines across all seed voxels. An alternative that can work here is to normalise the rows or columns of the matrix to sum to one.

Note that it is unusual to ask for these within-grey matter connectivity matrices as the orientational information is more reliable in the white matter. I think that your NxN matrices for within pallidum connectivity are going to be telling you indirectly about the coherence of the diffusion orientations (a little bit like in this paper : https://www.ncbi.nlm.nih.gov/pubmed/19332141 but with tractography instead of directly calculating coherence). So for example, if you have coherent orientations in the GP you should see a sparser NxN connectivity matrix than if you had random orientations. Whether this is an interesting measure for your application I don’t know :)

On 6 Mar 2019, at 09:17, Gayane Aghakhanyan <[log in to unmask]> wrote:

Dear FSL community,

I'm interested to compare the N by N connectivity probabilities of the globus pallidues externa (where N = 151 voxels) in two different groups of patients. I wish than to correlate the connectivity profile with the local TAU burden.

My idea was to run for each patient the probtrackx2 with matrix1 option (-- omatrix1) and with -x globus_pallidus_externa_mask. Than, if I understand correctly reading the mailing list queries, I should divide the 3rd column of my fdt_matrix1.dot file to waytotal to obtain the probabilities of the connectivities of each x voxel (1st column) to y voxel (2nd column). Next, I need to compare the matrices between two groups.
Is there any standardized approach for this purpose? code available for matlab or R, or any suggestions ?

Thanks a lot
Gayane

########################################################################

To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1