Hi Jonatan
If you add the —network option, then probtrackx2 will output a text file called fdt_network_matrix which will contain the info you are after, i.e. the roi-by-roi connectivity matrix.
Cheers
Saad
> On 2 May 2016, at 19:11, Jonatan Snir <[log in to unmask]> wrote:
>
> Hello experts!
> A quick question that I feel there must be an easy solution for:
>
> I used 90 ROIS (functional network segmentation in MNI152 space) for a seed to seed connectivity using probtracx2.
> The output (fdt_matrix1.dot) is huge and I followed the instructions according to the user guide. I just wonder now how I can use the known order of masks in creating a 2D matrix of connectivity between each seed and all other seeds (90 in total).
>
> I could probably create a script to organize the information as necessary but I want to make sure I understand all the files created and their content. I apologize if this is listed adequately elsewhere. If the question above is fuzzy please continue reading:
>
> at the moment fdt_matrix1.dot is loaded into matlab (x=load('fdt_matrix1.dot');) and is a 56818x56818 sparse double and after I type M=spconvert(x);) it becomes 39182973x3 double. How do I use this matrix, combined with coords_for_fdt_matrix1 to create a 2D matrix of connectivity between each seed and another?
>
> This is what I think I know: the coords_for_fdt_matrix1 file appears to be a 5x56818 matrix. I assume the first 3 columns are coordinates in the MNI space, the 4th column is the ROI number (there are 90 of those) and the last column is simply an ID in order. The 3 columns of the fdt_matrix.dot are as follows: the 1st and 2nd columns are the seed voxel identification number that can be mapped based on this value back to coords_for_fdt_matrix1 line number. the 3rd column is the number of tracks connecting the two out of the total allowable (I set up 500 for now but understand that 5000 is preferable).
>
> If the above is correct and there isn't a simple existant script to rearrange the data for a 2D seedxseed matrix please let me know I'm at least right and I will write a script for that.
> Thanks so much,
> Jonatan
|