 |
 |
Dear Moises,Hmm, that might exist. The 90 anatomical ROIs are divided from the cortical region in the standard space, but during non-linear and linear transformation to the native space, the ROIs might overlap on their margins. So in this way, it might explain the exceeding part?Another important point I haven't made clear until now is, in the condition that the network-mode seedmask list is exactly the same with the --wtstop list, will these ROIs really function well as I expected to be both seedmasks and termination masks? Will the streamlines initiated successfully and then stop going once they enter and exit any of a target mask?Best regards,XinyangDo they overlap?Dear Moises,The 90 network seedmasks are all volume ROIs in nifti format, which were transformed from the AAL90 atlas in standard space.Best regards,Xinyang
At 2018-03-21 22:05:35, "Moises Hernandez" <[log in to unmask]> wrote:
are these ROIs volumes (nifti) or surfaces (vtk/ASCII/GIFTI)?Surfaces are defined by vertices and triangles.If the whole cortex is divided into ROIs according to a list of vertices,and all the triangles from these vertices are included in each ROI,ROIs that are neighbour contains triangles that are common to more than one ROI.Thus, a streamline crossing these triangles will be counted for more than one ROIOn 21 March 2018 at 04:28, Xinyang Liu <[log in to unmask]> wrote:Dear Moises,Hi. Thanks a lot for your reply.Our purpose is to acquire the ROI*ROI connectivity probabilistic matrix based on the AAL90 volume atlas. The command we use is probtrackx2 network mode.Our expected way of counting connectivity is that once the streamline reached a target ROI, it could stop entering the next region and therefore prevent repeated counting.We first added the --stop parameter to the network tracing, however it turned out the streamlines could not initiate even from the seedmasks. Then we had another try by adding the --wtstop, which contained the same list as the -x seedmasks list. In this way we wanted to let the streamline stop once exiting the target ROI.Our command was:probtrackx2 --network -l -c 0.2 -s /.../T1w/Diffusion.bedpostX/merged -m /.../T1w/Diffusion.bedpostX/ nodif_brain_mask.nii.gz -x /.../seed90_sciblade.txt -o 100206_network90 --dir=/.../output_net90 --forcedir --opd --omatrix1 -P 500 --wtstop=/.../seed90_sciblade. txt In this --wtstop case, we can successfully get the non-zero fdt_network_matrix and waytotal matrix. And such termination mask did work well when comparing the result with and without it. However, the sum of each row in fdt_network_matrix is larger than corresponding row in waytotal matrix.
waytotal 3314994 1482202 1607910 1424172 1181556 1096645 1362465 1275803 404120 354270 471835 406840 This is out of our expectation because we suppose each streamline should only arrive at one target ROI and the above relationship should be less or equal.Therefore, we wish to make clear the exact working mode here. Any guidance would be very appreciated.Best regards,Xinyang
At 2018-03-21 15:41:24, "Moises Hernandez" <[log in to unmask]> wrote:
HI Xinyang,how big is this difference?are these ROIs volumes or surfaces?MoisesOn 20 March 2018 at 23:01, Xinyang Liu <[log in to unmask]> wrote:Dear Matteo,Thanks a lot for your reply.Then there is a contradiction here that I cannot make clear for a long time. When I set all network mode seedmasks exactly the same ones as in the --wtstop list (probtrackx2 ... -x /.../seed90_sciblade.txt ... --wtstop=/.../seed90_sciblade.txt ), it means that these masks can be used as both waypoint masks which streamlines would go into and terminate masks from which the streamlines would stop on exiting them. How did it came out that waytotal row values were lower than corresponding sum of the fdt_network_matrix? To my understanding, in this situation, one streamline would reach only one target seedmask. I really cannot think out how it works here.Best regards,Xinyang
At 2018-03-21 01:32:42, "Matteo Bastiani" <[log in to unmask]> wrote:
Hi again,
yes, you can provide a text file with the different wt_stop masks separately.This will ensure that a streamline will be stopped when leaving one of the masks.
Cheers,Matteo
On 20 Mar 2018, at 14:40, Xinyang Liu <[log in to unmask]> wrote:
Hi, Matteo.
Maybe I should make my question a shorter one.So in the network mode prob-tracking, if I use the --wtstop parameter to make all the ROI seedmasks also as the wtstop masks, does it mean that the streamlines will stop on exiting every target seedmask? Because it seemed to be like this according to the output image. And also, the matrix seemed also work on that mode. Therefore I want to ask whether it is the case in theory.Look forward to your feedback. Thank you very much.
Best regards,Xinyang
At 2018-03-20 17:52:19, "Xinyang Liu" <[log in to unmask]> wrote:
Dear Matteo,
Thanks a lot for your kind feedback.I will try as you suggested. But before that, I still want to make clear some confusions about using --wtstop. Any guidance would be very appreciated.
1. According to one previous email about network mode tracking, our FSL expert said "if you use a wtstop mask for each ROI, the streamline will be able to exit from the ROI and enter into a second ROI. It will stop when it exits the second ROI." So will the streamlines in the second terminate ROI be counted in to the fdt_network_matrix after passing through the target ROI in the network mode?
2. If there is a --wtstop ROI mask between the a seed ROI and a target ROI in the network mode, will the streamlines passing through the --wtstop ROI and reach the target?
3. In a 90*90 ROI network, if I set the same 90 seedmasks as the --wtstop masks to avoid too far propagation and only focus on direct connections, how do the streamline goes? Because it seemed somehow different from the condition when defining one seed ROI and the rest 89 ROIs as --stop terminate masks as a pair-wise relationship.
Look forward to your feedback. Many thanks.
Best regards,Xinyang
At 2018-03-19 21:26:57, "Matteo Bastiani" <[log in to unmask]> wrote:
Hi again,
if you use the same stop mask as your seed mask, tracking won’t be initiated.Try defining some other anatomically meaningful stop masks that does not completely overlap with your seed masks, if you think you need those.
Hope this helps, cheers,Matteo
On 18 Mar 2018, at 12:20, Xinyang Liu <[log in to unmask]> wrote:
Dear Matteo,
Thanks a lot for your reply.Yes, the --wtstop will allow streamlines go into the next mask. Then I tried with the --stop parameter in the network mode prob-tracking. The command is as below.probtrackx2 --network -l -c 0.2 -s /.../T1w/Diffusion.bedpostX/merged -m /.../T1w/Diffusion.bedpostX/ nodif_brain_mask.nii.gz -x /.../seed90_sciblade.txt -o 100206_network90 --dir=/.../output_net90 --forcedir --opd --omatrix1 -P 500 --stop=/.../seed90_sciblade. txt
However, both the fdt_network_matrix and the waytotal matrix turned out to be zeros. Is there anything wrong with my script?
Best regards,Xinyang
At 2018-03-18 02:08:59, "Matteo Bastiani" <[log in to unmask]> wrote:
Hi Xinyang,
you can use stop masks to avoid streamlines reaching other locations.wt_stop masks allow propagation within this mask but terminate the streamline on exit. If your masks are neighbouring cortical regions, this might not help.
Hope this helps, cheers,Matteo
On 17 Mar 2018, at 12:28, Xinyang Liu <[log in to unmask]> wrote:
Dear Matteo,
Thank you very much for your kind and helpful reply.Then what should we do if we want to let the streamline stop when it has already reached one target ROI? Won't the --wtstop help?
Best regards,Xinyang
At 2018-03-17 20:22:35, "Matteo Bastiani" <[log in to unmask]> wrote:
Hi Xinyang,
that is because one single streamline can reach multiple target regions. The waytotal is the number of streamlines that go through all your other seed regions and waypoint masks (if you have used them).
Hope this helps, cheers,Matteo
On 17 Mar 2018, at 12:12, Xinyang Liu <[log in to unmask]> wrote:
Dear FSL experts,
Hi. We met a problem when computing the 90*90 ROI connectivity probabilistic matrix in probtrackx2 network mode. We planed to get the fdt_network_matrix and then use it to divide the waytotal matrix. However, we found that the sum of every row in fdt_network_matrix was a bit larger than the corresponding row of waytotal matrix. Our command was:
probtrackx2 --network -l -c 0.2 -s /.../T1w/Diffusion.bedpostX/merged -m /.../T1w/Diffusion.bedpostX/ nodif_brain_mask.nii.gz -x /.../seed90_sciblade.txt -o 100206_network90 --dir=/.../output_net90 --forcedir --opd --omatrix1 -P 500 --wtstop=/.../seed90_sciblade. txt
And the outcomes turned out to be:
waytotal 3314994 1482202 1607910 1424172 1181556 1096645 1362465 1275803 404120 354270 471835 406840 ...... .......
They were supposed to be the same. Could you please give us some guidance why these two items are different? Thank you very much.
Best regards,Xinyang