Hi Gari - The authors of probtrackx would be able to give you more information on how to normalize probtrackx outputs, but I'd suggest taking a look at the methods section of this paper as an example:
S. B. Eickhoff, S. Jbabdi, S. Caspers, A. R. Laird, P. T. Fox, K. Zilles, and T. E. J. Behrens, “Anatomical and Functional Connectivity of Cytoarchitectonic Areas within the Human Parietal Operculum,” J Neurosci, vol. 30, no. 18, pp. 6409–6421, May 2010.
Hope this helps, a.y
On Mon, 24 Feb 2014, Garikoitz Lerma-Usabiaga wrote:
Thanks Anastasia!
On Thu, Feb 6, 2014 at 1:11 AM, Anastasia Yendiki ayendiki@nmr.mgh.harvard.edu wrote:
Hi Gari - You're talking about thresholding the probtrackx maps, right?Yes If so the 20% at which the tracula maps are thresholded by default is probably too high for probtrackx, I'd try something like 2%. (Tracula is global and probtrackx is local tractography, so their outputs are different beasts.)
OK, thanks! The problem is that at 2% I think the tracts I obtain are too big. I attached a subject example (S_02) with thresholds at 2%, 5%, 20% and 30% in order to have an idea of what do you think it would be normal. The file is called tracts.png.
Inline image 2
Are the waytotal values very different among subjects?I think there is quite high variability, but I don't know what's normal. In the next figure, values.png, you can find the values:
- Waytotal values for the two new tracts
- Max values in the tracts after normalizing
- Seed mask size
Inline image 1
Are the anatomical seed mask sizes very different?
I don't think so. They are in values.png as well if you want to take a look.
Did you take that into account? Probtrackx will generate X sample paths for each voxel in the seed mask.
I didn't take either the waytotal or seed mask or waypoint mask sizes into account when thresholding. How would you recommend doing that?
Did you threshold the tracula tracts before using them as waymasks?
Yes, using fslmath I used -thrp 20 first to threshold them and after that I binarize it to create the waypoint.
Thank you very much for your help again, Gari
a.y
On Tue, 14 Jan 2014, Garikoitz Lerma-Usabiaga wrote: Hi Anastasia, it worked perfectly and I could run probtrack between my anatomical seed and using waypoints from Tracula tracts. I have another doubt right now. The probabilities in each new tract are very different. I divided every value with waytotal using fslmaths -div in order to normalize them, and obtained something like this: subject: S_20, Max Value after normalizing with waytotal: 0.75533 subject: S_23, Max Value after normalizing with waytotal: 0.27883 subject: S_25, Max Value after normalizing with waytotal: 0.60251 subject: S_27, Max Value after normalizing with waytotal: 0.32355 Should I consider the same absolute value for all tracts in order to threshold them or should I threshold all tracts independently with the 20%? I want to check visually the existence of the tract between the anatomical and the tract, with the same probability across subjects, and afterwards obtain the FA values in order to correlate with behavior. Thanks again! Gari On Mon, Jan 13, 2014 at 7:55 PM, Anastasia Yendiki <ayendiki@nmr.mgh.harvard.edu> wrote: Hi Gari - Yes, fslmaths should be able to do it, with -thrp 20. a.y On Mon, 13 Jan 2014, Garikoitz Lerma-Usabiaga wrote: Thanks Anastasia, I finally got it, I had to do mri_vol2vol twice to have it in the diffusion space. I have an additional question: I want to create waypoint masks for probtracx from Tracula tracts (for example, uncinate). Should I use mri_binarize on path.pd.nii to create the mask? Should I use 20% of the maximum value as the threshold in order to do so? (a little help with the command line will be greatly appreciated) thanks again! Gari On Sun, Jan 12, 2014 at 3:08 AM, Anastasia Yendiki <ayendiki@nmr.mgh.harvard.edu> wrote: Has mri/hippo_subfield_in_diffusion_space.nii been actually transformed to diffusion space by applying the anatorig2diff transformation? I'm assuming that your diffusion data is not 0.5mm resolution? You can search for the name of the transformation in trac-all.log to see examples of how it is applied to volumes. On Sat, 11 Jan 2014, Garikoitz Lerma-Usabiaga wrote: Hi Anastasia,the problem is with mri/hippo_subfield_in_diffusion_space.nii, which originally is mri/hippo_subfield.mgz, the output of running recon-all -hippo-subfields. So as far as I understand it is not in anatomical space, according to MRI info, it is (one example): - LIA - 85x97x141 - 0.5x0.5x0.5 So, how could I convert it to the diffussion space in FSL format in order to be able to use it in probtrackx? Should I do 2 steps or 3 steps? 1.- mri_convert to anatorig space (brainmask.mgz space, LIA, 256x256x256) 2.- anatorig2diff in order to change to FSL format and to diffusion space? (3.- would be the previous one in two steps, first from anatorig to anat and then from anat to diff) Any indication of the best programs to do all the change will be appreciated (I've seen in trac-all.log that mostly fsl programs are being used to do the flips, and then tkregister2 and bbregister for the registrations) Thanks for the help! Gari On Sat, Jan 11, 2014 at 1:31 AM, Anastasia Yendiki <ayendiki@nmr.mgh.harvard.edu> wrote: Hi Gari - So if I understand correctly, your inputs to probtrackx are: mri/hippo_subfield_in_diffusion_space.nii dmri.bedpostX/merged dmri.bedpostX/nodif_brain_mask If these all come from the output of tracula, aren't these all in LAS orientation? a.y On Thu, 9 Jan 2014, Garikoitz Lerma-Usabiaga wrote: Hi Anastasia,I want to use a hippocampal subfield as a seed in FSL. I am having problems understanding how to change spaces around... 1.- Is this command ok? mri_vol2vol --mov dmri/dtifit_FA.nii.gz --targ mri/hippo_subfield_in_anatomical_space.mgz --inv --interp nearest --o mri/hippo_subfield_in_diffusion_space.nii --reg dmri/xfms/anatorig2diff.bbr.dat --no-save-reg 2.- I would like to visualize it in freeview in anatomical space to check if it is ok, if I load nu.mgz and hippo_subfield_in_anatomical_space.mgz and then load hippo_subfield_in_diffusion_space.nii with the registration file dmri/xfms/anatorig2diff.bbr.dat, should I visualize it ok? If I don't register it, they show to be displaced (more than expected). 3.- Afterwards I want to run probtracx with the following command (my doubt is mainly about the files to use, since FSL files are RAS, freesurfer anatomical are LIA and freesurfer diffusion are LIA, are all the files in the same format and FSL is ok with it?): /usr/local/fsl/bin/probtrackx --mode=seedmask -x mri/hippo_subfield_in_diffusion_space.nii -s dmri.bedpostX/merged -m dmri.bedpostX/nodif_brain_mask -o fdt_paths_hippo_subfields --dir=FSL_ConnResults_Hipp -c 0.2 -S 2000 --steplength=0.5 -P 5000 --opd many thanks! Gari The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.