Dear Tracula team,
I have noticed in a multisubject tracula analysis, that the posterior density in many if not all paths in the path.pd.nii.gz file is just a bright one-dimensional curve, rather than a true distribution. The path.map.nii.gz is empty. This is true for both flirt and bbreg registered versions in many subjects. Trac-all finished wiithout errors. So did Recon-all.
Could it actually be a memory (RAM) problem ? We are running tracula on a cluster with only 2.5 GB per core.
thx for any help and kind regards, andi
Hi Andi - The most common problem that would cause all or almost all pathways to fail is an incorrect gradient table. You can do a sanity check on that by looking at the primary eigenvectors from your tensor fit (dmri/dtifit_V1.nii.gz). See also my answer to Fernando earlier today.
Hope this helps, a.y
On Tue, 30 Oct 2012, Andi Heckel wrote:
Dear Tracula team,
I have noticed in a multisubject tracula analysis, that the posterior density in many if not all paths in the path.pd.nii.gz file is just a bright one-dimensional curve, rather than a true distribution. The path.map.nii.gz is empty. This is true for both flirt and bbreg registered versions in many subjects. Trac-all finished wiithout errors. So did Recon-all.
Could it actually be a memory (RAM) problem ? We are running tracula on a cluster with only 2.5 GB per core.
thx for any help and kind regards, andi
Hi Andi - I managed to track this down by looking at the complete data set that you sent me. Your bvals file has no b=0, there's only b=2075 with some b=5 interspersed. The b=5 volumes must be your baseline volumes but unfortunately there's no way for the program to know this right now.
As a hack for now I'd just set the 5's to 0's in dmri/bvals and rerun the path reconstruction step, it should be a fine approximation for that step. I'll add support to this for the next version.
Sorry about that! a.y
On Tue, 30 Oct 2012, Andi Heckel wrote:
Dear Tracula team,
I have noticed in a multisubject tracula analysis, that the posterior density in many if not all paths in the path.pd.nii.gz file is just a bright one-dimensional curve, rather than a true distribution. The path.map.nii.gz is empty. This is true for both flirt and bbreg registered versions in many subjects. Trac-all finished wiithout errors. So did Recon-all.
Could it actually be a memory (RAM) problem ? We are running tracula on a cluster with only 2.5 GB per core.
thx for any help and kind regards, andi
Dear Freesurfers,
I am not very experienced with Freesurfer and would like to construct a Diffusion Tensor Image. I have the .bvecs, .bvals and .img files which I believe should suffice. I am aware that the format is slightly old but I believe it should still be fine. I attempted to use dt_recon and freeview but could not quite get the results I anticipated. I would really appreciate any help.
Many thanks in advance, Yaniv
Hi Yaniv - What results did you anticipate and what results did you get?
a.y
On Thu, 20 Dec 2012, Yaniv Kaufman wrote:
Dear Freesurfers,
I am not very experienced with Freesurfer and would like to construct a Diffusion Tensor Image. I have the .bvecs, .bvals and .img files which I believe should suffice. I am aware that the format is slightly old but I believe it should still be fine. I attempted to use dt_recon and freeview but could not quite get the results I anticipated. I would really appreciate any help.
Many thanks in advance, Yaniv _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
freesurfer@nmr.mgh.harvard.edu
-
Anastasia Yendiki -
Andi Heckel -
Yaniv Kaufman