Hi Michele - If there are large parts of the brain mask missing that
overlap with one of the tracts, the reconstruction of that tract will keep
failing. You should investigate why the aseg failed in that part of the
brain. It's hard to guess without seeing the images. If there's a large
artif
Hi Michele - It seems like it's failing in the right uncinate again. From
the error, I would guess that part of that tract is getting left out of
the brain mask or something of that sort. Can you please check how that
area looks in the aparc+aseg and in the diffusion data?
a.y
On Thu, 18 Se
Hi Barbara - It means that, depending on the peculiarities of a specific
data set, one tract may be trickier than others and the algorithm may need
to retry multiple initializations to get it right. It would be nice if it
figured that out by itself and reinitialized without you having to do it
Dear all,
What does this mean? That tracula is more robust when running it seperately
for the tracts I am interested in?
Thanks,
Barbara
On Mon, Sep 15, 2014 at 4:44 PM, Michele Cavallari <
cavallari.mich...@gmail.com> wrote:
> errata corrige: it did work! I had to re-run the whole thing includi
errata corrige: it did work! I had to re-run the whole thing including only
that specific tract...and eventually it worked.
On Fri, Sep 12, 2014 at 11:19 AM, Michele Cavallari <
cavallari.mich...@gmail.com> wrote:
> Tried...unfortunately it didn't improve
>
> On Thu, Sep 11, 2014 at 4:56 PM, Chri
Tried...unfortunately it didn't improve
On Thu, Sep 11, 2014 at 4:56 PM, Chris Watson <
christopher.wat...@childrens.harvard.edu> wrote:
> You can try just re-running "trac-prep -prior" and "trac-paths" for the
> L uncinate and no other tracts
>
> On 09/11/2014 02:58 PM, Michele Cavallari wrote
You can try just re-running "trac-prep -prior" and "trac-paths" for the
L uncinate and no other tracts
On 09/11/2014 02:58 PM, Michele Cavallari wrote:
So, I re-ran the case with the set reinit option. It "half"
worked!...in the sense that the new results show the left uncinate
right, but the
So, I re-ran the case with the set reinit option. It "half" worked!...in
the sense that the new results show the left uncinate right, but the
right-side one is still a dot (see screenshot of the brain view from the
bottom).
Any further suggestion?
Thanks.
[image: Inline image 2]
On Thu, Sep 1
Thanks! it's running...
On Thu, Sep 11, 2014 at 1:31 PM, Barbara Kreilkamp
wrote:
> Hi Michele,
>
> Don't think there is anything wrong with the attached dmrirc.tutorial file.
> You definitely need to add the '-c' flag infront of the path to your
> configuration file.
> Right now it reads the p
Hi Michele,
Don't think there is anything wrong with the attached dmrirc.tutorial file.
You definitely need to add the '-c' flag infront of the path to your
configuration file.
Right now it reads the path to your file as a flag, which is of course
not what you want.
Best,
Barbara
On 11/09/
Hi Anastasia,
I am probably doing something wrong with the syntax of the dmrirc file.
I get this error message
trac-all -prior
/Users/michele/Desktop/tracula/sages_diff_test_tracula/sages_uncinate_dmrirc.tutorial
ERROR: flag
/Users/michele/Desktop/tracula/sages_diff_test_tracula/sages_uncinate_d
Thanks, Michele. Hard to tell what's causing this, perhaps a bit of
distortion orbitofrontally. You may be able to fix this type of thing by
reinitializing the tract reconstruction: Add "set reinit = 1" to your
configuration file, set the pathlist to include only the left and right
uncinate,
uploaded (and activated).
Thanks!
On Wed, Sep 10, 2014 at 1:14 PM, Anastasia Yendiki <
ayend...@nmr.mgh.harvard.edu> wrote:
>
> Hi Michele - The anatomical segmentation does look good, but from the
> screenshot the DWI data seems to be noisy in the orbitofrontal area, which
> may be affecting th
Hi Michele - The anatomical segmentation does look good, but from the
screenshot the DWI data seems to be noisy in the orbitofrontal area, which
may be affecting the uncinate. It's hard to tell just from one slice.
If you upload all the tracula output directories of this subject (dmri,
dmri.
Hi Ludy - If your gradient table is formatted in 3 rows you need to
either:
1. Convert it to 3 columns so you can use it with the 5.3 version of
tracula, which requires the gradient table to be formatted in columns
OR
2. Download the tracula update that can use gradient tables formatted in
r
Hi,
I was having similar errors as Michele Cavallari regarding "error reading
/path/to/subject/dmri/dwi_frame," but I'm not sure it's related to my bvecs
file. I did try reconfiguring my bvecs file into columns instead of row just in
case, but that didn't solve the problem. It really just looks
Thanks Anastasia. I figured out the right path, and I'm now running the 2nd
step of tracula pipeline on that case.
There was a problem with subdirectories (my fault).
And also the missing 'SUBJECTS_DIR' syntax in the setenv command, as both
you and Barbara pointed out.
Thanks again for your help
Hi Michelle - Try replacing this line:
setenv
/Users/michele/Desktop/tracula/sages_diff_test_tracula/SD40_1/FSout_P0401_e20130813_s002_Cor7minMDEFT_T1_b
with:
setenv SUBJECTS_DIR /PHShome/my738/FSout
(Or whatever the location of your freesurfer recons is.)
a.y
On Wed, 3 Sep 2014, Mic
Hi Michelle - Based on your recon-all.log, your SUBJECTS_DIR that contains
the freesurfer recon of your subject is /PHShome/my738/FSout. But based on
you trac-all.log, you're using:
SUBJECTS_DIR /Users/michele/Desktop/tracula/sages_diff_test_tracula
So tracula cannot find the freesurfer recon
Hi Barbara and Anastasia,
thanks to your suggestions and help I was able to run the pre-processing
step of tracula.
I was using FreeSurfer 5.3, but didn't download any update. That's probably
why the row configuration was not working. Anyway, I used the column
configuration after manual correction
Hi Michelle - Your column-formatted gradient table (original_columns.bvec)
shows up as having only 2 rows in unix (e.g. with the cat command). That's
sometimes the case for text files saved from Excel or other non-unix
applications. When I open your file in TextEdit on a mac and then
copy-past
Hi Michele,
Yes, tracula would have complained at the very first call, if the dwi
data was not there.
So here is the problem, you write:
setenv
/Users/michele/Desktop/tracula/sages_diff_test_tracula/SD40_1/FSout_P0401_e20130813_s002_Cor7minMDEFT_T1_b
in the very first line in your c
Hi Barbara,
I checked the path. The folder with the trac-all output is is the very same
folder of the dwi dicoms.
Attached is the dmrirc file I am using, with all the specs.
Thanks
On Wed, Sep 3, 2014 at 12:59 PM, Barbara Kreilkamp
wrote:
> Hi Michele,
>
> You would have to make sure that the
Hi Michele,
You would have to make sure that the files are written to the same
folder for both commands (recon-all and trac-all)
Right now it seems you wrote the output of recon-all to
/PHShome/my738/FSout/FSout_P0401_e20130813_s002_Cor7minMDEFT_T1_b/
and trac-all output to
/Users/michele
Hi Michele,
This might be the solution - try the command 'more' in unix on your column
bvec and bval, they should each have one (bval) and three (bvec) columns
and no strange symbols. But you have these symbols in the bvec-file
(probably because you have extra formatting information, that is not n
Hi Michele,
Do you use the dicoms or the nifti files as input to TRACULA?
Also, I think the complete trac-all.log could be even more helpful.
I am a newbie myself but I am having a similar problem, so I am trying to
find a solution as well :).
Best,
Barbara
On Fri, Aug 29, 2014 at 4:48 PM, Mich
Hi Barbara,
thanks for your reply. I checked the dwi series on a viewer and it looks
fine: I have 1960 dicoms, corresponding to 56 slices x 35 directions
specified in the bval/bvec. Also, I was able to obtain the FA and MD maps
through fsl using the very same dicoms and bval/bvec.
Any suggestion?
T
Dear Michele,
I cannot find anything wrong with your bvec and bval files.
Seeing as you had an errormessage related to the dwi (the one about the
dwi_frame.nii.gz): Did you check that your dwi data have the same amount
of volumes as entries in bvec and bval?
Good luck,
Barbara
On 28/08/2014
28 matches
Mail list logo
