incomplete whole-brain tractography

[Alistair Perry]

Hello,

After visualising the original tractography results (non-filtered) with CMTK below, something has obviously gone wrong in that only a relatively smalled amount of fibers have been tracked.

The input to diffusion toolkit was generated using dti_recon, and a wm mask was generated during the earlier stages of the pipeline.

 

runCmd     : Running: /usr/lib/python2.6/site-packages/cmp/binary/linux2/bit64/DTB_streamline –dir /home/alistair/networkD/projects2/2896/wave4/CMP/raw_diffusion/dti_0/dti_dir.nii –wm /home/alistair/networkD/projects2/2896/wave4/CMP/fs_output/HR__registered-TO-b0/fsmask_1mm__8bit.nii  –out /home/alistair/networkD/projects2/2896/wave4/CMP/fibers/streamline.trk –angle 60  –seeds 32

 

Could you pinpoint the inaccuracies in tractrography due to the DTI image and the mask, as you can see they have different dimensions to each other.

 

INFO     : 2012-11-27 14:04:04,331 : main       : -> Reading 'DIR' dataset…

INFO     : 2012-11-27 14:04:04,331 : main       : dim   : 120 x 120 x 69 x 12

INFO     : 2012-11-27 14:04:04,331 : main       : pixdim: 2.0000 x 2.0000 x 2.0000

INFO     : 2012-11-27 14:04:04,331 : main       : [ OK ]

INFO     : 2012-11-27 14:04:04,331 : main       : -> Reading 'MASK' image…

INFO     : 2012-11-27 14:04:04,331 : main       : dim   : 240 x 240 x 138

INFO     : 2012-11-27 14:04:04,332 : main       : pixdim: 1.0000 x 1.0000 x 1.0000

 

Or could it be perhaps due to an error a stage during dti_recon, as you can see here after inserting a custom gradient table.

 

DEBUG    : 2012-11-27 13:46:07,987 : runCmd     : Running: dti_recon /home/alistair/networkD/

projects/2896/wave4/CMP/raw_diffusion/DTI_resampled_2x2x2.nii.gz /home/alistair/networkD/proj

ects/2896/wave4/CMP/raw_diffusion/dti_0/dti_ -b0 1 -b 2400  -gm /home/alistair/networkD/dti/b

s/2896_w4.gradient -ot nii

INFO     : 2012-11-27 13:46:08,488 : main       : Need at least 6 gradient directions to reco

nstruct DTI data.

INFO     : 2012-11-27 13:46:08,488 : main       : Cannot load gradient matrix from /home/alis

tair/networkD/dti/bs/2896_w4.gradient

DEBUG    : 2012-11-27 13:46:07,987 : runCmd     : Running: dti_recon /home/alistair/networkD/projects/2896/wave4/CMP/raw_diffusion/DTI_resampled_2x2x2.nii.gz /home/alistair/networkD/projects/2896/wave4/CMP/raw_diffusion/dti_0/dti_ -b0 1 -b 2400  -gm /home/alistair/networkD/dti/bs/2896_w4.gradient -ot niiINFO     : 2012-11-27 13:46:08,488 : main       : Need at least 6 gradient directions to reconstruct DTI data.INFO     : 2012-11-27 13:46:08,488 : main       : Cannot load gradient matrix from /home/alistair/networkD/dti/bs/2896_w4.gradient
 

I manually changed the Maximum b-value to 2400.

 

I have visualised all the segmentations and registrations, and they are all correct.

 

Thanks.

 

 

[Qingyang Li]

Hi Alistair,

I don't think the tractography is incomplete. 

The figure you attached seems to be filtered by your viewer.

Here is some notes I took when I was learning to view fibers with Trackvis.

1. open TrackVis 

  File->open track. -> Load 

  $project_dir/$subID/$tp/CMP/fibers/streamlne_final_freesurferaparc.trk

2. slice filter

     change the value after "Y" to half of it's default (if it is 129 change to 64...).

     expand "Y" by clicking the triangle in front of it, then change the thickness to 129. This will gave you the space to see all the fibers in the brain.

good luck!

Yang