Automatic Bundle Segmentation

This section describes the integrated automatic virtual dissection of the main brain tracts based on auto_tracto. This tool was implemented using ANTs registration tools. The main purpose of the script 03_auto_tracts.sh is to split a tck tractogram into 35 white matter tracts, using already established automatic dissection protocols, which were manually tuned for optimal performance.

Derived from a full brain tractogram, 35 bundles are virtually dissected using the LANIREM protocols. The quality of the full brain tractogram will determine the quality of bundle separation. It is highly recommended to provide a tractogram with more than one million streamlines, and quality controlled for any errors. Strategies such as anatomically-contstrained tractography (ACT) and spherical deconvolution informed filtering of tractograms (SIFT), available in MRTrix3, should aid in obtaining such high-quality tractograms.

../../_images/autotract.png
Processing steps
Usage
Outputs
  • Non-linear (SyN) registration of the native FA map to the FA atlas (FMRIB58_FA_1mm)
  • Apply transformation to each bundle protocol to register them to the native FA space (DWI)
  • Verify stop criterion conditionals
  • Filter each white matter bundle according to the dissection protocols
There are two ways to run the automatic bundle segmentation. The first is integrated within the main script mica-pipe under the -SC module by adding a specific flag for this purpose, as shown in the following example:
Integrated usage within mica-pipe -SC
1
2
3
4
5
6
7
mica-pipe \
     -sub <subject> \
     -ses <session> \
     -out <outputDirectory> \
     -bids <BIDS> \
     -SC -autoTract \
     -tracts 40M

Automatic bundle segmentation can also be performed using the stand alone script 03_auto_tracts.sh. Note that the stand alone script requires that all inputs are in DWI space.

Example usage for standalone script
1
2
3
4
5
6
03_auto_tracts.sh \
     -tck sub-01_full_brain_tractogram.tck \
     -outbase sub-01_tract \
     -mask sub-01_binary_brain_mask.nii.gz \
     -fa sub-01_FA_map.nii.gz \
     -weights sub-01_full_brain_tractogram_weights.txt
Options Description
-tck path Full path to full tractogram file tck (ideally SIFTED)
-outbase string Base name for all outputs
-mask path Binary brain mask in the subject’s DWI space
fa path FA map in subject’s DWI space. This input is used for registration to the template

Additional optional arguments can also be specified:

Optional argument Description
-keep_tmp If specific, the temporary directory will not be delted after processing is completed
-tmpDir path Specify location of temporary directory
-minStreamlinesPerVoxel num Streamlines are truncated if voxel contains less than this number of streamlines. Default is 1
-robust This option to run a ROBUST registration, at the expense of more computation time
-weights path Use this option if you calculated a weights file from SIFT2
threads num Number of threads (Default is 6)
version` Print software version

Directories created by this script will be in the selected outbase. The outputs generated by the micapipe integrated script are located in <outputDirectory>/micapipe/<sub>/dwi/auto_tract

dwi/auto_tract/
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_AC.tck
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_AF_L.tck
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_AF_R.tck
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_atlas2fa.nii.gz
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_CC_MID.tck
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_CGFP_L.tck
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_CGFP_R.tck
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_CGH_L.tck
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_CGH_R.tck
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_CG_L.tck
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_CGR_L.tck
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_CGR_R.tck
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_CG_R.tck
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_CST_L.tck
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_CST_R.tck
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_FA_L.tck
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_FA_R.tck
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_FMA.tck
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_FMI.tck
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_FX_L.tck
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_FX_R.tck
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_IFOF_L.tck
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_IFOF_R.tck
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_ILF_L.tck
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_ILF_R.tck
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_MLF_L.tck
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_MLF_R.tck
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_OR_L.tck
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_OR_R.tck
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_SLF_L.tck
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_SLF_R.tck
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_summary.txt
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_TAPETUM.tck
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_UF_L.tck
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_UF_R.tck
├── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_VOF_L.tck
└── <sub>_space-dwi_desc-iFOD2-40M-SIFT2_VOF_R.tck

The description (desc-) refers to the algorithm used to generate the tractogram (here, iFOD2), its number of streamlines (here, 40M) and the filtering algorithm (here, SIFT2).

Full names for here tract and their associated acronyms are listed below:

Acronym Tract name
AC Anterior Commissure
AF Arcuate Fasciculus
CC_MID Corpus Callosum middle portion
CGFP Cingulum, fronto-parietal portion
CGH Cingulum, parahippocampal portion
CG Cingulum, whole
CGR Cingulum, rostral and subgenual
CST Corticospinal tract
FA Frontal aslant
FMA Forceps major of corpus callosum
FMI Forceps minor of corpus callosum
FX Fornix
IFOF Inferior fronto-occipital fasciculus
ILF Inferior longitudinal fasciculus
MLF Middle longitudinal fasciculus
OR Optic radiation
SLF Superior longitudinal fasciculus
UF Uncinate fasciculus
VOF Vertical occipital fasciculus