Mutual Nearest Neighbor dMRI-based parcellation

dMRIParcellation is a python toolbox for parcellating whole cortical surface using diffusion Magnetic Resonanse Imaging (dMRI) information and the Mutual Nearest Neighbor condition.

The parcellation algorithm needs the mesh connectivity, the tractograms of the seeds (vertices of the mesh) in NIfTi format ".nii.gz", and the coordinate of the seeds in the diffusion space in ".txt" file. The file name of tractogram, of seed i with coordinate (x,y,z), is be in the format of the probabilistic tractography of FSL i.e. file_prefix_x_y_z.nii.gz.

Requirements

1-h5py(>=2.6.0)

2-scipy(>=0.17.1)

3-numpy(>=1.11.3)

4-nibabel(>=2.1.0)

Cite

Brahim Belaoucha, Maurren Clerc and Théodore Papadopoulo, “Cortical Surface Parcellation via dMRI Using Mutual Nearset Neighbor Condition”, International Symposium on Biomedical Imaging, Apr 2016, Prague, Czech Republic. 2016.

Author

Belaoucha Brahim

Installation

python setup.py install

Parcellating whole cortex

MNNparcellation is the command used to parcellate the cortical surface. Check the file "./details/how_to_use.pdf" for more details or MNNparcellation --help.

Example

An example of parcellating a part of temporal lobe can be obtained by running:

sh example_parcellation.sh

The results are saved in ./data/Results/