Francois Special
Fusion of four tractography/connectome pipelines (TractoFlow, RBx-Flow, Tractometry-Flow, Connectoflow)
Francois Special is a fusion of four pipelines: Tractoflow, RBx-Flow, Tractometry-Flow, and Connectoflow. Given PreQual-preprocessed diffusion data (.nii.gz, .bval, .bvec), a T1 image, and SLANT labels, it runs a full connectomics workflow: FOD estimation, probabilistic tractography, and connectome and tractometry construction, with reporting (report.html, report.pdf) and QC.
Usage
The repository ships with a Singularity recipe (singularity/singularity_recipe_francois_special.def) and the code in code/. Inputs are assumed to be PreQual-preprocessed, with a b-value in the range 800-1200 and at least 12 directions (tensor) or 800-3000 and at least 32 directions (fODF). Key parameters are documented in the repository README (sh_order, dti_shells, fodf_shells, pft_seed, local_seed, algo, nb_run, vote_ratio). The pipeline was optimized for bundles with large spatial extent extracted with RecoBundlesX from the RecoBundlesX atlas; to fully QA the data, inspect the label maps in MI-Brain.
The pipeline builds on the following:
Guillaume Theaud, Jean-Charles Houde, Antoine Bor, Francois Rheault, Francis Morency, and Maxime Descoteaux. “TractoFlow: A robust, efficient and reproducible diffusion MRI pipeline leveraging Nextflow & Singularity” NeuroImage, 2020.