What Larmor does

Larmor is a Python library for diffusion MRI, with a viewer, BrainViewer. Each method is also a command of the lar command line, described in the command-line reference. The papers behind the methods are listed under Citing.

Preprocessing

• Denoising diffusion images with MP-PCA (lar denoise).

• Removing Gibbs ringing (lar degibbs).

• A binary brain mask from a diffusion image (lar mask).

Models

• DTI and DKI, fitted by ordinary or weighted least squares (lar dti, lar dki).

• Fibre orientation distributions (FODs) by spherical deconvolution, with the GRL and mFOD methods and the dRL, non-negative least squares and LASSO solvers (lar fod), and their conversion between SH bases (lar sh-convert).

• The peaks of the FODs (lar peaks).

Tractography

• Deterministic and probabilistic tracking through an FOD image (lar track).

• Editing tractograms: selecting streamlines by regions, length and weight (lar tract-edit).

• Profiles of maps such as FA along a bundle (lar tract-profile).

• Track density images (lar tract-density) and structural connectomes (lar connectome).

• Applying registrations computed by ANTs, FSL and MRtrix3 to images, tractograms and surfaces (lar transform).

Files

Larmor reads and writes NIfTI, MRtrix, TCK, TRK, VTK, FreeSurfer and GIfTI files, and reads MRIToolkit results saved as MATLAB files (lar convert, lar info); File formats lists them all.

The viewer

BrainViewer shows volumes, parcellations, tractograms, FODs, peaks and cortical surfaces in 2D and 3D (lar bv). See Viewer and Agents.