MatMRI: A GPU enabled package for MRI image reconstruction and processing

MatMRI is a package for MRI image reconstruction and processing using only core or toolbox MATLAB^® functions (i.e., no mex files or pre-compiled binaries). This philosophy is used to enable easy integration with built-in MATLAB^® GPU functionality, and to minimize dependencies. It currently supports:

• Reconstruction
  • non-Cartesian regridding
  • iterative SENSE (Cartesian or non-Cartesian)
  • iterative SENSE for higher order models that may include:
    • a B0 map
    • time varying spherical harmonics of phase accrual
  • compressed sensing with wavelet xform
• diffusion MRI fitting
  • spatially regularized diffusion kurtosis fitting with an axially symmetric model (nii2kurt.m)

Also see https://gitlab.com/cfmm/matlab/matmri

Cite as:

1. Varela-Mattatall G, Dubovan PI, Santini T, Gilbert KM, Menon RS, Baron CA. Single-shot spiral diffusion-weighted imaging at 7T using expanded encoding with compressed sensing. Magn Reson Med. 2023 Apr 10. doi: 10.1002/mrm.29666
2. Baron CA (2021, February 2). MatMRI: A GPU enabled package for model based MRI image reconstruction. Zenodo. http://doi.org/10.5281/zenodo.4495476

Additionally, please reference the following works for usage of the below functions:

• nii2kurt: Hamilton, J., Xu, K., Geremia, N., Prado, V. F., Prado, M. A. M., Brown, A., & Baron, C. A. (2024). Robust frequency-dependent diffusional kurtosis computation using an efficient direction scheme, axisymmetric modelling, and spatial regularization. Imaging Neuroscience, 2, 1–22.
• harmonicsFromRaw: Dubovan PI, Gilbert KM, Baron CA. A correction algorithm for improved magnetic field monitoring with distal field probes. Magn Reson Med.
• findDelAuto: Dubovan PI, Baron CA. Model-based determination of the synchronization delay between MRI and trajectory data. Magn Reson Med. 2022 Sep 26. doi:10.1002/mrm.29460.
• nufftOp: Baron CA, Dwork N, Pauly JM, Nishimura DG. Rapid compressed sensing reconstruction of 3D non-Cartesian MRI. Magn. Reson. Med. 2018;79:2685–2692.
• sampHighOrder: Wilm BJ, Barmet C, Pruessmann KP. Fast higher-order MR image reconstruction using singular-vector separation. IEEE Trans Med Imaging. 2012 Jul;31(7):1396-403. doi: 10.1109/TMI.2012.2190991. Epub 2012 Mar 14. Erratum in: IEEE Trans Med Imaging. 2012 Sep;31(9):1833.

Version notes:
0.1.09: added nii2kurt.m (kurtosis fitting assuming axial symmetry, with optional spatial regularization). Improved performance for sampHighOrder "interpolation" option.
0.1.07: improved GPU utilization for sampHighOrder.m. Considerably faster now, with slightly increased memory usage.