The major first release v1.0.0 of hMRI-toolbox.

[v1.0.0]

Added

• option to choose different models and parameters for B1-correction of MTsat
• set default WM percent value in hmri_defaults
• spatial processing: add explicit mask creation and fix implicit mask (0 to NaN in float images)
• update FIL seste seq parameters in get_metadata_val_classic
• denoising module-first part: Java-Matlab interface for LCPCA denoising
• read EffectiveEchoTime in new TerraX Dicom format
• save LCPCA-denoising supplementary files as nifti instead of .mat
• parameter error maps
• robust combination of two runs using error maps
• add .orig file extension to .gitignore to prevent merge artifacts to be pushed
• update GUI code to enhance documentation for Proc. Smoothing
• denoising module (lcpca): set mat_intent fields to input file values instead of spm_create_vol setting of 'aligned'
• denoising module-second part: MPPCA denoising
• added imperfect spoiling correction coefficients for common FIL protocols (3T and 7T).
  Note that these coefficients are computed without the small angle approximation.
• Update defaults with new recommended options: no more small angle approximation, and using weighted least squares R2* fitting

Fixed

• replace datestr(now) with datetime('now') in line with MATLAB recommendation
• fix crash if input images have different matrix sizes, and warn
• make B1-map creation using 3DEPI SE/STE and AFI methods fall back to defaults without sidecar files, rather than crash
• Modify the filenames as files are copied to RFsensCalc to prevent overwriting in further processing
• batch interface now enforces the number of B1 input images correctly for B1 mapping methods which only need two images
• more informative error if optimization toolbox not present during NLLS R2* calculation
• fix 3D-EPI B1 mapping not using b1defaults for Triotim scanner
• use cell- instead of char- array to accommodate filenames of unequal length in RFsens
• prevent missing B1 map for MTsat spamming the log
• fix when no TE provided in 3DEPI SE/STE B1 mapping data
• fixes compatibility with spm/spm required due to refactoring that removed TEMPLATE field
• do not log ISC-applied to R1 in case of no-B1-corr and UNICORT
• account for diffusion while gradients are off when calculating imperfect spoiling correction parameters
• apply shift and diffusion operators in the correct order when calculating imperfect spoiling correction parameters
• Default AFI TR2/TR1 value now matches typical sequence order

Breaking changes

• Old, unused imperfect spoiling correction coefficients have been removed and replaced with new ones.
  If you need the old coefficients, you will need to add them back using a local defaults file.
• Fix inconsistency between new implementation of PD and old T2* weighting removal method
• Update imperfect spoiling correction with upstream bugfixes in EPG-X so computed coefficients will be different
• Default AFI TR2/TR1 value now based on the second input image having the longer TR, which may not be the case for older data

Pre-release for use and testing for the major public release v1.0.0 of hMRI-toolbox:

[beta-v1.0.0]

Added

• option to choose different models and parameters for B1-correction of MTsat
• set default WM percent value in hmri_defaults
• spatial processing: add explicit mask creation and fix implicit mask (0 to NaN in float images)
• update FIL seste seq parameters in get_metadata_val_classic
• denoising module-first part: Java-Matlab interface for LCPCA denoising
• read EffectiveEchoTime in new TerraX Dicom format
• save LCPCA-denoising supplementary files as nifti instead of .mat
• parameter error maps
• robust combination of two runs using error maps
• add .orig file extension to .gitignore to prevent merge artifacts to be pushed
• update GUI code to enhance documentation for Proc. Smoothing
• denoising module (lcpca): set mat_intent fields to input file values instead of spm_create_vol setting of 'aligned'
• denoising module-second part: MPPCA denoising
• added imperfect spoiling correction coefficients for common FIL protocols (3T and 7T).
  Note that these coefficients are computed without the small angle approximation.
• Update defaults with new recommended options: no more small angle approximation, and using weighted least squares R2* fitting

Fixed

• replace datestr(now) with datetime('now') in line with MATLAB recommendation
• fix crash if input images have different matrix sizes, and warn
• make B1-map creation using 3DEPI SE/STE and AFI methods fall back to defaults without sidecar files, rather than crash
• Modify the filenames as files are copied to RFsensCalc to prevent overwriting in further processing
• batch interface now enforces the number of B1 input images correctly for B1 mapping methods which only need two images
• more informative error if optimization toolbox not present during NLLS R2* calculation
• fix 3D-EPI B1 mapping not using b1defaults for Triotim scanner
• use cell- instead of char- array to accommodate filenames of unequal length in RFsens
• prevent missing B1 map for MTsat spamming the log
• fix when no TE provided in 3DEPI SE/STE B1 mapping data
• fixes compatibility with spm/spm required due to refactoring that removed TEMPLATE field
• do not log ISC-applied to R1 in case of no-B1-corr and UNICORT
• account for diffusion while gradients are off when calculating imperfect spoiling correction parameters
• apply shift and diffusion operators in the correct order when calculating imperfect spoiling correction parameters

Breaking changes

• Old, unused imperfect spoiling correction coefficients have been removed and replaced with new ones.
  If you need the old coefficients, you will need to add them back using a local defaults file.
• Fix inconsistency between new implementation of PD and old T2* weighting removal method
• Update imperfect spoiling correction with upstream bugfixes in EPG-X so computed coefficients will be different

[v0.6.1]

Fixed

• The local config files have been converted to scripts for compatibility with compiled version
• function-evaluate SPM-struct (preproc8.val) for SPM development version compatibility.
• copy acquisition metadata to TE=0 volumes in Results/Supplementary folder after map creation so they can be used as input to the toolbox if needed

[standalone-hMRItoolboxv0.6.1.zip]

Compiled with:

-Matlab version: 9.14.0.2239454 (R2023a) Update 1
-OS version: Debian GNU/Linux 11 (bullseye)

Added

• support for reading RepetitionTime from individual file metadata for AFI B1-mapping data (i.e. support for qMRI-BIDS formatted data)

Fixed

• issue #5: fixed version check for compiled toolbox
• QUIQI check: dependence on stats toolbox
• issue #14 (Spatial processing: Inverse deformation field moved along with forward deformation field to requested folder)
• issue #59: both the qform and the sform of the first PD-weighted image are now propagated to the quantitative maps, rather than just the sform

Breaking changes

• AFI B1 mapping data must be entered in the opposite order to previously (for sequence programmers: the assumption is now made that the order of alTR strictly reflects the order of acquisition, rather than being sorted). A warning will be printed if the toolbox detects that the data might have been provided in the wrong order (see changes made in #41). Note that while this is a breaking change, it should make data sorting more logical.

Added

• double angle mapping (DAM) B1 mapping option
• option to test ambiguous angle ranges beyond [0,90] and [90,180] degrees for SE/STE B1 mapping (nAmbiguousAngles b1 defaults parameter)
• various B1 input checking
• smoothing and masking options for all appropriate B1 mapping methods

Fixed

• issue #42 (Non-matching filename length can cause AFI B1 calculation crash)
• issue #46 (Toolbox expects SE/STE input sorted by echo and flip angle whereas BIDS sorts by flip angle then echo)
• issue #26 (Possible bug in calculation of B1 maps with SE/STE EPI)
• SE/STE B1 mapping calculation can now no longer select both ambiguous angles for a given SE/STE pair

Added

• weighted least squares R2* fitting
• cleaner input/output functions for map creation
• several unit tests
• nulling of bright voxels during unified segmentation to improve segmentation of anonymised data

• QUIQI added
• visual quality control tool added

fixes:

• Backwards compatibility restored: TR, TE and FA can now again be read from the nifti description field if needed

• Provides option to remove the small angle approximation when calculating R1 and PD
• Bugfix for reverse phase encoding direction 3D-EPI on Siemens scanners
• Experimental BIDS support

• added back accidentally removed B1 mapping options