Skip to content

FSL 6.0.7.16, 12th December 2024

fsl-base 2408.0 -> 2411.0

Detailed history

  • BF: Adjust fslversion to check for different variants of conda env update commands when searching through env update history for "official" updates
  • MNT: Allow some make vars to be overridden by the env/prior to inclusion of default.mk
  • RF: Adjust construction of compiler flags so that $FSLDEVDIR/include takes precedence over $FSLDIR/include
  • DOC: Changelog, readme
  • BF: Typo in .c build rule - variable is ${CC}, not ${C}

fsl-data_omm 1.0 -> 1.1

The intent of fsl_get_standard is to allow users and scripts to dynamically select standard templates, instead of hard-coding paths to (e.g.) $FSLDIR/data/standard/MNI152_T1_2mm.nii.gz. Assuming that the OMM templates are installed into $FSLDIR/data/omm/Oxford-MM-1/, users/scripts can select the OMM templates by setting the FSL_STANDARD environment variable, e.g.:

$ fsl_get_standard
/usr/local/fsl/data/standard/MNI152_T1_2mm.nii.gz
$ export FSL_STANDARD=omm/Oxford-MM-1
$ fsl_get_standard
/usr/local/fsl/data/omm/Oxford-MM-1/OMM-1_T1_head.nii.gz

Or by using the --standard option, e.g.:

$ fsl_get_standard --standard omm/Oxford-MM-1
/usr/local/fsl/data/omm/Oxford-MM-1/OMM-1_T1_head.nii.gz
- Update transformation names for specificity Since there is a major breaking change in the naming convention (i.e., swapping the to/from meanings) I thought it would be better to just rename the files entirely. This will lead to scripts using these warps failing rather than mistakenly using the wrong transformations, which is preferable.

Detailed history

  • Add skeletonised FA for TBSS
  • Correct naming convention of warps NB! This is a breaking change! Essentially, the filenames have been swapped for MNI->OMM and OMM->MNI transformations to reflect *the direction in which they transform data*, rather than the direction in which the warp is moving points.
  • MNT: Manifest files for use by fsl_get_standard
  • RF: Use T2_FLAIR rather than T2
  • Add QSM volume
  • Update README
  • Update transformation names for specificity Since there is a major breaking change in the naming convention (i.e., swapping the to/from meanings) I thought it would be better to just rename the files entirely. This will lead to scripts using these warps failing rather than mistakenly using the wrong transformations, which is preferable.

fsl-fugue 2201.4 -> 2201.5

  • Fix OSX Packing On OSX repeatedly switching between the Siemens/GEHC modes will break the layout/widget packing.

This commit fixes by updating the label text instead.

Detailed history

  • BF: Fix packing issue on reported mode changes
  • MNT: Removed obsolete OCMR/Varian mode
  • MNT: Removed incomplete OCMR option
  • MNT: Restored copyright

fsl-utils 2203.4 -> 2203.5

  • MNT: Call omp_set_num_threads in addition to openblas_set_num_threads BLAS threading is disabled by default in FSL tools, as it usually does not give any performance benefit. When using an OpenMP build of OpenBLAS, the openblas_set_num_threads function used to call omp_set_num_threads. But this has changed at some point, and now one must explicitly call omp_set_num_threads in order to control the number of threads used by OpenBLAS.

Detailed history

  • MNT: Call omp_set_num_threads in addition to openblas_set_num_threads, as the latter no longer calls the former, and openblas calls may otherwise use multiple threads (op openblas-omp builds)

fsl-xtract 2.0.5 -> 2.0.6

Detailed history

  • Merge pull request #9 from SPMIC-UoN/xpy
  • bug fix: was saving left scalar map prediction as left and right results
  • Merge pull request #8 from SPMIC-UoN/xpy
  • added KLD scipy version
  • update normalisation
  • Merge pull request #7 from SPMIC-UoN/xpy
  • updated to include make_subcort_bp
  • added normalisation in case of tract subsetting
  • additional script to support subcortical blueprinting
  • Updates to report to make it more user friendly
  • Updated to use nomat2 for subcortex.
  • Updated to use nomat2 for subcortex. Added intent code for CIFTI (3006)
  • added hidden queue option