Lifespan Processing Bugfix Release
Various bug fixes, minor functionally additions, example script tweaks, and minor across script harmonization, including:
- Bash command syntax compatibility with standard MacOS installations
- Fix bug in
ReApplyFixPipeline
, which affects the volume (NIFTI) data, in which the non-highpass-filtered volume data was being used, rather than the hp-filtered volume data. - Make the polynomial detrending option that was added to
hcp_fix_multi_run
more numerically robust, and limit it to 5th order or less. - Adapted
RestingStateStats
to use a similar approach to the other pipelines for invoking matlab. - Created a script (in global/matlab) to compile all matlab code.
- Modified
MSMAll
scripts to support use of multi-run FIX-cleaned data. - Bug fixes so that calls to
topup
andflirt
work in the context of FSL 6.0+ (necessitated by changes in how FSL 6.0+ treats multi-frame inputs in certain circumstances). [This is not fully resolved, see below, and #118]. - Revisions to
SetUpHCPPipeline.sh
to allow some reliance on $PATH environment variable. - Some improvements to the logging functions.
- Add --delete-intermediates flags to FIX and ReApplyFix family of scripts.
Known issues:
- Input NIFTI files with a TR that does not use units of seconds result in incorrect TR of outputs - please set the TR in units of seconds before processing (#124)
- FSL 6.0.0 and 6.0.1 changed how they handle reference images when resampling compared to 5.x, and it is likely that not all pipeline scripts take care to use a single-frame file as reference image (#118)
- Due to a limitation in FIX's argument parsing,
hcp_fix
cannot currently be run with motion regression disabled (#108)