Lifespan Functional Preprocessing Release
This release is the version used for HCP Development and HCP Aging data functional preprocessing.
Added features and fixed various bugs for processing Lifespan HCP data:
- New debugging library, with error trapping turned on in all scripts as a default
- Increase dilation in fMRI subcortical segmentation to avoid invalid CIFTI files being produced, plus add checking for mismatching atlas subcortical segmentation and subject's brain mask (#163)
- Improve/fix masking of dMRI data for voxels outside of the FOV (#129)
- Add assessment of spatial coverage completeness within both fMRIVolume and fMRISurface (#138)
- Add capability to use FLAIR style acquisitions within FreeSurferPipeline.sh (#149)
- Improvements to dimensionality estimation of multi-run FIX (MR-FIX) (#150)
- Extend single-run FIX to support the MotionRegression=FALSE condition (#157)
- Compiled matlab now uses R2017b (same version as used for the current FSL FIX compiled matlab) (#159)
- dMRI QC report generated using FSL's eddy “QUAD” QC tool (#155)
- Added option to use custom brain image or mask in PreFreeSurfer pipeline (#133)
Added an option to process legacy (non-HCP style) data (#133, #134, #135, #136, #156). Specifically to:
- Allow structural processing without a high resolution T2w image
- Allow functional processing without susceptibility distortion correction
- Enable use of slice timing correction
- Enable use of expanded fMRI masks
- Enable use of a reference fMRI run with linear or additional nonlinear registration to the reference fMRI run
Note: The listed legacy processing functionality has to be explicitly enabled using the --processing-mode
parameter. The added options acknowledge the need to process legacy style datasets. However, in these cases, the quality of the results may suffer. Please see related usage information, warnings and processing mode information provided in pipelines for more details.
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)