Repository navigation
0.3.0
Release 0.3.0
Brand new release 🎉 !
This release includes many new features, most notably the GPU profile, enabling GPU acceleration for eddy, tracking, and BIBSnet! The pipeline now supports a unified atlas scheme (either the built-in Brainnetome atlas or external user-supplied atlases) across the full developmental age range (more details are available in the documentation). Additionally, reading BIDS datasets now uses nf-bids, enabling faster loading and more complete BIDS support. Other (and not less important) changes are summarized below:
Added
- Changed segmentation in young participants to use BIBSnet and infant freesurfer (#128)
- Now use
nf-bidsplugin to read BIDS dataset for faster and more efficient data collection (#124) - Add GPU support via a config profile for module
eddyandlocaltracking(#121) - Add support for user-defined external atlases in ‘fsLR‘ 32k space and simply mapping from template to subject space (#120)
- Add RMSE and NRMSE maps to freewater and noddi output (#118)
- Support for
nextflow26.04.0 ([#116]) - Quality control section for connectivity data ([#116])
- Quality control visualization section for bundles ([#116])
- Improved visualization of cortical/subcortical segmentation results in native subject space ([#116])
Fixed
- Issues when multiple anatomical runs were found and not linked with the DWI acquisition(s) + tests for bidsignore (#126)
- Issues where
reconst_fw_noddiwould use too much RAM causing java heap errors (#118) - Issues where
NaNvalues would appear followingeddyand were makingamicomodels fail due to invalid voxels.NaNvalues are now quantified, reported, and converted to 0.
Changed
- Update to nf-core template
4.1.0(#129) - Update to nf-core template
4.0.3(#127) - Move to parameter typing and bump minimal nextflow version to
26.04.0(#122) - Update to nf-core template
4.0.2(#117)
Removed
- Remove M-CRIB-S as a segmentation method for younger participants in favor of BIBSnet and infantfs (#128)
Full Changelog: 0.2.2...0.3.0