22.0.0

Release Notes

New feature release in the 22.0.x series.

This release has been tested to work with Python 3.9 and FreeSurfer 7.2,
which are now bundled in the Docker image.

This release also features improvements to T2* and carpetplot reporting.

For the next release, we are investigating issues with susceptibility
distortion correction (SDC). Please check your results carefully and report
any issues you find.

Changes

• FIX: Update wrapper python path (#2783)
• FIX: Preserve collect_data behavior by using named args (#2754)
• FIX: Update fMRIPrep version in bibliography at run time (#2738)
• FIX: Plot carpetplot with CIFTI-specific colorbar (#2737)
• FIX: Two minor typos in report spec (#2708)
• ENH: Clip T2* values at 100ms to keep consistent histogram axes (#2781)
• ENH: Plot histogram of T2* values in gray-matter mask (#2778)
• ENH: Save T2starmap files in all requested output spaces, if calculated (#2776)
• ENH: Compare T2* map to BOLD reference (#2751)
• ENH: Add edge-regressors to confounds & crown to carpetplot (#2621)
• ENH: Add major/minor versions to base workflow name (#2716)
• DOC: Fix JSON typo in config in FAQ (#2771)
• DOC: Clarify calculation of confounding signals (#2724)
• MNT: Build on Python 3.9 environment (#2782)
• MNT: Seed ignore-revs file and script to tag new hashes with log entries (#2748)
• MNT: Require PyBIDS 0.15+ to allow fMRIPrep to preserve zero-padding in run entity (#2745)
• MNT: Upload artifacts after each step of Circle's workflow (#2736)
• MNT: Normalize code style of workflows.confounds (#2729)
• MNT: Ask for fmriprep-docker RUNNING line (#2670)
• DOCKER: Bundle FreeSurfer 7 (#2779)
• CI: Touch up CircleCI configuration (#2764)
• CI: Update package builds to use python -m build (#2746)

New Contributors

• @keithcallenberg made their first contribution in #2708
• @dillonplunkett made their first contribution in #2724
• @Remi-Gau made their first contribution in #2771

Full Changelog: 21.0.2...22.0.0

22.0.0rc4

This release candidate includes fixes for failing DVARS, running recon-all with T2w images, and misaligned GIFTI surfaces.

Full Changelog: 22.0.0rc3...22.0.0rc4

22.0.0rc3

This release candidate fixes an error that was encountered when filtering sessions with the --bids-filter-file.

22.0.0rc2

This release candidate includes missing FreeSurfer binaries that would result in recon-all sometimes failing.

22.0.0rc0

Release Notes

New feature release in the 22.0.x series.

This release has been tested to work with Python 3.9 and FreeSurfer 7.2,
which are now bundled in the Docker image.

This release also features improvements to T2* and carpetplot reporting.

Changes

• FIX: Update wrapper python path (#2783)
• FIX: Preserve collect_data behavior by using named args (#2754)
• FIX: Update fMRIPrep version in bibliography at run time (#2738)
• FIX: Plot carpetplot with CIFTI-specific colorbar (#2737)
• FIX: Two minor typos in report spec (#2708)
• ENH: Clip T2* values at 100ms to keep consistent histogram axes (#2781)
• ENH: Plot histogram of T2* values in gray-matter mask (#2778)
• ENH: Save T2starmap files in all requested output spaces, if calculated (#2776)
• ENH: Compare T2* map to BOLD reference (#2751)
• ENH: Add edge-regressors to confounds & crown to carpetplot (#2621)
• ENH: Add major/minor versions to base workflow name (#2716)
• DOC: Fix JSON typo in config in FAQ (#2771)
• DOC: Clarify calculation of confounding signals (#2724)
• MNT: Build on Python 3.9 environment (#2782)
• MNT: Seed ignore-revs file and script to tag new hashes with log entries (#2748)
• MNT: Require PyBIDS 0.15+ to allow fMRIPrep to preserve zero-padding in run entity (#2745)
• MNT: Upload artifacts after each step of Circle's workflow (#2736)
• MNT: Normalize code style of workflows.confounds (#2729)
• MNT: Ask for fmriprep-docker RUNNING line (#2670)
• DOCKER: Bundle FreeSurfer 7 (#2779)
• CI: Touch up CircleCI configuration (#2764)
• CI: Update package builds to use python -m build (#2746)

New Contributors

• @keithcallenberg made their first contribution in #2708
• @dillonplunkett made their first contribution in #2724
• @Remi-Gau made their first contribution in #2771

Full Changelog: 21.0.2...22.0.0rc0

21.0.2

Release Notes

This release includes a few bug-fixes for susceptibility distortion correction (SDC) and multi-echo (ME).
A few notable fixes include:

• Added tolerance for affine precision differences when using EPI fieldmaps.
• Removed hang-ups when reusing anatomical derivatives with ME data.
• Increased BOLD masking workflow robustness.

Changes

• DOCKER: Update multiarch deb package link (#2758)
• ENH: Add affine_tolerance flag to MergeSeries (nipreps/niworkflows#706)
• FIX: Initialize BIDS layout after cleaning working directory (#2741)
• FIX: Avoid double unwarping during resampling of processed multi-echo data (#2730)
• FIX: Account for potential lists of lists in multi-echo cases (nipreps/niworkflows#719)
• FIX: Improve reliability of BOLD masking workflow (nipreps/niworkflows#712)
• FIX: Relax tolerance for different affines when concatenating blips (nipreps/sdcflows#265)

21.0.1

Release Notes

Bug-fix release in the 21.0.x series.

This release re-enables reverse-PE EPI fieldmaps, i.e. EPIs acquired with the opposite phase-encoding direction to a BOLD series.
Additionally, the --use-syn-sdc option will now error by default if no PhaseEncodingDirection can be found for a BOLD series. To reduce this to a warning, use --use-syn-sdc warn. SyN-SDC will NOT be run, but the workflow will proceed.

NOTE:

topup runtime scales with the number of volumes used, so fMRIPrep defaults to using a maximum of 5 volumes from each of the BOLD series and the EPI fieldmap. To adjust this, please use the --topup-max-vols parameter.

Changes

• FIX: Create single fmap_select_std node for use with all ANAT estimators (#2692)
• FIX: Re-enable reverse-PE EPI fieldmaps (#2684)
• FIX: Pass boolean use_syn_sdc to find_estimators (#2685)
• FIX: Ensure AROMA inputs are uncompressed if --low-mem (#2657)
• ENH: Add --topup-max-vols flag to control TOPUP runtime (#2688)
• ENH: Allow --use-syn-sdc to take a "warn" option to avoid exiting when PE dir is unavailable (#2680)
• RF: Add extension field to differentiate from upcoming templates (#2613)
• WRAPPER: Map --bids-filter-file\s into Docker container (#2691)
• MNT: Silence warning from pybids; list packaging as dependency (#2689)
• MNT: Add netbase to Docker image to ensure correct datalad operation (#2682)
• MNT: Restore nitime to requirements, necessary for DVARS (#2678)

20.2.7 (Long-Term Support)

Release Notes

Bug-fix release in the 20.2.x LTS series.

Changes

• FIX: Clarify phase encoding direction, rather than axis (#2690)
• FIX: Ensure AROMA inputs are uncompressed if --low-mem (#2657)
• FIX: Add -basescale 1 parameter to avoid flirt scaling (#2624)
• WRAPPER: Map --bids-filter-file\s into Docker container (#2691)
• MNT: Add netbase to Docker image to ensure correct datalad operation (#2682)

21.0.0

Release Notes

A new series is finally here!

Some highlights of this version's new features include a substantial rework of fMRIPrep's
susceptibility distortion correction (including the addition of FSL's topup as the new
default PEPOLAR correction technique), improved flexibility for multiecho scans, and
dependency version upgrades within the Docker environment.

⚠️ CAUTION ⚠️

To follow BIDS Derivatives more closely, fMRIPrep's default output layout has changed.
To enable the legacy layout, use the --output-layout legacy flag.
For more information, please see https://fmriprep.org/en/latest/outputs.html#layout

As with all minor version increments, working directories from previous versions should not be reused.

Thank you for using fMRIPrep!

If you encounter any issues with this release, please let us know
by posting an issue on our GitHub page!

A full list of changes can be found below.

Changes

• DOC: Remove mention to epidewarp.fsl from NOTICE (#2629)
• DOC: Update description of output layout, add discussion of legacy mode (#2646)
• DOC: ME | Add missing output to documentation, improve boilerplate (#2608)
• DOC: Add --index-metadata to pybids layout faq (#2546)
• DOC: Add warning about slice timing correction in output documentation. (#2502)
• DOC: Update sample report (#2519)
• DOC: Refactor README and citing information (#2474)
• DOC: Point documentation at SDCFlows and remove SDC section (#2470)\
• DOC: Transfer duplicated documentation to www.nipreps.org (#2469)
• DOC: Better explanation on how spike regressors are generated (#2465)
• DOC: Clarify that res-2 entity does not mean 2mm (#2466)
• DOC: adds --mem metavar (#2378)
• DOC: skull stripping is forced by default (#2345)
• DOCKER: Shrink Docker image size (#2551)
• DOCKER: Strip ABI tag from libQt5Core.so.5 to prevent Singularity failures (#2535)
• ENH: Make bids the default output layout option (#2555)
• ENH: Mask multi-echo data with reference mask instead of echo-specific masks (#2349)
• ENH: Replace initial FLIRT with mri_coreg, use -basescale 1 for FLIRT-BBR (#2625)
• ENH: Add flag for STC reference time and set in all cases (#2520)
• ENH: Use BIDSLayoutIndexer and do not index unnecessary modalities (#2494)
• ENH: Allow users to download templateflow templates to directory (#2482)
• ENH: Detect 3D & too-short BOLD series, warn and skip run's workflow building (#2461)
• ENH: Modernize Dockerfile (#2409) (#2270)
• ENH: Leverage SDCFlows 2.x API (#2392)
• ENH: Tedana-friendly ME derivatives (#2575)
• ENH: Better integration of SDCFlows' unwarping (#2576)
• ENH: Refactor workflow to adopt SDCFlows' new correction workflow (#2547)
• FIX: Threshold boldref resampled values to be non-negative (#2630)
• FIX: Restore SyN-SDC (#2530)
• FIX: Confusing I/O field bold_mask in bold-to-bold resampling workflow (#2611)
• FIX: ME | Revise SDC compute graph (#2610)
• FIX: Propagate the SDC warp to resampling node also with ME (#2609)
• FIX: Optimize build job on CircleCI (#2605)
• FIX: Minor multiecho issues (#2574)
• FIX: Missing numpy import (#2533)
• FIX: Avoid eager imports from NiWorkflows causing API breaks (#2495)
• FIX: Split aCompCor into aCompCor, cCompCor, wCompCor (#2523)
• FIX: Avoid unnecessary connections based on branching logic (#2508)
• FIX: Permit missing TR to show PyBIDS error at workflow construction time (#2513)
• FIX: Simplify STC logic for too short BOLD series (#2489)
• FIX: Catch FreeSurfer error related to FIPS being enabled (#2490)
• FIX: DerivativesDataSink nondeterministic checksums fixed with niworkflows=~1.3.4 (#2458)
• FIX: Address dependency incompatibilities by pinning specific packages (#2463)
• FIX: Unprotected import of sentry_sdk, which is not a dependency (#2460)
• FIX: Clarify phase encoding direction, rather than axis (#2302)
• MAINT: Conform fast-track outputs to BIDS ordering (#2663)
• MAINT: Reduce redundant processing of ds005 (#2649)
• MAINT: Finalize the work we started with the refactor of the Docker image (#2601)
• MAINT: Run full workflows on tags (#2593)
• MAINT: Stage Python environment of Docker image from nipreps/miniconda (#2581)
• MAINT: Stop NeuroDebian dependency in Dockerfile (#2578)
• MAINT: Run pytest through GitHub actions (#2529)
• MAINT: Relicense +20.3.x - BSD-3-Clause -> Apache License 2.0 (#2325)
• MAINT: Add missing OASIS30 WM/BS probsegs (#2471)
• MAINT: Update BIDS validator to 1.8.0 (#2443)
• MAINT: Address the problems of a sloppy merge (#2468)
• MAINT: Containers - remove /root/.npm (#2464)
• MAINT: Remove traces of Xenial base (#2459)
• MAINT: Revise CircleCI's cache tags to avoid mixups between builds (#2457)
• MAINT: Back port CircleCI configuration from dev branch (#2456)
• MAINT: Increase minimum tedana version (#2366)
• MAINT: Docker hub has disabled password authentication (#2448)
• MAINT: Add Lea Waller to contributors.json (#2304)
• MAINT: Move away from nearly deprecated images (#2295)
• MAINT: Migrate from poldracklab to nipreps orgs (#2290)
• MAINT: Ease CI packaging tests (#2472)
• RF/FIX: Iterate over echo indices, not filenames, simplifying iteration logic (#2651)

21.0.0rc2

Release Notes

A new release candidate of the upcoming 21.0.x series.
Included are some vital bug fixes, as well as improvements to susceptibility distortion correction and multi-echo derivatives.

Changes

• DOC: ME | Add missing output to documentation, improve boilerplate (#2608)
• ENH: Tedana-friendly ME derivatives (#2575)
• ENH: Better integration of SDCFlows' unwarping (#2576)
• ENH: Refactor workflow to adopt SDCFlows' new correction workflow (#2547)
• FIX: Restore SyN-SDC (#2530)
• FIX: Confusing I/O field bold_mask in bold-to-bold resampling workflow (#2611)
• FIX: ME | Revise SDC compute graph (#2610)
• FIX: Propagate the SDC warp to resampling node also with ME (#2609)
• FIX: Optimize build job on CircleCI (#2605)
• FIX: Minor multiecho issues (#2574)
• MAINT: Finalize the work we started with the refactor of the Docker image (#2601)
• MAINT: Run full workflows on tags (#2593)
• MAINT: Stage Python environment of Docker image from nipreps/miniconda (#2581)
• MAINT: Stop NeuroDebian dependency in Dockerfile (#2578)