0.6.0

Release Notes

New feature release in the 0.6.x series.

This release respects recent changes to the BIDS-StatsModels draft specification to support fixed-effects meta-analysis (FEMA) contrasts, and renames "AutoContrasts" to "DummyContrasts".

Provisional support for F-tests has been added.

Additional rearchitecting by Dylan Nielson provides significant speedups for large datasets by caching BIDS layout information.

CHANGES

• SPEC: Add fixed effects (FEMA) contrasts (#191)
• SPEC: Change AutoContrasts to DummyContrasts (#197)
• FIX: Don't pass ignore and force index to init_fitlins_wf (#202)
• FIX: BIDSelect confusion between Nifti and JSON (#193)
• FIX: Set smoothing_fwhm after creating next level. (#190)
• FIX: Recognize cosine columns without underscores (#185)
• ENH: Update logging levels (#198)
• ENH: Add option to load BIDS layouts from database file (#187)
• ENH: Add memory estimate for first-level models, enable memory management (#199)
• ENH: Avoid casting BOLD data to float64 if possible (#196)
• ENH: Add F-tests (#195)
• ENH: Drop missing model inputs (#183)
• RF: Abstract interfaces to simplify swappability (#188)

0.5.1

Release Notes

Bug fix release to work with PyBIDS 0.9.4+.

CHANGES

• FIX: Expand entity whitelist (#182)
• FIX: Don't validate generated paths (#180)

0.5.0

Release Notes

This release features significant improvements to reporting and documentation,
including a Jupyter notebook to demonstrate usage. Example models are now in
the main branch of the repository, and annotated in the documentation.

CHANGES

• FIX: Smoothing level, check length (#157)
• FIX: mask parameter to FirstLevelModel is deprecated for mask_img (#158)
• ENH: Add ds000117 model (#163)
• ENH: Move to single-page report (#161)
• ENH: Add task-vs-baseline contrast to ds003 (#160)
• ENH: Reporting cleanups (#155)
• DOC: Curate reports and models (#153)
• DOC: Add example running through DS003 (#152)
• MAINT: BIDSLayout.get() parameter "extensions" deprecated (#167)
• MAINT: Update pybids dependency, package name (#166)

0.4.0

Release Notes

This release produces effect, variance, statistic (t or F), Z-score, and p-value
maps at every level, and enables smoothing at higher levels if preferred.

Additionally, documentation has been added at https://fitlins.readthedocs.io and
versioning/packaging issues have been resolved.

CHANGELOG

• FIX: Do not install FitLins as editable in Docker (#137)
• ENH: Save design matrix as TSV to output directory (#143)
• ENH: Enable smoothing at any analysis level (#135)
• ENH: Produce all available statistical maps from each analysis unit (#131)
• ENH: Add version to non-release Docker images. (#136)
• DOC: Flesh out documentation (#147)
• DOC: Build API docs on RTD (#146)
• DOC: Create Sphinx documentation with API autodocs (#145)
• MAINT: Drop Python 3.5 support (#140)
• CI: Run FitLins with coverage (#144)
• CI: Test FitLins on OpenNeuro DS000003, preprocessed with fMRIPrep 1.3.2 (#141)

0.3.0

This release restores reports at the second level and higher, and enables isotropic
smoothing with the nistats backend. Reporting has also been refactored to reduce
clutter in the outputs.

With thanks to Karolina Finc, Rastko Ciric and Mathias Goncalves for contributions.

• FIX: Restore level 2+ reports (#130)
• FIX: Remove uninformative metadata from derivative filenames (#129)
• FIX: Re-enable analysis level selection (#120)
• FIX: Switch plot colors to conventional blue for negative, red for positive (#108)
• ENH: Save crashfiles as text in working directory (#121)
• ENH: Add naive isotropic smoothing (#104)
• REF: Delegate isotropic smoothing to nistats (#118)
• DOC: Update README with latest help text, remove smoothing disclaimer (#119)
• MAINT: Add contributors to Zenodo (#122)
• MAINT: Consolidate configuration (#113)
• MAINT: Pybids 0.8 compatibility (#109)
• MAINT: Use numpy 1.15 to accommodate pytables (#106)

0.2.0

Release Notes

This release marks a substantial refactoring in the wake of BIDS Derivatives RC1, fMRIPrep 1.2.x and pybids 0.7.0.

Reports at second level and higher are currently broken, but we're at a point where neuroscout is depending on the current code base, the user base is increasing, and it's worth having a starting point for considering new features.

With thanks to Alejandro de la Vega, Adina Wagner and Yaroslav Halchenko for contributions.

CHANGES

• FIX: Allow derivatives to be a boolean value (#91)
• FIX: Restore report generation (#88)
• ENH: Plotting improvements (#89)
• ENH: Allow selecting for no space --space '' option (#96)
• ENH: Allow selecting for desc with --desc option (#95)
• MAINT: Depend on unreleased pybids commit (#99)
• MAINT: Pybids 0.7.0 compatibility (#84)

0.1.0

Release Notes

This release moves FitLins to a Nipype workflow and provides a set of Nipype interfaces for interacting with BIDS Models and the nistats statistical package.

CHANGES

• FIX: Correctly handle missing confounds (#73)
• ENH: Set loop_preproc during model loading (#66)
• REF: Second-level workflow (#30)
• DOC: Example model (#63)
• MAINT: Update nipype, grabbit and pybids dependencies (#70)

0.0.6

Hotfix release.

Changelog

• FIX: Explicitly create working directory (#61)

0.0.5

Changelog

• FIX: Limit NaN imputation and use mean non-zero value (#57)

0.0.4

With thanks to Alejandro De La Vega for many contributions.

Changelog

• ENH: Allow models without non-HRF variables (#55)
• ENH: Make dataset_description optional (#51)
• ENH: Loop over preproc files, instead of raw BOLD files (#50)
• ENH: Add --n-cpus option to CLI (#49)
• ENH: Run datasinks on main thread (#39)
• ENH: Enable derivative label to tag pipelines (#37)
• ENH: Make dataset description (#29)
• ENH: Add trivial dataset_description.json (#31)
• ENH: Run auto model by default (#26)
• ENH: Rewrite first level analysis as Nipype workflow (#16)
• ENH: Add acq, rec, run and echo to output patterns (#20)
• FIX: Second level contrast computation, and dense/sparse transformation issues (#48)
• FIX: Include run in derivative names (#43)
• FIX: Force string input to snake_to_camel (#41)
• FIX: Versioneer adjustments (#36)
• FIX: Create group level results dir (#22)
• RF: Simplify entry-points, restore preproc discovery (#38)
• MAINT: Update pybids 0.6.3, grabbit 0.2.1 (#52)
• MAINT: Manage version with versioneer (#35)
• MAINT: Neuroscout changes / pybids updates (#28)
• MAINT: Add nipype dependency, remove unused code (#27)