25.0.1

[handwerkerd]

Release Notes

🐛 Bug Fix and (un)breaking change

• Fixing unflipped PCA component weights by @handwerkerd in Fixing unflipped PCA component weights #1234
  Properly flip ICA mixing matrix time series #1212 was released with v25.0.0 and included an error where the PCA mixing matrix time series, but not the weight maps were multiplied by -1. This altered the ICA results in undocumented and probably improper ways. This PR fixes that bug. For identical parameters, running v24.0.02 and v25.0.1 in the same environment should result in nearly identical ICA mixing matrices (except for some time series being multipled by -1). People using v25.0.0 should update.

Breaking Change

• Replace --mir with --gscontrol mir in ica_reclassify by @tsalo in Replace --mir with --gscontrol mir in ica_reclassify #1222
  Changed so a parameter option in ica_reclassify more closely matches the naming scheme in tedana

Enhancements

• Add gscontrol figures by @tsalo in Add gscontrol figures #1221
• Add external regressor-mixing matrix correlation heat map by @tsalo in Add external regressor-mixing matrix correlation heat map #1226
  The above two PRs add new figures to the report to help users between understand and quality check results if they use external regressors or global signal regression.
• Add --dummy-scans parameter to CLIs by @tsalo in Add --dummy-scans parameter to CLIs #1232
  tedana now has a parameter --dummy-scans that can be used to remove the first X volumes of fMRI data from within tedana.

Project maintenance

• Drop Python 3.8 support by @effigies in chore: Drop Python 3.8 support #1228
  Python 3.8 is considered end-of-life and it was starting to cause some backwards compatibility issues.

Full Changelog: 25.0.0...25.0.1

---

This discussion was created from the release 25.0.1.