Nilearn 0.12.0 is out

Remi-Gau · June 23, 2025, 7:32am

Hello everyone

We have just released Nilearn 0.12.0!

This is a major release with some exciting new features:

Better handling of surface data via the functions from the nilearn.image and nilearn.masking packages (mean_img,math_img, apply_mask…)
Add support for “transparent thresholding” plotting that not only highlights statistically significant regions but also includes subthreshold locations. See this new plotting example to know how to use it.
Surface plots can now plot both hemispheres. This is demonstrated in this example: Making a surface plot of a 3D statistical map - Nilearn
Image comparison can now be done on both volume and surface data and can also be done via Bland-Altman. Used in this example: First level analysis of a complete BIDS dataset from openneuro - Nilearn
The default colormap has been changed to “RdBu_r” or “gray” for most plotting functions.
Label maskers can now accept a BIDS formatted ‘look up tables’ to more easily map an atlas region name, its index and its color.
More uniform look for all nilearn masker and GLM reports. Improved user interface of the GLM reports (GLM reports, masker reports)
Fixes the behavior of nilearn.decoding.Decoder where it used to set the score to zero if all coefficients are zero, irrespective of the scoring metric used.
Improve compliance of nilearn maskers, GLM and decoder estimators with scikit-learn estimators guidelines.

Update from PyPi

pip install --upgrade nilearn

Thanks to our new contributors !!!

bthirion · June 23, 2025, 9:45pm

Thx Rémi for the coordination work !

sdiebolt · June 25, 2025, 12:14pm

Thanks for the awesome work! Although not targeted, the functional ultrasound imaging (fUSI) field benefits a lot from many of Nilearn’s features!

Remi-Gau · June 25, 2025, 12:32pm

glad to hear: let us know if there are things we can do to help more that could be within our scope.