[REF] Use nilearn and images instead of arrays

ME-ICA / mapca

A Python implementation of the moving average principal components analysis methods from GIFT

GNU General Public License v2.0

6 stars 8 forks source link

[REF] Use nilearn and images instead of arrays #35

Open tsalo opened 3 years ago

tsalo commented 3 years ago

Closes #31 and closes #34.

Changes proposed:

Allow utils._subsampling() to accept either an array or an image. Ultimately, we may want to limit this to just images, but I haven't done that yet.
Use nilearn's masking functions instead of internal implementations.

tsalo commented 3 years ago

This changes the number of components identified in the integration test from 57 to 74, so it's definitely impacting behavior, which is why I opened this as a draft. I will try to dig into it at some point, but if anyone notices anything, please feel free to comment.