nilearn
nilearn copied to clipboard
nilearn
Machine learning for NeuroImaging in Python
This PR is designed as a first pass at solving performance issues as noted in [#3266](https://github.com/nilearn/nilearn/issues/3266). This PR is rather minimal in scope in that it only changes the internal...
Closes #3334 Changes proposed in this pull request: - Add conditional so that binder knows to look for the notebooks in `/dev` or `/stable` - Binder now uses jupyterlab instead...
When trying to open binder I get a 404 error. This is probably due to the new file structure and cleaning of https://github.com/nilearn/nilearn.github.io in https://github.com/nilearn/nilearn.github.io/pull/3. I am also curious about...
Closes #3321. Changes proposed in this pull request: - Replace rendered URLs with sphinx internal referencing - If external referencing is needed, use intersphinx mapping or external links (in `doc/conf.py`)
Working on #3315 made me think that there might be a better way to handle URLs. For example, some links repeat many times making it hard to manage them. I...
Closes #3159. Instead of adding a new example that uses `permuted_ols` with a one-sample design, I added more documentation about designs and linked to an existing one-sample example using `non_parametric_inference`...
Closes #2726, closes #3132, and closes #2724. Changes proposed in this pull request: - Fix docstrings of masker classes' `inverse_transform` methods to make it clear that 1D inputs are allowed,...
Closes #3303. Changes proposed in this pull request: - Add stable directory to fix broken links
See https://github.com/nilearn/nilearn/pull/3300#discussion_r915643466 and https://github.com/nilearn/nilearn/pull/3327#discussion_r939007061 Relevant steps: - Move existing fixtures to one file (`conftest.py`?) - Create new fixtures where generated data is reused - Better naming to improve reusability
The TFCE method we implemented in #3196 is very slow. The probabilistic TFCE measure implemented in MNE seems to be faster, although I haven't directly compared them. We should look...