I don't know if it improves the quality of the thesis, but it helped me move forward.

From my point of view, this is the most difficult part of science... For my thesis, I made a very detailed plan (almost to the point where each sub-section was a sentence). Then I wrote a few sub-sections every day. This allowed me to quantify my progress and have clear small goals.

Any way to save these settings so that they are loaded by default when importing matplotlib ?

Also, about the splitting, what's the rationale for using the first 15 participants versus a random sample of 15? A random split would be more robust and reduce potential biases.

It’s always a good idea to share your analysis and to track changes, but as a reviewer, I wouldn’t necessarily take it as definitive proof of a "replication" analysis. In principle, anyone could analyze the full dataset offline and simply say they didn’t by publishing the part of the code they want

docs.astral.sh/uv/

Great resources! Thanks a lot for the time and energy put into this!

+ they ask us to subscribe to mailing list in the process 🙃

On my side: Reviewer Assignment since September 13th

Nice job @cbrnr.bsky.social !

To demystify things, she doesn't control her computer with EEG (i.e. signal coming from the brain). She simply contracts her jaw muscles. These muscles create a much stronger electrical signal (EMG) than the one generated by the brain. The algorithm therefore detects only the muscle contractions.

228 healthy participants with Raw and preprocessed EEG recordings ( + MRI, behavior, phenotype)

fcon_1000.projects.nitrc.org/indi/retro/M...

MNE-BIDS ?