How to compute Weight Projection after decoding

đź’¬ Support & Discussions
1

Hi everyone,
I want to compare the weights between conditons after decoding, now I know the theory and background behind this problem (A=w*Cov(X)), but I don’t know how to compute in MNE_python. Anyone know this?

2

Hello @jkb521 and welcome to the forum!

Please take a look at the following example and tutorial:

https://mne.tools/stable/auto_examples/decoding/linear_model_patterns.html#decoding-in-sensor-space-using-a-logisticregression-classifier

https://mne.tools/stable/auto_tutorials/machine-learning/50_decoding.html#temporal-decoding

Best wishes,
Richard

3

Thanks, Richard,
Now I can use the get_coef to get the patterns of the trained model, but in My situation is that I trained on one dataset, and then used this model to test on another datasets, that means I should also get the patterns of other datasets and compare them, how can I get that? If I can only get the pattern of trained data, I can not compare them between conditions.
Thanks again