TimeFrequency with machine learning

💬 Support & Discussions
, ,
1

Hello everyone,
I want to use TimeFrequency in mne.decoding for my epochs, which is in a shape of (100, 24, 1020), when apply mne.decoding.TimeFrequency to it, the return array is in the shape(100, 24, 12, 1020)
My question is how to reshape this for training purposes in machine learning, as you know the shape of y is equal to (100,) when I want to reshape to (100, 24 * 12 * 1020) the shape changes to (100, 293760).
In my opinion this is wrong , so what should I do?

Thank you so much for your help.

Regards,

2

hi @Raz

I would not do things like logistic regression on TFR outputs. A better approach to decode time-frequency / induced content is to use things made for it such as CSP mne.decoding.CSP — MNE 1.1.1 documentation

you have an example at the bottom of the page

Alex

2 Likes