Shall I use tsfresh for EEG data

[blrk]

Hi, I am working in EEG data. I am extracting the features using python code. Does tsfresh supports feature extraction in EEG?

[nils-braun]

Hi! I think there are two dimensions to your question:

• Does tsfresh technically work with EEG data? Yes. As long as you have the time series data in a format applicable to tsfresh, it will work.
• Are the extracted features meaningful for EEG data? This depends more on your target, than on your data. In principle I would propose to try it out :-) (maybe on a smaller subset of your data is too large). The feature extraction methods in tsfresh come from various domains - I am very positive that some of them will also be useful for your use case.

Question to the community: does someone have a nice use-case with EEG data to show?

PS: it might also be a good idea to have a look into papers citing tsfresh. We have over 150 papers that used tsfresh, maybe there is one that has a similar use case?

[blrk]

Awesome, Ok, I will try it out. It's very nice to know that 150 papers used tsfresh.

[nils-braun]

Thanks for trying it out!
I did find this: https://gist.github.com/F-A/ebd0cf72fb4e8b43d6d278db776f5824

And this: https://www.researchgate.net/publication/347827761_Detection_of_an_Autism_EEG_Signature_From_Only_Two_EEG_Channels_Through_Features_Extraction_and_Advanced_Machine_Learning_Analysis/references

Maybe it helps!

[blrk]

Thank you so much