CSP on resting state data

✉️ Mailing List Archive (read-only)
1

External Email - Use Caution

I'd like to have a general discussion on applying common spatial patterns
on resting eeg.

I have the following questions:

Is CSP used on resting state eeg data at all? Does it have discriminative
capacity over such data as much as it does in event-based eeg data?

If CSP can be applied to resting eeg, what is the correct way to use CSP on
resting eeg data in a binary setting ? Can someone outline some pseudocode
using the mne.decomposition.csp module for this scenario. (I had posted
some of my code already on this forum. )

Thank you
Supraja J
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.nmr.mgh.harvard.edu/pipermail/mne_analysis/attachments/20190906/0e045575/attachment.html

2

External Email - Use Caution

you need 2 conditions to learn CSP filters.

what would they be on rest?

Alex

3

External Email - Use Caution

Sure. The two are two different forms of progression of a neurodegenerative
disease (milder and more aggressive) on a sample of some N patients.

Is it possible at all to use CSP successfully with this data ? I have since
tried many ways of serving this data to the CSP routine in MNE with little
success.

I have tried concatenating rawarrays like mentioned in this example:
https://mne.tools/stable/auto_examples/decoding/plot_decoding_csp_eeg.html.

Perhaps, it will be of great help if another such example is given for
rest EEG of 2 different conditions on different patients.