[Mne_analysis] SSP for EEG

[Marina Papoutsi]

Dear Matti and MNE users

Recently, it was suggested to me that I could apply SSP to EEG data as a 
way of removing some of the physiological etc noise. I wanted to check 
whether my understanding is correct and ask some questions about how to 
do this.

My limited understanding of this is that I need to create an SSP 
projection and then apply that to the raw data. In the MNE manual it 
mentions to create the SSP projection from empty room recordings, but I 
assume that, in the case of EEG, recordings from a resting recording 
could do. I have about 3min of rest with eyes open and 3min of rest with 
eyes closed before the onset of the experiment. Would 3min be enough?

At the moment, the only way that I can see to create this projection is 
through mne_browse_raw. Is there any other function that I could use?

Also, if and when I do manage to generate an SSP projection, I assume it 
would be include all sensors (GRDS, MAGS and EEG), so then I could use 
mne_project_raw to apply the projection, which again I assume will be 
including the EEG.

Lastly, after the projection has been applied to the data, I assume that 
the data have been fully transformed and, if I then need to export them 
(in sensor space) into matlab, the transformation would be preserved.

Does that sound ok? Please let me know if there is anything I am missing 
and if you have any other advice, word of caution etc. I just wanted to 
check whether my understanding is correct.

Many thanks,
Marina