Hi everyone,
I just ran ICA on my EEG data acquired through BioSemi. I’ve preprocessed it with a (separate ICA preprocessing pipeline) 1-30 Hz filter, epoched -0.2 to 0.4 and referenced to mastoids (EXG1 and EXG2). These are the components I get after running ICA:
001 and 003 look like standard EOG activity to me, but I have no idea what 000 means. Does anyone know what that component means?
Thanks,
Nathan
how did you reference your data?
it seems you have constant latent source
Alex
Hi Alex,
Do you mean how I did it in MNE? If so, I used the “set_eeg_reference” function":
raw_motoraud.set_eeg_reference(ref_channels=[‘EXG1’, ‘EXG2’])
That was the line that I specified.
During the experiment, the participant wore two external electrodes at the left/right mastoids (EXG1/EXG2 respectively). They also wore headphones over the cap, if that is relevant (although if that was the case, I guess I would expect to see the same ICA component in other eeg sessions as well, and I don’t).
Thanks!
The topomap alone does not contain enough information to interpret its origin. Can you also show the time course as well as an epochs image and the PSD via ica.plot_properties(raw)? It could be ECG, but it’s unusual that it is the first component. It could also be related to your reference, I’d be curious to see if anything changes for average referenced data.
Oh yes, sorry, I should have included that before. I’ll include it now:
This is the plot_properties for the mastoid referenced data. I’ve also included the ICA topos and plot_properties for a nose-referenced set.
It looks slightly different (it seems to have slightly lower power in the higher frequencies for the PSD), but I’m still not sure what it means.
Thanks,
Nathan
Can you show the time course (for example the first 20s)?
