ICA mne python - combining mags and grads

✉️ Mailing List Archive (read-only)
1

Hello everyone,

I have an ICA-related question for the mne-python community. This time
more theoretically oriented.

All examples and tutorials on the mne python pages for ICA artefact
rejection procedures combine both magnetometers and gradiometers
together when performing ICA and finding artefact components. What would
be an argument for combining all sensors like this as opposed to doing
ICA on grads and mags separately (which seems to be a recommended way in
other packages)? Is this simply for saving computational time?

Thanks for your advice!

Ana

2

hi Ana,

yes we combine grads and mags but we put them on the same scale
by whitening before feeding them to the ICA function. So AFAIK this
works quite well.

HTH
Alex

3

Hi Ana,

it is indeed mainly for simplifying things and reducing the number of
solutions and files to handle. But nothing speaks against fitting separate
ICA solutions.

Denis