MEG eye blinks

✉️ Mailing List Archive (read-only)
1

Hello,

We are using the example script plot_find_eog_artifacts.py to find the
blinks in MEG data and we have been trying to use mne.write_events to save
the blink markers into the original fif. Part of the reason is that these
are actually resting data so we wanted to just identify the blinks. If we
want to use 4 sec epochs to estimate the broadband power spectrum in
auditory cortex, what would you recommend we do stepwise? If we epoch the
data first would this fit more naturally with the way blinks are currently
identified and rejected?

::vincent
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.nmr.mgh.harvard.edu/pipermail/mne_analysis/attachments/20121211/95c1791e/attachment.html

2

hi vincent,

I would not add the events in the raw fif file as it means duplicating
the data on disk
and you need these only to compute an SSP projection.

A simple approach is to use 4s epochs with regularly spaced events and reject
4s windows that violated epochs rejection parameters.

You have a function for this:

https://github.com/mne-tools/mne-python/blob/master/mne/event.py#L262

Once you have epochs you can compute PSD in the source space and more. See
latest examples :

https://github.com/mne-tools/mne-python/tree/master/examples/time_frequency

hope this helps
Alex