[Eeglablist] Removing bad channels before average referencing - how stringent should I be?

[Makoto Miyakoshi]

Dear Jenny,

For you, I gave major update on this wiki page.
https://sccn.ucsd.edu/wiki/Makoto's_preprocessing_pipeline
See the steps 6-9.

> but I'm not sure how stringent to be.

See Nima's PREP paper (Bigdely-Shamlo et al., 2015). There is a figure that
illustrates what happens when you include bad channel to average reference.

> Clearly I would remove channels recorded from electrodes with a
consistently bad connection, but what about channels that show slight 60 Hz
interference, or channels that were poorly connected for the first 20% of
the trials, but were then fixed and look fine for the rest of the task?

If 10% of datapoint rejection cannot clean the data, give up the channel.
That's the rule of thumb. See Delorme et al. (2007) NeuroImage.

However, I recommend you use clean_rawdata(). This has ASR, artifact
subspace reconstruction, which is a very smart algorithm.
http://sccn.ucsd.edu/eeglab/plugins/ASR.pdf

Makoto

On Wed, Aug 24, 2016 at 8:35 AM, Jenny Bress <jennifer.bress at gmail.com>
wrote:

> I recently began working in a lab that uses a dense-array cap and average
> referencing. I typically work with a smaller number of electrodes and use
> mastoid referencing, so this is a bit new to me. I know I should be
> removing bad channels before re-referencing to the average in order to
> avoid noise propagating to other channels, but I'm not sure how stringent
> to be. Clearly I would remove channels recorded from electrodes with a
> consistently bad connection, but what about channels that show slight 60 Hz
> interference, or channels that were poorly connected for the first 20% of
> the trials, but were then fixed and look fine for the rest of the task? Any
> advice would be much appreciated.
>
> _______________________________________________
> Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
> To unsubscribe, send an empty email to eeglablist-unsubscribe at sccn.
> ucsd.edu
> For digest mode, send an email with the subject "set digest mime" to
> eeglablist-request at sccn.ucsd.edu
>



-- 
Makoto Miyakoshi
Swartz Center for Computational Neuroscience
Institute for Neural Computation, University of California San Diego
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20160825/3ab2f991/attachment.html>