[Eeglablist] Reference to linked mastoids - Neuroscan

Andreas Widmann widmann at uni-leipzig.de
Thu Mar 23 09:40:47 PDT 2017

Hi Tara,

> I usually use a 64 channel cap with averaged mastoids as my reference.  Regrettably, I recently obtained pediatric 20 channel caps and learned afterwards that they have linked mastoids. I would return them for unlinked mastoids, but I've already lost a lot of time on this project.
> I've been unable to get EEGLAB to recognize the linked electrodes as the reference.
Not sure what you mean by getting „EEGLAB to recognize the linked electrodes as the reference“. The implicit reference (here linked mastoids) is „recognized" implicitly.

> M1 & M2 were not present in the channel locations.  I used the M1-M2 channel locations from my 64 channel file to append these channel locations in EEGLAB (added channels 21 & 22).  Then I tried to reference the mastoids in 2 ways:
> 1. From the GUI: I clicked on Re-reference data to channels, retain old reference channels in data and add current reference channel back to data.  M1-M2 were available.  However, when I added them, I received an error message, "Aborting - you must enter one or more reference channels."
> 2. I also tried entering the following from the command line:
> EEG = pop_reref (EEG, [21,22], 'keepref', 'on'); 
> I have two questions.  (1) Is there anyway to troubleshoot this problem? (2) If I can't fix it, will it be a problem that the reference will be unknown for ICA, even though there was a reference when the data was recorded?
Could you please specify more precisely what the problem is? Just some speculations:

Do you refer to the distortion of the scalp topography introduced by linking the channels (e.g. described by Luck, 2014, pp.162: „However, physically linking the wires from these two electrodes creates a zero-resistance electrical bridge between the hemispheres, which may distort the distribution of voltage over the scalp and reduce hemispheric differences. This also invalidates the use of source localization techniques.“)? This distortion unfortunately cannot be corrected as any re-referencing will only change the reference potential but never the shape of the topography.

You wil also not be able to recover the signal at M1 and/or M2 and use that for re-referencing to averaged mastoids. However, your linked mastoid data are already as close to averaged mastoid reference as you can get (given the recordings). To my understanding, you should be able to either directly compute ICA (if rank is ok; check!) or compute ICA on the average referenced data (see Makoto’s pre-processing pipeline for recommendations and caveats).

I would, however, indeed recommend to return the caps for unlinked ones for future recordings.
Hope this helps,

> Thanks,
> Tara Davis
> -- 
> Tara M. Davis, Ph.D. CCC-A
> University of South Alabama
> Associate Professor
> Department of Speech Pathology & Audiology
> _______________________________________________
> 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

More information about the eeglablist mailing list