[Eeglablist] reference one more time

Philip Michael Zeman pzeman at alumni.uvic.ca
Mon Jun 6 17:45:33 PDT 2005


Rita-

Check to see if when you manually invert EEG.icawinv (the inverse of the weight matrix) that you get or don't get any singularity warnings.

As you said, there is a dependency there.

I personally would prefer to leave the A1 A2 references in there for localization purposes however then there is this trade-off you speak of.  It also seems wrong to just remove one of the reference channels (A2) and then look at the spatial maps because we lose electrode symmetry.  Maybe the solution is to use A1 and A2 for data at acquisition time and use the inion or the naison as the reference.  Comments??


Phil Zeman 


  ----- Original Message ----- 
  From: Rita Ceponiene 
  To: eeglablist at sccn.ucsd.edu 
  Sent: Thursday, June 02, 2005 3:43 PM
  Subject: [Eeglablist] reference one more time


  I wonder whether it is better to use a single-mastoid referenced data or average-mastoid referenced data when training ICA on continuous data? When one mastoid is used for the reference, another mastoid is typically left in the dataset - can this induce a bias in the component configuration, especially over the back of the head? On the other hand, when average of both mastoids is used, both channels have data in them, which is, however, 100% correlated and smaller in magnitude than data on other channels. Will this affect ICA or not (since the manipulation was linear)? If it will, should both mastoid channels be deleted from the data?

  Thank you-

  Rita Ceponiene    



------------------------------------------------------------------------------


  _______________________________________________
  eeglablist mailing list eeglablist at sccn.ucsd.edu
  http://sccn.ucsd.edu/mailman/listinfo/eeglablist
  Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
  To unsubscribe, send an empty email to eeglablist-unsubscribe at sccn.ucsd.edu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20050606/415cf0c1/attachment.htm


More information about the eeglablist mailing list