[Eeglablist] baseline correction of cluster ERP in STUDY

[Tom Campbell]

Hi, I ran it all again with slightly different parameters, making sure to use the prestimulus baseline for std_precomp. Maybe this is where baseline correction of components belongs. The baselining of the cluster ERPs now looks good, so problem solved, I think. Thanks to those of you who put any thought into it. Best regards, Tom. 
 From: tom_campbell75 at hotmail.com
To: eeglablist at sccn.ucsd.edu
Subject: baseline correction of cluster ERP in STUDY
Date: Wed, 22 Feb 2012 01:59:45 +0000








Dear Scott, Arno, & colleagues,
 
After plotting the dipoles of all cluster centroids in pop_clustedit, I notice there is one cluster with a large apparently additive ERP difference between conditions for that cluster in the pre-stimulus baseline. This pre-stimulus difference occurs even though I have baseline corrected at the epoching stage. I would find significant post-stimulus permutation in this circumstance unconvincing. My inclination is, for each cluster, to baseline correct in STUDY, via the command line, each of the components that contribute to that cluster centroid ERP, as I think would baseline correct the cluster centroid ERP. The place where it seems to make sense to baseline correct component-wise is for each centroid component comp in cluster y STUDY.cluster(1,y).erpdata(:,comp) and to do so for all clusters. Accordingly would it follow to baseline correct the corresponding ERSP and ITC  (using compute common ERSP baseline)?  I do not know what this would do to dipole locations. Please, is there a better approach to this baseline issue or a theoretical-analytical reason why such baseline correction is not in the GUI in STUDY?
 
I notice there is the facility to baseline correct cluster ERPs in std_envtopo.m.
 
Best regards,
Tom.

  		 	   		   		 	   		  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20120224/84766e66/attachment.html>