[Eeglablist] EOG

[Arnaud Delorme]

Timothy O'Keefe wrote:

>That seems odd to me because we are able to import bipolar VEO and HEO
>into EEGLAB (we use Digitize acquisition software whcih records into
>erpss format). They both come up as channels 63 and 64 (as they are
>recorded) AND we have imported and exported (without running ICA) and
>the output yeilds identical datasets. I don't know if that could be a
>result of how MANSCAN codes bipolar electrodes? Though Arno did say that
>"EEGLAB can not yet handle properly bipolar channels". That is confusing
>to me because we have had no problems as of yet removing blinks. 
>
The current problem is that EEGLAB does not store a reference for each 
channel. As a result, it can not make the difference between eye 
channels recorded in a bipolar montage and eye channels recorded using 
the same reference as the EEG scalp electrodes. To solve this problem, 
in the future, we intend to include a reference for each channel in the 
channel structure. This is why EEGLAB cannot handle eye channels properly.

What would happen if you run ICA on all channels (including bipolar 
channels)?
We expect that ICA might dedicate (in theory) some components to account 
for the different references. In theory, we think that ICA component 
should be able to resolve the multiple reference problem and find some 
kind of common reference.
However no study has been done on this yet.

>Does
>readses.m fix this (Arno, you should know about readses.m)?  I'm not so
>sure cuz Iv'e used the read_erpss function and that reads them in fine
>too. 
>
Neither readses() nor read_erpss() fix this problem (readses() is not an 
eeglab function; it was kindly sent to us to import compressed ERPSS 
data but, since it could not handle data events properly, we chose to 
augment our read_erpss() function to import compressed ERPSS data).

>Also, we reference everything to the nose inc. HEO and VEO.
>
We think that this is the best option if you are planning to include 
these channels in the ICA decomposition.

Arno
-- 

*Arnaud Delorme, Ph.D.*
Computational Neurobiology Lab, Salk Institute
10010 North Torrey Pines Road
La Jolla, CA 92037 USA

*Tel* : /(+1)-858-458-1927 ext 15/
*Fax* : /(+1)-858-458-1847/
*Web page *: www.sccn.ucsd.edu/~arno <http://www.sccn.ucsd.edu/%7Earno>


-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20030829/886016ee/attachment.htm