[Eeglablist] Artifact Rejection

Arnaud Delorme arno at salk.edu
Thu Jul 28 16:29:52 PDT 2005


Edward Justin Modestino, M.Phil. wrote:

>Hello,
>In the EEGLAB tutorial on artifact rejection, it mentions the default trim
>percentage is 5%.  Are there any references for using this default of 5%? 
>Or is this based on general statistics of a 95% confidence, or something
>similar?
>  
>
This percentage correspond to the percentage of epochs the experimenter 
judge to be artifactual. If the data quality is bad, you might want to 
increase it, but 10% is usually an acceptable value for most datasets.

>If I were to perform a amplitude thresholding on a channel by channel
>basis, what amplitude in +/- microvolts would be considered an upper
>limit?
>
Yes, you may select specific channels for amplitude thresholding 
("electrode(s)"  edit box).

>Should this be dynamic and thus be a different microvolt +/-
>amplitude value cutoff for each channel? Are there any reference papers
>for this?
>  
>
Usually the same value is used for different (often eye) channels.

>I have one reference Picton et al. (2000), an article about guidelines for
>ERPs which includes recording standards and publication criteria.  In this
>article on page 138, it mentions that a criterion level of +/- 200
>microvolts could be used to remove trials.  It mentions this may be too
>Conservative for sleep studies which may have higher amplitudes.  This +/-
>200 microvolts sounds far too liberal to me.  I believe the eye movements
>I have already removed from the data were much less than that!
>
The standard value I use is 50 microvolt although I do not have a 
reference at hand. Eye movement amplitude also varies widely from 
subject to subject, from probably 30 microvolt to 300 microvolts. For 
the purpose of rejecting eye blink artifacts, I think it is better to 
adjust this value on a subject to subject basis.

Best regards,

Arno
-- 

*Arnaud Delorme, Ph.D.*
Swartz Center for Computational Neuroscience, INC, University of San 
Diego California
La Jolla, CA92093-0961, USA

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

    If I try to be like him, who will be like me?

        /Yiddish proverb/  

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20050728/b712a5bc/attachment.html>


More information about the eeglablist mailing list