[FieldTrip] Artifact rejection in realtime analysis

[paymando- morientes]

Dear field trippers
I am trying to design my first real time loop for an EEG experiment. The question that I have now is that how should I deal with artifacts such as eye blinks.
Firstly, I think rejecting data segments in real time analysis is pointless because if an epoch is artifactual and can not represent the classified features,  it could simply get the label (epoch rejected) in the classification section and the script then moves to the next segment. 
Secondly, ICA is too slow to be implemented in an online loop.
So how should artifacts be dealt with inside a real time analysis? Are there any ways for correcting eye blinks other than ICA? 
Can you give me any suggestions?

THANK YOU ALL!
payman
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20140731/d36103c7/attachment.html>