[Eeglablist] Artifact frojm EGI system

[Eric Landsness]

I am writing for advice on how to remove a large artifact from my data.

I have what I believe to be a sweating artifact with the EGI system.  It is characterized by a rapid depolarization (0.5 seconds, hundreds of microvolts) followed by a slow hyperpolarization on the order of 3-4 seconds.  I can filter it out with a highpass filter with a cutoff with 3 or 4 Hz.  However, for my analysis I am interested in the frequency range of 0.5-2.0 Hz and therefore a HPF is not feasible.  I am looking for a way to selectively remove this artifact while still maintaining the integrity of the true physiological data.

I have uploaded the EEGlab data set to my website along with a screen shot of the data.  Please note that due to the nature of the DC recording we normally apply a 1st order 0.1 high pass filter to the data before we do anything else, but I wanted to give you the data as is.  Sampling freq = 500Hz 

http://mywebspace.wisc.edu/landsness/MonsterArtifact/

This is a 10 minute data segment extracted from an 8 hour study.  

For computational power I have a 64 bit Windows machine with 16 gigs of RAM.

Are there any alternatives besides PCA/ICA such as adaptive filter?

Thanks,
Eric