[Eeglablist] Using ICA to remove skin potentials?

[Scott Makeig]

Steve -

In our experience, using ICA to model data with large <<1 Hz features gives
less physiologically interpretable ('dipolar') components. I assume this is
because typical sources of near-DC potential changes (including 'the skin
potentials' you refer to) may not be spatially stationary, but may instead
e.g. flow across the scalp. ICA must model such phenomena using a number of
(or 'subspace of') components, each representing as it were an 'overlapping
movie frame' view of the phenomenon's scalp map.

Perhaps it would be of interest to high pass filter the data (say > 1 Hz),
then decompose the high pass and complementary low pass data separately by
ICA. How do the component maps relate to each other?  (Using matcorr() to
examine this).

Scott Makeig

On Wed, Feb 22, 2012 at 7:27 PM, Stephen Politzer-Ahles <
politzerahless at gmail.com> wrote:

> Hello,
>
> I was wondering if anyone knows about (or could point me to references
> about) whether it's feasible and valid to use ICA to remove low-frequency
> artifact like skin potentials. I have a set of recordings that has lots of
> problems with these, and while I've found that a 0.5 Hz high-pass filter
> can attenuate them somewhat, I'm also aware that high-pass filtering can
> have some big consequences. In the past I have used ICA to remove EOG
> artifact and I know of others who have used it for things like alpha,
> heartbeat, and muscle-related artifacts, so I'm curious about whether I can
> use it to remove skin potentials and what kinds of distortions it might
> create. But before I do that I wanted to see if others use it for that, or
> if it sounds crazy.
>
> Thank you,
> Steve Politzer-Ahles
>
> --
> Stephen Politzer-Ahles
> University of Kansas
> Linguistics Department
> http://www.linguistics.ku.edu/
>
> _______________________________________________
> Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
> To unsubscribe, send an empty email to
> eeglablist-unsubscribe at sccn.ucsd.edu
> For digest mode, send an email with the subject "set digest mime" to
> eeglablist-request at sccn.ucsd.edu
>



-- 
Scott Makeig, Research Scientist and Director, Swartz Center for
Computational Neuroscience, Institute for Neural Computation; Prof. of
Neurosciences (Adj.), University of California San Diego, La Jolla CA
92093-0559, http://sccn.ucsd.edu/~scott
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20120223/ba831db8/attachment.html>