<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=utf-8">
</head>
<body bgcolor="#FFFFFF" text="#000000">
Dear EEGLAB mailing list subscribers,<br>
<br>
I would like to ask for your help on a project that will likely be
of use to you in your research or any other endeavor involving ICA
decomposition of EEG data. I am an Electrical Engineering Ph.D.
student at UCSD and I work at the Swartz Center for Computational
Neuroscience, where EEGLAB is developed.<br>
<br>
Here at the SCCN, we have gathered a lot of ICA decomposed EEG data.
My plan is to use it to create a multi-class EEG independent
component (IC) classifier, one you can run and trust the results as
much as if a domain expert had personally labeled them for you. The
secret to making such a classifier accurate is in the data used to
train it. Including results from many different experiments,
electrode montages, subjects, and EEG devices — as well as tentative
classifications by many judges — should allow a level of
generalization that makes the resulting classifier usable on any
dataset. A lot of people I’ve spoken with are excited by this
prospect, as I hope you are too.<br>
<br>
If you are interested in participating in this project, please
browse <a href="reaching.ucsd.edu:8000/tutorial">this website</a> (
<a class="moz-txt-link-freetext" href="http://reaching.ucsd.edu:8000">http://reaching.ucsd.edu:8000</a> ) and suggest classifications for as
many ICs as you have time for. Each IC is represented by a figure
showing several measures (scalp map, equivalent dipole location,
mean spectrum, erpimage, etc.). There is a tutorial on the site that
will tell you more about the project and how to use it — and also a
guide to discriminating several types of EEG IC processes. <br>
<br>
If you have comments, questions, or suggestions, please give me
feedback through the website or at this email address. We will make
use of all data entries, regardless of how many get labeled — and
each IC you label will make the classifier more useful.<br>
<br>
Thank you,<br>
Luca Pion-Tonachini
</body>
</html>