<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=utf-8">
</head>
<body text="#000000" bgcolor="#FFFFFF">
Hello EEGLABlist,
<div class="moz-forward-container">
<p>I write to all of you with news about the ICLabel project. For
reference, the ICLabel project is my attempt at creating an
automated EEG IC classifier that is both high-accuracy and
multi-class (discerns brain from muscle, eye, heart, etc.). In
the past, I've written about the <a moz-do-not-send="true"
href="reaching.ucsd.edu:8000/tutorial">ICLabel website
(reaching.ucsd.edu:8000/tutorial)</a> having hosted a
competition based on who has submitted more EEG IC labels to the
project as well as an expansion of the website as an educational
and training tool. Now, I have some exciting new announcements
of features that had been previously requested.</p>
<p><b>If you are wondering how proficient you are at labeling EEG
ICs, you can get a rigorous rating based on a statistical
crowd-labeling model</b>. When you sign into your account and
view your ICLabel profile, there will be a colored table showing
your estimated proficiency. Note that you have to have submitted
a nominal amount of labels before you are incorporated into the
model and that the accuracy of the model's estimate of your
proficiency depends on how many labels you've submitted (more is
better). I would suggest a few hundred, which doesn't take as
long as you would imagine.</p>
<p>There are also some features still in the works, such as
labeling specific types of component of your choice as well as a
comparison of classifier performances.<br>
</p>
<p>I also want to announce that <b>we will soon be hosting </b><b>another
competition</b>! More information to come soon...</p>
<p>Sincerely,</p>
<p>Luca Pion-Tonachini<br>
Swartz Center for Computational Neuroscience<br>
and Department of Electrical Engineering at University of
California San Diego<br>
</p>
</div>
</body>
</html>