<html xmlns:v="urn:schemas-microsoft-com:vml" xmlns:o="urn:schemas-microsoft-com:office:office" xmlns:w="urn:schemas-microsoft-com:office:word" xmlns:m="http://schemas.microsoft.com/office/2004/12/omml" xmlns="http://www.w3.org/TR/REC-html40">
<head>
<meta http-equiv=Content-Type content="text/html; charset=utf-8">
<meta name=Generator content="Microsoft Word 12 (filtered medium)">
<style>
<!--
/* Font Definitions */
@font-face
{font-family:Calibri;
panose-1:2 15 5 2 2 2 4 3 2 4;}
@font-face
{font-family:Tahoma;
panose-1:2 11 6 4 3 5 4 4 2 4;}
@font-face
{font-family:-webkit-sans-serif;
panose-1:0 0 0 0 0 0 0 0 0 0;}
/* Style Definitions */
p.MsoNormal, li.MsoNormal, div.MsoNormal
{margin:0cm;
margin-bottom:.0001pt;
font-size:12.0pt;
font-family:"Times New Roman","serif";}
a:link, span.MsoHyperlink
{mso-style-priority:99;
color:blue;
text-decoration:underline;}
a:visited, span.MsoHyperlinkFollowed
{mso-style-priority:99;
color:purple;
text-decoration:underline;}
p
{mso-style-priority:99;
mso-margin-top-alt:auto;
margin-right:0cm;
mso-margin-bottom-alt:auto;
margin-left:0cm;
font-size:12.0pt;
font-family:"Times New Roman","serif";}
span.EmailStyle18
{mso-style-type:personal-reply;
font-family:"Calibri","sans-serif";
color:#1F497D;}
span.apple-style-span
{mso-style-name:apple-style-span;}
.MsoChpDefault
{mso-style-type:export-only;}
@page Section1
{size:612.0pt 792.0pt;
margin:72.0pt 72.0pt 72.0pt 72.0pt;}
div.Section1
{page:Section1;}
-->
</style>
<!--[if gte mso 9]><xml>
<o:shapedefaults v:ext="edit" spidmax="1026" />
</xml><![endif]--><!--[if gte mso 9]><xml>
<o:shapelayout v:ext="edit">
<o:idmap v:ext="edit" data="1" />
</o:shapelayout></xml><![endif]-->
</head>
<body lang=EN-US link=blue vlink=purple>
<div class=Section1>
<p class=MsoNormal><span style='font-size:11.0pt;font-family:"Calibri","sans-serif";
color:#1F497D'>Dear Manousos,<o:p></o:p></span></p>
<p class=MsoNormal><span style='font-size:11.0pt;font-family:"Calibri","sans-serif";
color:#1F497D'><o:p> </o:p></span></p>
<p class=MsoNormal><span style='font-size:11.0pt;font-family:"Calibri","sans-serif";
color:#1F497D'><o:p> </o:p></span></p>
<p class=MsoNormal><span style='font-size:11.0pt;font-family:"Calibri","sans-serif";
color:#1F497D'>FastICA is known to be sensitive to its initialization (the
weighting matrix is randomly initialized, or specified by the user …). When a
non-convergence is met, it simply starts off with a reinitialisation of the
weighting vector for that component. This can easily be explained by the use of
Newton-like gradient algorithms that are optimal for local minima, though not
for global optimization. It is a sort of simplified Monte Carlo estimation. The
best is to run fastICA multiple times and to cluster the results to obtain its
consistent components (see the ICASSO package: </span><a
href="http://www.cis.hut.fi/projects/ica/icasso/">http://www.cis.hut.fi/projects/ica/icasso/</a><span
style='font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D'>),
however you will use the relatively ‘fast’ property of fastICA.<o:p></o:p></span></p>
<p class=MsoNormal><span style='font-size:11.0pt;font-family:"Calibri","sans-serif";
color:#1F497D'><o:p> </o:p></span></p>
<p class=MsoNormal><span style='font-size:11.0pt;font-family:"Calibri","sans-serif";
color:#1F497D'>Personally, I think that the problem of convergence is rather
due to the real data dimensionality (the number of underlying sources) than to
the component’s strength (if with strength, you mean variance), since ICA is
blind to source variance (it has unit variance sources as a solution, due to
the ambiguity). The dimensionality of biomedical datasets is a huge topic of
research, and I know that e.g. Dr. C. Hesse (</span><span
class=apple-style-span><i><span style='font-size:10.0pt;font-family:"-webkit-sans-serif","serif";
color:black'><a
href="https://embs.papercept.net/conferences/scripts/abstract.pl?ConfID=8&Number=1763"
target="_blank">Model Order Estimation for Blind Source Separation of
Multichannel Magnetoencephalogram and Electroencephalogram Signals (I)</a></span></i></span><span
style='font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D'>) or
prof. A. Cichocki (a.o. </span><a
href="http://www.bsp.brain.riken.jp/publications/1997/Cich-Sabala-CSCOSnolta97.pdf">http://www.bsp.brain.riken.jp/publications/1997/Cich-Sabala-CSCOSnolta97.pdf</a>)
<span style='font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D'>have
made some attempts for the estimation of the data dimensionality. <o:p></o:p></span></p>
<p class=MsoNormal><span style='font-size:11.0pt;font-family:"Calibri","sans-serif";
color:#1F497D'><o:p> </o:p></span></p>
<p class=MsoNormal><span style='font-size:11.0pt;font-family:"Calibri","sans-serif";
color:#1F497D'>Concerning the “source strength” in a different sense, note that
<o:p></o:p></span></p>
<p class=MsoNormal><span style='font-size:11.0pt;font-family:"Calibri","sans-serif";
color:#1F497D'>1/ the strength in the objective function (the kurtosis value of
the source if one uses the ‘pow3’ option) plays a crucial role and thus near
Gaussian sources will be harder to recuperate than highly kurtic sources<o:p></o:p></span></p>
<p class=MsoNormal><span style='font-size:11.0pt;font-family:"Calibri","sans-serif";
color:#1F497D'>2/ if you use the ‘tanh’ nonlinearity in fastICA, than it is
only optimal for superGaussian sources; the function is thus sub-optimal for periodic
triangular/sinusoidal/… or near-uniform sources! See also the extension from
infomax to extended-infomax by Te-Won Lee.<o:p></o:p></span></p>
<p class=MsoNormal><span style='font-size:11.0pt;font-family:"Calibri","sans-serif";
color:#1F497D'><o:p> </o:p></span></p>
<p class=MsoNormal><span style='font-size:11.0pt;font-family:"Calibri","sans-serif";
color:#1F497D'>There also exist some alternatives to the FastICA algorithm (note
that it is only algorithmic, the theory remains the same), and I can advise you
the RobustICA package (</span><a
href="http://www.i3s.unice.fr/~zarzoso/robustica.html">http://www.i3s.unice.fr/~zarzoso/robustica.html</a><span
style='font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D'>).
This package has no initialization issues; it is faster, thanks to the
calculation of the optimal step-size and it converges naturally with an optimal
“fixed” point. Thus at least the issues of paragraph 1 in this mail could be
tested.<o:p></o:p></span></p>
<p class=MsoNormal><span style='font-size:11.0pt;font-family:"Calibri","sans-serif";
color:#1F497D'><o:p> </o:p></span></p>
<p class=MsoNormal><span style='font-size:11.0pt;font-family:"Calibri","sans-serif";
color:#1F497D'><o:p> </o:p></span></p>
<p class=MsoNormal><span style='font-size:11.0pt;font-family:"Calibri","sans-serif";
color:#1F497D'>Hope this helps,<o:p></o:p></span></p>
<p class=MsoNormal><span style='font-size:11.0pt;font-family:"Calibri","sans-serif";
color:#1F497D'><o:p> </o:p></span></p>
<p class=MsoNormal><span style='font-size:11.0pt;font-family:"Calibri","sans-serif";
color:#1F497D'><o:p> </o:p></span></p>
<p class=MsoNormal><span style='font-size:11.0pt;font-family:"Calibri","sans-serif";
color:#1F497D'>Ronald <o:p></o:p></span></p>
<p class=MsoNormal><span style='font-size:11.0pt;font-family:"Calibri","sans-serif";
color:#1F497D'><o:p> </o:p></span></p>
<p class=MsoNormal><span style='font-size:11.0pt;font-family:"Calibri","sans-serif";
color:#1F497D'><o:p> </o:p></span></p>
<div style='border:none;border-top:solid #B5C4DF 1.0pt;padding:3.0pt 0cm 0cm 0cm'>
<p class=MsoNormal><b><span style='font-size:10.0pt;font-family:"Tahoma","sans-serif"'>From:</span></b><span
style='font-size:10.0pt;font-family:"Tahoma","sans-serif"'>
eeglablist-bounces@sccn.ucsd.edu [mailto:eeglablist-bounces@sccn.ucsd.edu] <b>On
Behalf Of </b>Klados Manousos<br>
<b>Sent:</b> mercredi 7 janvier 2009 14:11<br>
<b>To:</b> eeglablist@sccn.ucsd.edu<br>
<b>Subject:</b> [Eeglablist] FastICA failure in converge<o:p></o:p></span></p>
</div>
<p class=MsoNormal><o:p> </o:p></p>
<p class=MsoNormal style='margin-bottom:12.0pt'>Dear EEGLAB users,<br>
<br>
I use FastICA v. 2.5 in EEG signals. My problem is that fastica does not
converge in 28 of my 54 datasets. When fastica stucks in one component, it
tries to converge five times (making 1000 iteration each time) and finaly it
appears me the following message. <br>
<b><br>
Component number ## did not converge in 1000 iterations.<br>
Too many failures to converge (6). Giving up.<br>
Adding the mean back to the data.</b><br>
<br>
I raised the number of interations to 10000 but then the algorithm becomes very
slow (so it loses it's fast ability) and in some datasets doesn't converge
again...<br>
<br>
I would like to ask you why is that happening? <br>
<br>
I read that <b>"The fastICA method failed completely to find the target
sources, since the Newton iteration always converged to one of the much
stronger confound sources."</b> and that <b>"Even for very low
signal strengths, where the fastICA method fails to converge to the target
sources"</b><a href="http://cds.ismrm.org/ismrm-2004/Files/000493.pdf">
here</a>. EEG signals have very low strength...so that is the answer to my
previous question or it's something more than EEG signal's strength?<br>
<br>
Also i noticed that in the same dataset, fastica can converge ones and does not
converge other time...Is it that logical?<br>
<br>
Thank you in advnace<br>
Have a happy new year with my best wishes<br>
<br>
Manousos<br>
<br clear=all>
<br>
-- <br>
Klados A. Manousos<br>
Graduate Student, Research Assistant<br>
Group of Applied Neurosciences<br>
Lab of Medical Informatics, Medical School<br>
Aristotle University of Thessaloniki<br>
Thessaloniki, Greece<br>
_________________________________________________<br>
Tel: +30-2310-999332<br>
Website :<br>
<a href="http://lomiweb.med.auth.gr/gan/index_en_files/Page609.htm">http://lomiweb.med.auth.gr/gan/index_en_files/Page609.htm</a><br>
Blog: <a href="http://appliedneurosciences.blogspot.com">http://appliedneurosciences.blogspot.com</a><o:p></o:p></p>
<p><span style='font-size:10.0pt;font-family:"Arial","sans-serif"'>No virus
found in this incoming message.<br>
Checked by AVG - http://www.avg.com<br>
Version: 8.0.176 / Virus Database: 270.10.5/1883 - Release Date: 1/8/2009 6:05
PM</span><o:p></o:p></p>
</div>
</body>
</html>