The data_length/channels^2 = k > 30 heuristic was based on observing our 256-channel data decompositions. For fewer channels, I don't think this (k > 30) may be necessary (though a small k will very likely prove a problem!). <div>
<br></div><div>But for small numbers of channels, data length is not so much an obstacle -- for example, decomposing 32 channels with k=30 would require only 30*32^2= 32K data points: At a sampling rate of (say) 256 Hz, this would be ~2 minutes of data. With 128 channels, the equivalent heuristic would be ~32 minutes.<div>
<br></div><div>I hope one of us will be able to do a proper study, decomposing data subsets of various lengths and measuring mutual information reduction and dipolarity (see Delorme et al., PLoS One 2012). Also, Jason Palmer has been working on theoretic lower bounds on ICA accuracy at given data lengths and channel numbers. I'll summarize result with him when possible.</div>
<div><br></div><div>Scott Makeig<br><br><div class="gmail_quote">On Sat, Mar 17, 2012 at 5:17 PM, Tarik S Bel-Bahar <span dir="ltr"><<a href="mailto:tarikbelbahar@gmail.com">tarikbelbahar@gmail.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div>1. not sure if you have to decimate, have seen some people to do this to better meet requirements for good ICA decompositions.</div>
<div>2. don't think higher sampling rate will give you better ICA. I think it's more of a matter that ICA is fed data that gives an accurate and lengthy representation of the whole "Data space".</div>
<div>3. quote from earlier eeglablist post: "<span>the key factor is how much data you have (</span><span>timepoints</span><span> / </span><span>channels</span><span>^2). If this is > 30 (or near to it), then we find it preferable to return all possible components (since pca does a rather poor job of separating sources)."</span></div>
<div><span>So if below this threshold there is some reason for adding more time to the protocol, or reducing channels, or decimation</span></div><br>
<br><br><div class="gmail_quote"><div><div class="h5">On Tue, Mar 13, 2012 at 6:00 AM, Modestino, Edward J *HS <span dir="ltr"><<a href="mailto:EJM9F@hscmail.mcc.virginia.edu" target="_blank">EJM9F@hscmail.mcc.virginia.edu</a>></span> wrote:<br>
</div></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div><div class="h5">
<div lang="EN-US" link="blue" vlink="purple"><div><p class="MsoNormal"><span style="font-family:"Times New Roman","serif";color:#1f497d">Dear EEGLAB experts,<u></u><u></u></span></p><p class="MsoNormal">
<span style="font-family:"Times New Roman","serif";color:#1f497d"><u></u> <u></u></span></p><p class="MsoNormal"><span style="font-family:"Times New Roman","serif";color:#1f497d">(1) Is it true that ICA must be subject, like all the model-based spectral analysis methods, to a recommendation that one decimate to the lowest frequency capable of representing the actual signal content of the data without alienation effects? <b>Does one <u>NEED</u> to decimate the data before running ICA?</b> For instance, we have a data set recorded at 1,000 Hz. Do we need to decimate this to approximately 128 or 256?<u></u><u></u></span></p>
<p class="MsoNormal"><span style="font-family:"Times New Roman","serif";color:#1f497d"><u></u> <u></u></span></p><p class="MsoNormal"><span style="font-family:"Times New Roman","serif";color:#1f497d">(2 ) According to the formula that Dr. Makeig gave to determine the optimal amount of data, <b>#timepoints/(#channels)^2</b>, it would appear that a <b>higher sampling rate will give better ICA results</b>. Is this the case?<u></u><u></u></span></p>
<p class="MsoNormal"><span style="font-family:"Times New Roman","serif";color:#1f497d"><u></u> <u></u></span></p><p class="MsoNormal"><span style="font-family:"Times New Roman","serif";color:#1f497d">(3) Finally, using this formula, <b>#timepoints/(#channels)^2</b>, is there a <b>threshold or cutoff</b> one needs to be exceeded to have the optimal amount of data to run ICA. Simply doing the equation without any way to interpret the output is not helpful.<u></u><u></u></span></p>
<p class="MsoNormal"><span style="font-family:"Times New Roman","serif";color:#1f497d"><u></u> <u></u></span></p><p class="MsoNormal"><span style="font-family:"Times New Roman","serif";color:#1f497d">Thanks for your help,<u></u><u></u></span></p>
<p class="MsoNormal"><span style="font-family:"Times New Roman","serif";color:#1f497d">Dr. Modestino<u></u><u></u></span></p><p class="MsoNormal"><span style="font-family:"Times New Roman","serif";color:#1f497d"><u></u> <u></u></span></p>
<p class="MsoNormal"><span style="font-family:"Times New Roman","serif";color:#1f497d">Edward Justin Modestino, Ph.D.<u></u><u></u></span></p><p class="MsoNormal"><span style="font-family:"Times New Roman","serif";color:#1f497d">Postdoctoral Research Associate<u></u><u></u></span></p>
<p class="MsoNormal"><span style="font-family:"Times New Roman","serif";color:#1f497d">Ray Westphal Neuroimaging Laboratory<u></u><u></u></span></p><p class="MsoNormal"><span style="font-family:"Times New Roman","serif";color:#1f497d">Division of Perceptual Studies<u></u><u></u></span></p>
<p class="MsoNormal"><span style="font-family:"Times New Roman","serif";color:#1f497d">Department of Psychiatry and Neurobehavioral Sciences <u></u><u></u></span></p><p class="MsoNormal"><span style="font-family:"Times New Roman","serif";color:#1f497d">University of Virginia<u></u><u></u></span></p>
<p class="MsoNormal"><span style="font-family:"Times New Roman","serif";color:#1f497d">Email: <a href="mailto:ejm9f@virginia.edu" target="_blank">ejm9f@virginia.edu</a><u></u><u></u></span></p><p class="MsoNormal">
<span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1f497d"><u></u> <u></u></span></p><p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1f497d"><u></u> <u></u></span></p>
</div></div><br></div></div>_______________________________________________<br>
Eeglablist page: <a href="http://sccn.ucsd.edu/eeglab/eeglabmail.html" target="_blank">http://sccn.ucsd.edu/eeglab/eeglabmail.html</a><br>
To unsubscribe, send an empty email to <a href="mailto:eeglablist-unsubscribe@sccn.ucsd.edu" target="_blank">eeglablist-unsubscribe@sccn.ucsd.edu</a><br>
For digest mode, send an email with the subject "set digest mime" to <a href="mailto:eeglablist-request@sccn.ucsd.edu" target="_blank">eeglablist-request@sccn.ucsd.edu</a><br></blockquote></div><br>
<br>_______________________________________________<br>
Eeglablist page: <a href="http://sccn.ucsd.edu/eeglab/eeglabmail.html" target="_blank">http://sccn.ucsd.edu/eeglab/eeglabmail.html</a><br>
To unsubscribe, send an empty email to <a href="mailto:eeglablist-unsubscribe@sccn.ucsd.edu">eeglablist-unsubscribe@sccn.ucsd.edu</a><br>
For digest mode, send an email with the subject "set digest mime" to <a href="mailto:eeglablist-request@sccn.ucsd.edu">eeglablist-request@sccn.ucsd.edu</a><br></blockquote></div><br><br clear="all"><div><br></div>
-- <br>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, <a href="http://sccn.ucsd.edu/%7Escott" target="_blank">http://sccn.ucsd.edu/~scott</a><br>
</div></div>