<div dir="ltr">Marco,<div>Thanks for your response, that makes sense to me!</div><div>I've tested ADJUST with several sample datasets, and it appears to be working well on our data.<br><div><br></div><div><div style="font-family:arial,sans-serif;font-size:12.800000190734863px">
Brian Erickson</div><div style="font-family:arial,sans-serif;font-size:12.800000190734863px">Ph.D. Candidate</div><div style="font-family:arial,sans-serif;font-size:12.800000190734863px">Applied Brain and Cognitive Sciences Program</div>
<div style="font-family:arial,sans-serif;font-size:12.800000190734863px">Drexel University, College of Arts and Sciences, Department of Psychology</div></div></div></div><div class="gmail_extra"><br><br><div class="gmail_quote">
On Fri, Jan 17, 2014 at 4:39 AM, Marco Buiatti <span dir="ltr"><<a href="mailto:marco.buiatti@gmail.com" target="_blank">marco.buiatti@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Dear Brian,<br>
<br>
while ICs low in the list might represent just noise, they might also<br>
be associated with low-variance artifacts like transitory jumps in one<br>
specific electrode, and these are the ones that ADJUST typically<br>
captures at the bottom of the list.<br>
<br>
In any case, since they represent low portions of the overall<br>
variance, I do not think that if they are just noise, eliminating them<br>
would be a problem, at least for a standard ERP analysis.<br>
<br>
Do however check that ADJUST works properly on your data before<br>
blindly apply it.<br>
<br>
Best,<br>
<br>
Marco (ADJUST author).<br>
<div><div class="h5"><br>
On 15 January 2014 23:43, Erickson <<a href="mailto:ericksonb.eng@gmail.com">ericksonb.eng@gmail.com</a>> wrote:<br>
> List,<br>
> I am creating a data processing pipeline using ADJUST to automatically<br>
> reject components. Before I let this monster loose on my data, I want to<br>
> make sure that I should trust ADJUST when it tells me that components low on<br>
> the list are artifacts. To explain:<br>
><br>
> ICA components are returned in order of decreasing mutual information. Thus,<br>
> component 1 has the highest MI, 2 the next highest, etc. At the EEGLAB<br>
> conference this year, Scott made the point that components very low on the<br>
> list contain very low MI and may not actually represent sources, but noise<br>
> not otherwise accounted for. Yet ADJUST does not seem to have any cutoff -<br>
> it will happily label component 84 out of a dimensionality of 84 as an<br>
> artifact. If this component is not reliable, I certainly don't want to<br>
> eliminate it from my data as an artifact.<br>
><br>
> Has anybody dealt with this issue? Perhaps by setting some cutoff (half the<br>
> dimensionality? or some measure of MI?) below which we no longer accept that<br>
> calling a component artifactual makes sense?<br>
><br>
> Thank you,<br>
> Brian<br>
><br>
> Brian Erickson<br>
> Ph.D. Candidate<br>
> Applied Brain and Cognitive Sciences Program<br>
> Drexel University, College of Arts and Sciences, Department of Psychology<br>
><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">eeglablist-unsubscribe@sccn.ucsd.edu</a><br>
> For digest mode, send an email with the subject "set digest mime" to<br>
> <a href="mailto:eeglablist-request@sccn.ucsd.edu">eeglablist-request@sccn.ucsd.edu</a><br>
<br>
<br>
<br>
--<br>
Marco Buiatti, PhD<br>
<br>
CEA/DSV/I2BM / NeuroSpin<br>
INSERM U992 - Cognitive Neuroimaging Unit<br>
Bāt 145 - Point Courrier 156<br>
Gif sur Yvette F-91191 FRANCE<br>
Ph: <a href="tel:%2B33%280%29169.08.65.21" value="+33169086521">+33(0)169.08.65.21</a><br>
Fax: <a href="tel:%2B33%280%29169.08.79.73" value="+33169087973">+33(0)169.08.79.73</a><br>
E-mail: <a href="mailto:marco.buiatti@gmail.com">marco.buiatti@gmail.com</a><br>
<a href="http://www.unicog.org/pm/pmwiki.php/Main/MarcoBuiatti" target="_blank">http://www.unicog.org/pm/pmwiki.php/Main/MarcoBuiatti</a><br>
<br>
***********************************************<br>
</blockquote></div><br></div>