Hi,<br>tsss is more tricky, if uses default 4 sec window then you different dimensionality every 4 sec.<br><br>Sheraz<br><br><div class="gmail_quote">On Thu, Jul 7, 2011 at 12:07 PM, Sylvana <span dir="ltr"><<a href="mailto:sylvana.schister@utah.edu">sylvana.schister@utah.edu</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">Hi all,<br>
<br>
Thank you so much for your quick reply. This is all very helpful. I am<br>
using tsss filtered data. After thinking it through all day, I<br>
understood that I have a dimensionality problem.<br>
<br>
Anyway, thanks again!<br>
Cheers,<br>
<font color="#888888">Sylvana<br>
</font><div><div></div><div class="h5"><br>
On Wed, 2011-07-06 at 22:18 -0600, Sheraz Khan wrote:<br>
> Hi Joseph,<br>
><br>
> Most likely cause of ill-conditioned in neuromag data is maxfilter<br>
> (signal space separation) which reduces the dimensionality from 306<br>
> (Channels in neuromag machine) to 64 (Spherical harmonics Multipolar<br>
> expansion of order 8). Also neuromag is consist of two different type<br>
> of sensors (Planar Gradiometer and magnetometer) having different<br>
> sensitivities, so either ica needs to run separately on them or<br>
> combined by making them equal through normalization.<br>
><br>
> Sheraz Khan<br>
> Martinos Center<br>
> Boston, USA<br>
><br>
><br>
><br>
> On Wed, Jul 6, 2011 at 9:27 PM, Joseph Dien <<a href="mailto:jdien07@mac.com">jdien07@mac.com</a>> wrote:<br>
> I've had this same problem with EEGlab's runICA implementation<br>
> when the rank of the data was less than the number of<br>
> variables (i.e., it's an ill-conditioned matrix, which is to<br>
> say it has multicolinearity). Including the reference channel<br>
> can do this (since it's just zeroes). Using PCA to reduce the<br>
> data to a subspace can help as well since the rank will now<br>
> equal the number of variables (namely the number of factor<br>
> scores will equal the number of retained factors). Before you<br>
> go the PCA route, it's worth trying to figure out why the data<br>
> is ill-conditioned. There is likely to be something wrong<br>
> with one of your channels. Maybe two of them are shorted<br>
> together or something along those lines. Or if you used mean<br>
> mastoid reference and included both mastoids, that can do it<br>
> too (since they will be exactly inversely correlated, one of<br>
> them is statistically redundant, reducing the rank by one).<br>
> My EP Toolkit artifact correction routine<br>
> (<a href="http://sourceforge.net/projects/erppcatoolkit/" target="_blank">http://sourceforge.net/projects/erppcatoolkit/</a>) takes all<br>
> these sorts of things into account when performing eyeblink<br>
> correction.<br>
><br>
><br>
> Cheers!<br>
><br>
><br>
> Joe<br>
><br>
><br>
><br>
><br>
> On Jul 6, 2011, at 8:44 PM, Sheraz Khan wrote:<br>
><br>
> > Hi Sylvana,<br>
> ><br>
> > Is the data SSS (maxfilter), I normally have this problem<br>
> > when data is SSS, try doing ICA with PCA (set at 64<br>
> > components).<br>
> ><br>
> > Sheraz Khan<br>
> > Martinos Center<br>
> > Boston, USA<br>
> ><br>
> > On Wed, Jul 6, 2011 at 2:09 PM, Stephen Whitmarsh<br>
> > <<a href="mailto:stephen.whitmarsh@gmail.com">stephen.whitmarsh@gmail.com</a>> wrote:<br>
> > Dear Sylvana,<br>
> ><br>
> > >From the top op my head: When running the ICA, try<br>
> > only computing it<br>
> > for the MEG or EEG channels. (cfg.channel = 'MEG';),<br>
> > so it doesn't run<br>
> > on reference sensors in case of MEG.<br>
> > I have hunch that might help out.<br>
> ><br>
> > Cheers,<br>
> > Stephen<br>
> ><br>
> ><br>
> ><br>
> > On 6 July 2011 19:37, Sylvana<br>
> > <<a href="mailto:sylvana.schister@utah.edu">sylvana.schister@utah.edu</a>> wrote:<br>
> > > Hi all,<br>
> > ><br>
> > > I am trying to use ICA to clean EOG and ECG<br>
> > artifacts of neuromag data.<br>
> > > I am running ft_topoplotIC(cfg, comp), where<br>
> > 'comp' is the output of<br>
> > > ft_componentanalysis.<br>
> > ><br>
> > > I get the following error message:<br>
> > ><br>
> > ><br>
> > > ??? Error using ==> surf at 74<br>
> > > X, Y, Z, and C cannot be complex.<br>
> > ><br>
> > > Error in ==> ft_plot_topo at 184<br>
> > > h = surf(Xi-deltax/2,Yi-deltay/2,zeros(size(Zi)),<br>
> > Zi, 'EdgeColor',<br>
> > > 'none', 'FaceColor', shading);<br>
> > ><br>
> > > Error in ==> ft_topoplotER at 753<br>
> > ><br>
> > ft_plot_topo(chanX,chanY,datavector,'interpmethod',cfg.interpolation,...<br>
> > ><br>
> > > Error in ==> ft_topoplotIC at 116<br>
> > > ft_topoplotER(cfg, varargin{:});<br>
> > ><br>
> > ><br>
> > > The complex variable is 'Zi', which is taken from<br>
> > 'comp'. I can modify<br>
> > > the function to take the real part of the variable<br>
> > (or the absolute<br>
> > > value), but I am not sure if this would be the<br>
> > correct thing to do. In<br>
> > > case this is a bug, I thought I should report it.<br>
> > Any comments?<br>
> > ><br>
> > > Thanks for your help,<br>
> > ><br>
> > > Sylvana Schister<br>
> > > Univ. of Utah<br>
> > > Dept. of Bioengineering<br>
> > ><br>
> > > _______________________________________________<br>
> > > fieldtrip mailing list<br>
> > > <a href="mailto:fieldtrip@donders.ru.nl">fieldtrip@donders.ru.nl</a><br>
> > ><br>
> > <a href="http://mailman.science.ru.nl/mailman/listinfo/fieldtrip" target="_blank">http://mailman.science.ru.nl/mailman/listinfo/fieldtrip</a><br>
> > ><br>
> ><br>
> > _______________________________________________<br>
> > fieldtrip mailing list<br>
> > <a href="mailto:fieldtrip@donders.ru.nl">fieldtrip@donders.ru.nl</a><br>
> > <a href="http://mailman.science.ru.nl/mailman/listinfo/fieldtrip" target="_blank">http://mailman.science.ru.nl/mailman/listinfo/fieldtrip</a><br>
> ><br>
> ><br>
> > _______________________________________________<br>
> > fieldtrip mailing list<br>
> > <a href="mailto:fieldtrip@donders.ru.nl">fieldtrip@donders.ru.nl</a><br>
> > <a href="http://mailman.science.ru.nl/mailman/listinfo/fieldtrip" target="_blank">http://mailman.science.ru.nl/mailman/listinfo/fieldtrip</a><br>
><br>
><br>
><br>
> --------------------------------------------------------------------------------<br>
><br>
><br>
> Joseph Dien,<br>
> Senior Research Scientist<br>
> University of Maryland<br>
><br>
><br>
> E-mail: <a href="mailto:jdien07@mac.com">jdien07@mac.com</a><br>
> Phone: <a href="tel:301-226-8848" value="+13012268848">301-226-8848</a><br>
> Fax: <a href="tel:301-226-8811" value="+13012268811">301-226-8811</a><br>
> <a href="http://homepage.mac.com/jdien07/" target="_blank">http://homepage.mac.com/jdien07/</a><br>
><br>
><br>
><br>
><br>
><br>
><br>
><br>
><br>
><br>
><br>
><br>
><br>
><br>
><br>
><br>
><br>
> _______________________________________________<br>
> fieldtrip mailing list<br>
> <a href="mailto:fieldtrip@donders.ru.nl">fieldtrip@donders.ru.nl</a><br>
> <a href="http://mailman.science.ru.nl/mailman/listinfo/fieldtrip" target="_blank">http://mailman.science.ru.nl/mailman/listinfo/fieldtrip</a><br>
><br>
<br>
<br>
</div></div></blockquote></div><br>