[Eeglablist] ICA Explanation please?

[Kelly.Kaneswaran]

I have a 20 minute dataset of UP/DOWN/LEFT/RIGHT finger movements from
144 channel biosemi system. We recorded the movements using 128 channels
EEG , 2 wrist which picked up ECG and 4 EOGchannels, the rest
unconnected.

 

I have selected a montage of 43 electrodes across C3 C4 and symmetry of
head line. I ve noticed peaks and troughs alternating at different
frequencies at different movements. I run ICA on up epochs and down
epochs and it gives me 43 components for each. Can someone tell me is
each component mach relating to the actual channel number as in is ICA
removing the noise from each channels contribution to the overall signal
matrix. As in if I llok at component 37 or C3 in my plot , is that the
filtered C3 main component with neighbouring electrode noise filtered
out.

 

Kelly

 

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20090121/6ae51b01/attachment.html>