BioSemi.nl

Hi to all,

A few weeks ago I decided to test the functionality of my electrodes, by doing the recomended measuring in salty water.

The data looked fine and everything seemed ok. Then I noticed, that there is a higher correleation between the data for higher sample frequencies.

e.g.: mean of correlation of the pairwise compared data of electrodes
2kHz: r=0.5
4kHz: r=0.63
8kHz: r=0.69
16kHz: r=0.73

I just want to know, whether this is normal. And if it is, why does this effect appear? I couldn't find any explanation.

I tested 8 electrodes (EX1-EX8) in an empty room in an old building without any electric cables or lines nearby. I also took care for the electrodes not to touch each other.

Did anybody else made this observation or knows what the source of this is?

Thanks a lot.
max

Did you use the unreferenced (raw) signals as basis of your analysis ?

The correlation is a measure for the amount of Common Mode in the signals. The CMS/DRL loop becomes less effective at higher frequencies (the loop amplification decreases proportionally to the frequency). So, it is to be expected that with a higher bandwidth, a relatively higher amount of Common Mode will be present in the signals.

Best regards, Coen (BioSemi)

Dear Coen,

thanks for your fast reply.

Yes, I did use the unreferenced signal. I thought, that if I reference the data, I would increase the correlation, as I add (or better said substract) a common component to the data.
Of course in the raw data there is the CMS/DRL as the common component. But I was just wondering, if then an increase of correlation for higher sample frequencys is normal.

Your explanation with the higher amount of Common Mode sounds quite convincing.

Thanks a lot.
Kind regards, max.