I have started analyzing a signal from a MEA experiment that consists of reading from electrodes inserted in a brain slice. Anyway, the signal is sampled at 2 KHz but the readings we have are very strange. It looks like it is very noisy and we cannot understand why it is so much noisy so one of our doubts was if the sampling frequency was to high that it incorporated many artifacts or folded down frequencies into our first half of the Fourier transform frequency spectrum.

I will upload a figure so that you can see it and opinions are appreciated!

enter image description here

The time domain plot of the signal is the shown below, as you can see, it is very noisy but we have the important part, the peaks.

enter image description here

  • 1
    $\begingroup$ Have you tried feeding your electrodes with a known signal (ie, noiseless) before realizing this kind of experiment ?Therefore you will know the bias of your experiment tools. Futhermore I do not understand why you are interested in looking for frequencies above 1kHz if your sampling frequency is 2kHz (based on Shannon Theorem) $\endgroup$ Sep 8 '20 at 9:24
  • $\begingroup$ If I could, I would have used a control signal for that same purpose but they didn't, I was just given this signal. And actually, I am interest on just looking up to 200Hz since the biologically speaking, the neurons won't go more than that. The signal looks very dirty to me. I will also upload a figure with the time domain of the recorded signal $\endgroup$ Sep 8 '20 at 9:39
  • $\begingroup$ Sorry, i thought that the scale was in kHz... What is your FFT length, did you apply this transform on the whole sequence ? $\endgroup$ Sep 8 '20 at 11:06
  • $\begingroup$ Looks like power line noise (surge) of some sort. $\endgroup$
    – Shannon
    Sep 8 '20 at 13:09
  • $\begingroup$ Yes, I applied the fourier transform to the full sequence and then plotted just the first half of the frequency spectrum. It looks like this high frequency noise is constant but even if I remove the frequencies above 100Hz, the noise is still there. $\endgroup$ Sep 8 '20 at 15:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.