I can't understand "spikes of noise" in my EMG signal

BonzoBonzo
edited August 2018 in Cyton
Hello,

I attach two images that should be explicative.

In the first pisture there is the EMG signal from one channel; i didn't contract any muscle, so the "spikes" you see is all unwanted. (the signal is acquired [I suppose at 250 Hz] and then filtered with a notch filter at 50 Hz + a highpass filter at 20 Hz.)

In the second picture you can see the spectrum (fft) of the previous signal.

Can anyone help me to understand such unwanted noise/artifact? Any of you experienced the same thing? Could it be possible that my Cyton board is simply "broken" in some way?

Thank you very much in advance.

Roberto

image
image

Comments

  • wjcroftwjcroft Mount Shasta, CA
    Bonzo, hi.

    What happens when you follow exactly, all the steps in the tutorial?


    Do you see the ECG, EMG, EEG examples as shown?

    It might be the case that your electrode connections are not optimal. Try following the suggestions in the tutorial.

    Regards,

    William

  • BonzoBonzo
    Hi William, thank you for your answer.

    I connected the electrodes in the following way:
    • 4 differential channels (channel = couple of electrodes connected to N and P pins);
    • ch1 -> 1N & 1P, ch2 -> 2N & 2P, ch3 -> 3N & 3P, ch4 -> 4N & 4P.
    • In the pictures o the previous post I only plotted ch1 (but i connected ch1 and ch2 on a couple of extensor/flexor muscles of the forearm.)
    I think my configuration should be in accordance with the explanation of the cyton getting started tutorial.

    Additionally, note that I can see the muscle contraction when i contract my muscle, but the "unwanted spikes" randomly appear.

    Any further suggestion?

    Thanks for your support.

    Regards,
    Roberto
  • BonzoBonzo
    In order to add more detail to my last post, I just acquired the signals from the cyton without connect anything to any pin.

    Here you can see that the unwanted spikes are still present:
    image

    And in this zoom you can see that such spikes all have the same shape:
    image

    Thanks for your attention.

    Regards,

    Roberto
  • BonzoBonzo
    Sorry for posting three times in a row, but I want to add further information, so that maybe I can be better helped.

    I just acquired the raw data from the Cyton board without any filtering.

    As you can see in the image below, there are some impulsive-like "outlier" values during the acquisition of the signal. The oscillatory behaviour I have shown before is simply due to the notch+high-pass filtering.
    image

    All the big peaks you see are unwanted, and zooming they have this shape:
    image

    This seems a very strange behaviour, it looks like the board has problems. But I wait your opinion, I really hope I am wrong.

    Thanks, regards,

    Roberto
  • wjcroftwjcroft Mount Shasta, CA
    edited August 2018
    Are you connecting the Bias / Ground? Are you adjusting the CHAN SET settings, as shown in the tutorial?

    I'm suggesting you go through ALL the examples in the tutorial, because it covers ALL the required steps. For example, can you see EEG and ECG as well as EMG? Have you ever seen EEG or ECG with this board?

    Because you did not mention Bias, nor altering the settings, it sounds like you may be missing points from the tutorial.

    The Bias / Ground must be attached in all circumstances. The amplifier uses it to center the differential inputs. The settings must be adjusted to alter the scale factor.
  • wjcroftwjcroft Mount Shasta, CA
    Also suggest showing the GUI screens rather than your graphs (Matlab?), as that is one less step of processing.
  • BonzoBonzo
    Hi William,

    I tried out the bias and now it seems that is working properly!

    Thank you, I already red all the tutorial, but from my understanding I was thinking that for a single differential channel the ground pin was not necessary! Definitely wrong!

    Thanks again for your useful help and for your time!! :)

    Cheers,

    Roberto
Sign In or Register to comment.

© Vanilla Keystone Theme 2026