EMG Signal Noise decreases just by switching around the ground and measuring electrodes. Why?

agwat5agwat5 Melbourne
edited July 2019 in Electrodes
I am currently measuring EMG signals using 3M red dot electrodes and am using a system called open Ephys which uses an Intan RHD 2132 amplifier.
While I am not using the OpenBCI components I believe my question is more generic so I have posted here.

I am using three electrodes one at the upper anterior forearm, one at the anterior wrist and one on the posterior side of the upper forearm. This is to record EMG signals to classify finger movements. 
I initially had the ground electrode at the wrist position, which gives a peak to peak noise of 10mV at 50Hz for the upper anterior forearm electrode, but when I switch the ground electrode to the upper forearm on the anterior side the same electrode pair gives noise of 0.75 mV at 50Hz.

Why would the same electrode give such different amount of noise just by switching the electrodes around, why wouldnt it be the same. 
For the amplifier the ground and references points are connected.

Thanks in advance
Andrew 

Comments

  • BillhBillh
    In general, the closer the reference electrode is to the active electrode the lower the noise. This is because the closer reference tends to be closer in potential, at any given time, to the potential being registered by the active electrode.

  • agwat5agwat5 Melbourne
    Sorry I should have been clearer.
    The distance of the electrodes are the same.
    If have the GND/Ref electrode at position A and the active electrode at position B then I am getting 10mV peak to peak of noise, but if the GND/Ref is in position B and the active electrode is in position A I am getting only 0.75 mV of noise. So the path between to the two setups are the same and are just reversed.

    I would of have expected for the polarity of the signal to be reversed and the noise to be the same level but for some reason the noise level has changed. 

  • wjcroftwjcroft Mount Shasta, CA
    Andrew, hi. 

    I'm not quite picturing your setup. Normally EMG is measured differentially between two electrodes on either side of the muscle body.

    image


    Photo above from:


    Your description sounds to me like you are NOT positioned on either side of the same muscle body. Hence noise level will be higher. Above photo shows a ground lead for each muscle group. In most cases a single ground I think would be sufficient. Paper itself might clarify.

    William
  • wjcroftwjcroft Mount Shasta, CA
    According to this list, Intan DOES make an amplifier more geared to EMG, with bipolar inputs, rather than a common reference.


    But that is apparently not what you have. You have the common reference version. So you can only really measure one muscle at time.
  • wjcroftwjcroft Mount Shasta, CA
    https://www.researchgate.net/post/please_explain_how_a_bipolar_EMG_recording_reduces_the_noise_that_would_have_been_collected_if_using_a_single_electrode
  • BillhBillh
    You initially said you were using 3 wires. Can you clarify whether 2 or 3 wires? If 3 wires, is the third wire attached to one of the other two in some way? That might make a difference.

  • agwat5agwat5 Melbourne
    Hi Billh I have a third wire which is connected to another active electrode, it doesnt matter If i have it connected or not in regards to what I see in the situation i described above. 
  • wjcroftwjcroft Mount Shasta, CA
    Andrew, did you read my earlier post in this thread? Do you have the EMG electrodes on opposite sides of the muscle body?
  • agwat5agwat5 Melbourne
    In a way no, I am using monopolar emg recording, this is in part because of the equipment I have, that being in the Intan amplifier. So I am using a single reference point for the 2 active electrodes. I have seen this set up in a research paper and I have gotten around 80% accuracy in live classfication of 7 finger movements so I believe this setup should be fine. 
  • BillhBillh
    edited July 2019
    The asymmetry with one setup versus the other may reflect the varying distance of your active electrode from the muscle being measured.

    In general EMG setups can benefit from having a third electrode as ground, and referencing both active and reference electrode to ground before measuring the difference between the active and reference electrode as the actual signal you want. This minimizes the static interference signals.

    But this may not matter here. If you can classify 7 finger movements the static should be sufficiently less than your signal so that your setup is fine.

  • wjcroftwjcroft Mount Shasta, CA
    Ground on some amplifiers is injecting an inverse common mode noise cancellation voltage. On ECG equipment 'ground' is also called DRL 'driven right leg'.

    Previously you said, "For the amplifier the ground and references points are connected."

    Agreeing with Bill's last comment, try running the ground separately.

Sign In or Register to comment.

© Vanilla Keystone Theme 2026