Inconsistencies in signal, not using Bias

perception_cognitionperception_cognition
edited August 2016 in OpenBCI_GUI
Hi there,

I have been recording using the OpenBCI 32bit Board, and have been observing some signal inconsistencies. The set-up is: 1 EEG electrode on the bottom SRB pin + 6 EEG electrodes - one on each of the neighbouring bottom pins. The next bottom pin, plus a corresponding pin on top of this bottom pin, each have an electrode for measuring heart rate (i.e., one channel and one reference).  There is nothing plugged into the BIAS pin. In the GUI, the bias is set to "don't include" on all channels; SRB2 is off for the heart rate channel; channel 8 is completely unselected. 

This set-up worked quite reliably the first few times I tried streaming data (i.e., saw what resembled EEG data in 6 channels, and one channel worth of heart rate data). More recently, I've noted some inconsistencies in the signal (the participant, board, and leads are kept perfectly still). For example, after setting up the dongle, board, electrodes, and OpenBCI GUI, I start successfully streaming the data (looks as expected). Then, without any change to the set-up, heart rate data is now in one or two of the EEG channels, blink data appears in multiple channels where it wouldn't be expected, etc. Furthermore, when I turn the board off and on/unplug and replug the dongle (after restarting the GUI), I sometimes see no signal whatsoever, or just the heart rate data, when I saw all channels' expected signals just minutes before (no changes to any part of the experimental set up). 

Does anyone know what this might be due to/know how to fix this issue? Thanks in advance!

Comments

  • wjcroftwjcroft Mount Shasta, CA
    You need to use the Bias pin, it's important for EEG common mode noise rejection, and also centering the differential amplifiers. Without Bias the amplifier may produce unpredictable results.
  • perception_cognitionperception_cognition
    edited August 2016
    Thank you for the helpful reply, wjcroft! Just to make sure I understand, I wanted to ask a few more questions:
    1) To confirm, the bias pin must be plugged in *and* bias must be included in the EEG channels, in the OpenBCI GUI? 

    2) Does the mandatory use of the bias pin apply even in cases when one records OpenBCI during movement, such as walking (i.e., when there would be significant abrupt changes and high frequency noise from the movements being made)?

    3) Is Scenario 1, having the bias pin unplugged and bias set to "not included" on the GUI, functionally identical to Scenario 2, having the bias pin plugged in and bias set to "not included" in the GUI? 

    4) How does the bias pin on SRB2 affect a heart rate channel and reference, amongst a number of EEG channels on SRB2? For example, SRB2 is populated with an EEG reference, N2P, N3P, and N5P (all EEG channels). N4P has one lead on the bottom (SRB2) pin and one lead on the top pin, to create a heart rate channel and reference. How does the bias pin + inclusion of bias in the GUI impact the heart signal in this scenario? I ask because I have seen heart rate enter other SRB2 channels when the SRB2 bias pin was not plugged in. I would have thought that the bias pin would only impact the EEG signal/wouldn't be related to heart rate entering other channels.
  • wjcroftwjcroft Mount Shasta, CA
    1. yes this is the default setup as described in the docs.openbci.com tutorial. It's the default because Bias must be connected somewhere, but not needed for every channel, such as your ECG or EMG.

    2. EEG under movement is usually problematic. Any cable motion or headset motion will result in artifacts. Some elastic cap configurations might be less prone if everything is tightly bound to the head.

    3. The tutorial explains WHY the SRB and Bias is not included on the ECG EMG channels: because the amplitude of those signals would intrude on the EEG channels. Bias line centers the differential amps with a known reference voltage; that is it's primary purpose. Secondarily it injects a small out of phase waveform to attempt to cancel power line noise (usually the strongest common mode noise source.)

    4. The pins closest to the board are IN*N pins (see data sheet), and by default use the common reference SRB2. The top pins are IN*P pins and can utilize SRB1 in a similar fashion, but consult the TI data sheet on the ADS1299 for more specifics.

    Bias and SRB* pins are separate concepts. And can be independently disconnected per channel. At least one channel needs Bias, and if you let your EEG channels default, you have that.

    I think if you follow the example given in the tutorial, just changing the CHAN SET items suggested, you'll have less artifacts and cross contamination between channels. Cable routing can also influence ECG EEG cross contamination. Your ECG leads should not travel parallel to the EEG leads for long distances.
Sign In or Register to comment.

© Vanilla Keystone Theme 2026