spikes in Brainflow Time Series, noise?

SystemSystem
edited September 2025 in Ganglion
This discussion was created from comments split from: Clarification on EEG Signal Units from Ganglion Using Python and BrainFlow.

Comments

  • Tony_7Tony_7

  • wjcroftwjcroft Mount Shasta, CA

    Hi, those are simply noise spikes, likely from electrical sources in your environment. Because of the regular spacing. Where are your reference and ground leads?

    William

  • wjcroftwjcroft Mount Shasta, CA

    @Tony_7 8:09PM 183.69.137.133
    hi，
    I am currently using Python and the BrainFlow library to obtain electroencephalogram (EEG) signals from the Ganglion board. I have obtained a total of 5 seconds of EEG data. But I found that there was a fluctuation every second before drawing. I ran one of the sample Python files on the Brainflow documentation site to draw and filter the channel data.

    Please comment on this separate thread. See comments above.

  • Tony_7Tony_7

    @wjcroft said:

    @Tony_7 8:09PM 183.69.137.133
    hi，
    I am currently using Python and the BrainFlow library to obtain electroencephalogram (EEG) signals from the Ganglion board. I have obtained a total of 5 seconds of EEG data. But I found that there was a fluctuation every second before drawing. I ran one of the sample Python files on the Brainflow documentation site to draw and filter the channel data.

    Please comment on this separate thread. See comments above.

    I used to think this was what is called the "boundary effect?"

  • wjcroftwjcroft Mount Shasta, CA

    Try moving your setup to an electrically 'quiet' area away from all such electronic equipment: power supplies, extension cords, LED lighting, CFL lighting, conduits, wireless, etc.

    I asked where you have your reference and ground.

  • Tony_7Tony_7

    I tested the electroencephalogram (EEG) data on the GUI using the Ganglion board. The reference electrodes were attached to both ears, and the waveform diagrams were displayed normally. Then, filter the Ganglion board in python through brainflow and display the data. Then the above-mentioned problems were discovered. The following is the code for my filtering

  • Tony_7Tony_7
    edited September 2025 
    DataFilter.detrend(padded_data, DetrendOperations.CONSTANT.value)
DataFilter.perform_bandstop(padded_data, sampling_rate, 48.0, 52.0, 2, FilterTypes.BUTTERWORTH_ZERO_PHASE, 0)
DataFilter.perform_bandstop(padded_data, sampling_rate, 58.0, 62.0, 2, FilterTypes.BUTTERWORTH_ZERO_PHASE, 0)
DataFilter.perform_bandpass(padded_data, sampling_rate, 5.0, 50.0, 2, FilterTypes.BUTTERWORTH_ZERO_PHASE, 0)
  • wjcroftwjcroft Mount Shasta, CA

    If the signal looks fine on the GUI, but you see noise on your Brainflow version, then suspect the filtering you are doing. You can use the GUI filter menu to see what the default filtering parameters are there.

    https://docs.openbci.com/Software/OpenBCISoftware/GUIWidgets/#filters

    Your 'filter order' parameter of '2' is somewhat low. Try perhaps 4 or 5 for sharper filter edges.

    As previously mentioned, this type of noise is commonly in the electrical environment.

Sign In or Register to comment.

© Vanilla Keystone Theme 2026