OpenBCI and Neuromore readings in hertz amplitudes

SkalbeSkalbe Latvia
in Software
Hi,
I'm currently working on getting the OpenBCI signals processed into hard number hertz amplitude output through Neuromore.
The current issue I'm facing is I don't really understand what it is that is coming out of the OpenBCI parent node and how to process it. I've looked at the default example that detects alpha waves, but I don't see any 'real' numbers passed to any of the nodes in the live graph. (You know, when the links between the nodes display the numbers passed inbetween them)
I have until now used the OpenBCI GUI and understand that what gets read from the hardware is the microvolt (uV) values and the activity is quantified in the electrode region simply by combining the value of the microvoltage data, is this correct or am I misunderstanding something?
What I`m getting under normal circumstances in Neuromore is something like a constant  0.00 - 0.03 from any OpenBCI parent node electrode pin, while in the OpenBCI GUI the normal activity was around 4-6 uV range. I tested the device 
For a time I was also under the impression that those same (4-6) microvolt values represented the signal frequency amplitude as well, but now I have some doubts.
What is the process behind converting the input uV values to a hertz frequency?
I have background in programming and I've read quite a lot of academic material on EEG and neurofeedback so I am quite confident in the lingo and the theory of the method, but there is something I feel I am missing with this Neuromore processing software and I'd be very thankful if there was someone who could shed some light on it. Thank you.

Comments

  • wjcroftwjcroft Mount Shasta, CA
    The raw values from the OpenBCI board contain a millivolt DC offset that must first be removed, before you can see the microvolt values. See,


    OpenBCI_GUI does this for you. In neuromore you must do this yourself with a filter element.

    William


  • SkalbeSkalbe Latvia
    I read that thread and understood the theory, but my technical application in Neuromore is lacking I think.
    What I have is the starting classifier that comes with Neuromore for the OpenBCI and nowhere in the setup or any tutorial does it say anything about setting up any manual DC filters. As far as I understand the example should work out of the box. Or am I mistaken?
    This here is a running classifier with all the filters that have banding set up and such. I found the Settings on the upper right and took down the uV/Div to 1 so that comes in in the original scale, if that's relevant. I have my OpenBCI hooked up here and running normally. The graph lines actually fluctuate at runtime as normal when I blink my eyes, for example, and the alpha detection that the example is supposed to perform also works fine (when I close my eyes and relax). This leads me to believe that it is working as intended, but the numbers don't change whatever I do. The relevant part of the classifier is the one that goes into the FFT node and then into the frequency bands and this is where I believe I should see some real signal magnitude numbers, but it's all zeros. I don't understand why. I don't see where the magic happens. =/

    image
  • wjcroftwjcroft Mount Shasta, CA
    Sorry, but I have not run the neuromore examples recently. Your screen grab does not show the complete circuit (classifier) diagram with all the connections, including the OpenBCI channels.

    As you refer to a "DC filter", more typically this is just accomplished with a desired bandpass filter, say from 8 to 12 hz which is alpha. Such a bandpass would automatically remove the DC components below .5 hz.

    Another type of filter sometimes used is a "high pass" filter, for example at .5 hz. This would pass all frequencies above .5 and remove those below.

    The fact that neuromore is showing you the correct response with your eyes closed, must mean that it is setup correctly.

    If you are looking for another tutorial on VPLs and filtering, take a look at this BrainBay tutorial, many of the same ideas apply,

  • SkalbeSkalbe Latvia
    Thank you, wjcroft.
    I cross-studied OpenVibe as well as BrainBay and discovered that both work as intended. I built some custom BrainBay designs and tried recreating them in Neuromore. Neuromore still has a better communication (for me) built in so I would like to stick with it.
    Do you know if there`s a resource somewhere that describes what all the LED lights on the OpenBCI mean? Like when a green small one is blinking or if the large red LED is supposed to be on at all times when the device is being run by some interface?
    This is what it all looks like running in the OpenBCI GUI which seem all fine:
    Then when I turn that off and launch Neuromore, my spectrogram view and all of the uV values in all channels go flat no matter what I do:
    This is my classifier:
    Do you see anything that I don't?
    Can somebody please point me in the right direction why my Neuromore does not work? I posted on Neuromore support forums as well, but that community seems very much dead.
  • SkalbeSkalbe Latvia
    Update: Neuromore representative confirmed that there is some kind of a serial comm problem on their end.
Sign In or Register to comment.

© Vanilla Keystone Theme 2026