Band Power always at maximum ?

edited March 14 in Ganglion

Hello,
I recently got a starter kit with Ganglion board and dry electrodes, and am getting started with the setup after going throught tutorials. I noticed that the "Band Power" graph is always at it's maximum. Is there a way to adjust to this?

Also noticed that the Time Series & FFT plot are producing graphs no matter if I have electrodes placed on my head or not. Should this be happening?

Thanks,
Ivan

Comments

  • wjcroftwjcroft Mount Shasta, CA

    With electrodes disconnected from the scalp, you are just picking up environmental electrical noise, which is unpredictable. You will only get valid EEG when the channel, reference and ground are connected to the scalp or ear lobes. Turn off unused channels by togglling the Time Series channel number button.

    Valid EEG is mainly below 100 microvolts. So your Band Power graph is just getting swamped. Are you following the steps in this tutorial?

    https://docs.openbci.com/GettingStarted/Boards/GanglionGS/

  • edited March 11

    Thanks on this one as well, William. You said "Valid EEG is mainly below 100 microvolts. So your Band Power graph is just getting swamped. Are you following the steps in this tutorial?" I found this one:

    Just looked at one you sent and do not see any difference there except for the fact that my electrodes are dry (just like in the video). It seems the instructor in the video also had the same issue with Band Power being at max on all 5 graphs.

    What could cause the Band Power Graph to be swamped, as you said?

    Thanks,
    Ivan

  • wjcroftwjcroft Mount Shasta, CA

    Are you following the steps in the tutorial link I gave? Do you have cup electrodes and paste to contrast how that works? With dry electrodes, skin pressure hugely determines conductivity. So if you are using a headband kit, try increasing tension. You have not mentioned on any of your comments where your reference and ground are located.

  • edited March 11

    Hi William, here are answers one by one:

    • Are you following the steps in the tutorial link I gave? - Yes, I am.
    • Do you have cup electrodes and paste to contrast how that works? - I got this kit, which as far as I can tell does not include cup electrodes. https://shop.openbci.com/products/low-cost-biosensing-starter-bundle
    • With dry electrodes, skin pressure hugely determines conductivity. So if you are using a headband kit, try increasing tension.

      • Will give it a shot.
    • You have not mentioned on any of your comments where your reference and ground are located. - I am not sure what reference and ground mean - I just have 4 electrodes on a hand band, that's all. I do have ear clips - one on each ear. Do you mean them?

  • wjcroftwjcroft Mount Shasta, CA

    "I am not sure what reference and ground mean"

    At the bottom of the link I gave,

    https://docs.openbci.com/GettingStarted/Boards/GanglionGS/

    Is this section:

    https://docs.openbci.com/GettingStarted/Boards/GanglionGS/#connect-yourself-to-openbci

    Which then links to:

    https://docs.openbci.com/GettingStarted/Biosensing-Setups/EEGSetup/

    Which in turn links to:

    https://docs.openbci.com/AddOns/Headwear/HeadBand/

    Study the last link carefully. All EEG requires, in addition to the channel wires / electrodes -- two additional wires to the reference and ground electrodes. If you are missing these then EEG cannot work at all. EEG is measured differentially between the channel electrode and the reference electrode.

    In your headband kit, they include two ear clips, one is for reference and the other ground. These must be connected to the Ganglion.

  • Thank you, William. I read throught the material you sent and followed to the T. Still the same result, though.

    One thing I noticed is that the background noise (as you have named it, I believe) is very large in my output vs. what's in the examples. So here's my output (same as I posed before) where electrodes are not hooked up to anything at all.

    And here's output from one of the examples in one of the links you gave, which is hooked up:

    Seems rather drastic to me. Any idea what can be causing this?

  • wjcroftwjcroft Mount Shasta, CA

    Did you follow the instructions to change the Ganglion channel switches? They default to NOT using reference pin.

    https://docs.openbci.com/Ganglion/GanglionSpecs/#inverting-input-select-switches

  • edited March 12

    I figured it out!! Or most of it, I think. Lol!

    • It's the EMF from my laptop and electricity in the house! I did an experiment where had the sensors and the board close to the laptop, and then further and further away. The background rose higher and higher, the closer I got to it. Background noise went from 600 or so uVrms to about 3 as I moved away from the laptop!
    • I also experimented with turning the master breaker switch in the house - that did some good, but not nearly as much. Background went from 3 to about 1.5.
    • Also practiced tensing my muscles and relaxing. During tensing, the uVrms went up to above 400-600, during relaxation, it gradually went down to 100 to 20 or so.

    So a follow-up question - I am not always going to be in a perfectly relaxed state - muscles or mind. Is there a way to scale the band power graph to go above 100?

    PS: I did switch Ganglion channel switches to the down position after pealing the sticky tape.

  • wjcroftwjcroft Mount Shasta, CA

    You need to get away from external EMF sources. Then your band powers will drop below 100 microvolts.

  • Ok, thanks, William. Did some experimenting and seems to be working now.

  • Hi William,
    I was waiting for some hadware from the main office, so just restarted experimenting.

    While far from an EMF source, I get the following output. It changes, of course, up and down a bit, but that's about the average.
    Next image is the output with 2 electrodes on, which shows uVms little less then double the backround and band power a bit higher then double (again, this goes up and down, but this seems to be close to the average).

    So essentially the background is about 1/2 to 1/3 of the final output. Is this accurate? Shouls I be getting higher uVms and Band power after connecting electrodes?

  • wjcroftwjcroft Mount Shasta, CA

    As I mentioned before, dry electrode amplitude tends to be substantially lower than 'wet' electrodes such as cups/paste or gelled type. Dry amplitude is greatly influenced by the amount of pressure on the electrode. Try tightening the band(s) to see if that gives you stronger signal.

    Since I think you just have one band, the alpha will be stronger in the rear at O1, O2. With eyes closed, you should see the alpha production increase.

    EEG below 5uV tends to have a lot of noise, so hopefully you can get some more amplitude.

  • edited March 26

    Ok, thanks. I tried pressing electrodes harder against skin, but the results are about the same. What amplitude would you expect with wet electrodes you mentioned?

  • wjcroftwjcroft Mount Shasta, CA

    Have you tried the alpha test. That is what is important.

    https://docs.openbci.com/GettingStarted/Biosensing-Setups/EEGSetup/#4-alpha-brain-waves-eeg

    You will get more amplitude and less noise with wet electrodes, but hard to quantify. Try to make what you have work for you.

    The forehead flat electrodes will probably produce more amplitude than the comb electrodes. But alpha production at forehead is not great.

Sign In or Register to comment.