Detecting Alpha Waves

chipaudettechipaudette
in Software
Hi All,

I've been analyzing some data that I recorded from the new OpenBCI V3 boards.  These recordings are mostly simple exercises in generating Alpha rhythms induced by closing my eyes.  In my previous BCI hacks, I've used these Alpha rhythms to trigger lights and to control robots.  I figured that I'd explore in a little more detail different refinements on algorithms to get solid Alpha detection.

Here's the starting point...simple threshold detection:


Chip

Comments

  • chipaudettechipaudette
    Did some follow-up...I introduced "Guard Bands" to reduce my false alarm rate while maintaining my overall sensitivity.  Works pretty well!

  • chipaudettechipaudette
    Now I'm having some serious fun.  ROC Curves!  I love pronouncing the acronym "ROC" as a word instead of spelling it out as letters.  Makes it seem so tough!  

    I made some ROC curves comparing my "Alpha+Guard" algorithm versus the based "Alpha only" algorithm that I first used.  The ROC curve is excellent at illustrating the superiority of the newer algorithm. It will also be a great tool for making even more improvements.

    If you're interested, check it out at:


    Chip
  • chipaudettechipaudette
    Another follow-up...tuning my detection thresholds using a bunch more EEG data, instead of just the one EEG recording used in the earlier post...


    Chip
  • TJHSSTResearchTJHSSTResearch
    edited March 2015
    Hi Everyone! My research partner and I are currently trying to detect alpha waves using the OpenBCI hardware, but have been unsuccessful thus far. At this point, we are not sure if its a problem with the hardware or something we are doing wrong. If anyone has had experience working with the alpha waves, we would love to hear from you. Please let us know! 
  • wjcroftwjcroft Mount Shasta, CA
    TJHSS, hi. I merged your post into this existing thread. Chip has a number of blog posts on this subject see previous links. There is also another post on alpha detection on board.

    http://openbci.com/forum/index.php?p=/discussion/278/alpha-detector-on-board-or-laptop

    https://www.google.com/search?q=eeghacker+alpha

    William
  • chipaudettechipaudette
    Hi @TJHSSTResearch,

    After you take a look at the material discussed here, but sure to post what you've tried...especially where you put the electrodes on your head, where you plugged them into your OpenBCI board, and what you did to try to get your Alpha waves.  If you can share any pics or any of your recorded data (via dropbox or Google Drive or someplace), that would be very helpful.  Also, you should try more than one person...not everyone shows eyes-closed Alpha waves.

    Chip
  • TJHSSTResearch2TJHSSTResearch2
    Hi guys,

    I am TJHSSTResearch's research partner. We are still at a similar point in our project, unable to detect alpha waves. Attached is a dropbox folder containing our EEG data as well as several pictures we took while collecting data. We also tried with another research subject, but unfortunately obtained the same result. Is there anything obvious that we are doing wrong?


    Thanks in advance for your help!
  • wjcroftwjcroft Mount Shasta, CA
    @TJHSSTResearch @TJHSSTResearch2 , thanks.

    I took a quick look at your photos. Did not look at your data. Just from the photos of the OpenBCI electrode wiring, it doesnt look like you followed our tutorial,

    http://docs.openbci.com/tutorials/01-GettingStarted#getting-started-w-openbci-v-connect-yourself-to-openbci-2-connect-your-electrodes-to-openbci

    In other words, all your EEG electrodes should be plugged into the pins closest to the board. (The IN*N pins). The tutorial picture does show one wire on a IN*P pin, but that is the two wire differential setup for ECG.

    Maybe I'm mistaken and you intended to use the IN*P pins for some reason?

  • chipaudettechipaudette
    I looked at the photos, too.  wjcroft has good eyes...it does indeed look like you plugged into the wrong pins on OpenBCI.  I get this wrong sometimes, too.  With the default configuration of OpenBCI (as set via the GUI), you need to plug into the bottom row of pins (those closest to the board).

    Sorry for any confusion!

    Chip

  • TJHSSTResearch2TJHSSTResearch2
    Hello,

    Thanks a lot for all of your help so far - they were definitely plugged in incorrectly. We repeated the test, plugging them in correctly this time, and got better results. We think that there are alpha waves present in the attached screenshots, but they are not as clear as the ones in the tutorial. Could you please confirm that what we are seeing are indeed alpha waves?

    Here's a Dropbox link with pictures, screenshots, and our raw data: https://www.dropbox.com/sh/70524f06pohz7tg/AACemYh3xk9-ubc6HTmGCP6Da?dl=0 .

    Thanks again for your help!
  • wjcroftwjcroft Mount Shasta, CA
    TJ, hi. Your EEG signal on channel 7 (O1) looks quite noisy. Do you have any ideas why that might be? Is it possible that your electrode on O1 might not be reaching the skin. But instead is more attached to the hair? There is a video tutorial on electrode attachment here,

    http://openbci.com/forum/index.php?p=/discussion/176/placing-electrode-sensors-on-the-scalp

    What you need to do is push the hair away from the O1 site on all sides until you can see the scalp skin. Then push down there. I'd also recommend you use an elastic or velcro headband, to guarantee best contact. The scotch tape you show on the hair at O1 is not really that effective for making skin contact.

    Otherwise it looks like you are close, keep us posted.
Sign In or Register to comment.

© Vanilla Keystone Theme 2026