Averaged or single reference electrode?

ryansryans
in Cyton
Simple question:  Is the Cyton with Daisy using an average of earlobes or a single earlobe as reference when it is set up exactly according to the tutorial?

According to the TI ADS1299 datasheet, the versatile patient bias drive block allows the average of any electrode combination to be chosen in order to generate the patient drive signal.  The reference voltage for the bias drive can be chosen to be internally generated [(AVDD + AVSS) / 2].  Alternatively, EEG electrodes can be measured in a referential montage in which each of the electrodes is measured with respect to a single reference electrode.

Based on the OpenBCI documentation, it could seem like the latter:  a single reference electrode at the earlobe.   However, the documentation further states the bias uses destructive interference waveform techniques to eliminate the “common mode noise”.  This leads me to believe the developers chose to use the ADS1299 multiplexer to average both of the earlobes [(AVDD + AVSS) / 2].
 
Searching this forum, I have read that the Cyton is using a “linked earlobe reference”, which I believe is referring to “averaged earlobes”.  *To avoid ambiguities, the term “linked” should only be used when two references such as earlobes or mastoids are physically linked.

Thanks.

Comments

  • wjcroftwjcroft Mount Shasta, CA
    Ryan, hi.

    The Cyton tutorial shows the white electrode (SRB2, reference) connected to one earlobe. The Black (Bias) to the other earlobe. The reference in this case is NOT averaged between the earlobes, it's just using the single earlobe.

    The other common reference setup is to use a Y-cable (you could make one of these yourself from male-female header jumpers). Using the Y to connect both earlobes, then the tail of the Y connects to SRB2. This is the linked ears reference commonly used in QEEG neurofeedback.

    Bias / Ground is by default looking at the average of all the IN*N (bottom row pins) to derive it's common mode Bias (180 degree out of phase) injection signal presented on the Bias lead.

    More discussion of Reference schemes in the Electrodes Category on the forum.

    Cyton (if you are looking at tutorials) does NOT by default mention linking the earlobes. But if you read through various posts on the forum you can infer that using a physical linked ear setup is common practice in many applications.

    William
  • ryansryans
    Hello William,

    Averaged Earlobes or Averaged Mastoids as reference are the best practice in my line of ERP research.  This is typically done on the the software side in EEGLab.  Considering the ADS1299 is capable of doing this internally, it would be nice to implement without sacrificing any inputs to record earlobes.

    In the Cyton's current state with firmware 3.0.0, is it possible to do this?  
    If so, how would one go about wiring for this?

    Or, is this something that would need to be changed on the ADS1299 side via RLD_SENSP and RLD_SENSN ?

    Kind Regards,

    Ryan
  • wjcroftwjcroft Mount Shasta, CA
    Ryan, I believe you could do what you want by simply sacrificing one channel to record the other earlobe, as you say, it sounds like you have the Daisy. This is actually how Brainmaster (Discovery amp, 25 channel, QEEG) computes their algorithmic default "linked ears" reference. I'm still not clear that the ADS1299 can do what you want using the Bias as the other reference. In many typical cap montages, Bias/Ground is applied (Electro-cap for example) at AFz, on the midline to in effect, spread the common mode injection evenly to both hemispheres.

    You will have to research this (in the TI ADS1299 hardware pdf) if you want to change the firmware. Did you look at the CHAN SET tab in the GUI? It might have enough flexibility to adjust some of the ADS settings.

    I'm going to mention AJ Keller @pushtheworld, the main developer, to see if he has any suggestions.
  • ryansryans
    Dear William,

    I did have a look at the hardware settings in Processing GUI.  There are the options for each channel:  BIAS_MEAS, BIAS_DRN, BIAS_DRP

    I suppose I could sacrifice two channels, one for each earlobe, and assign BIAS_MEAS to them.  The confusion is then:  is the ADS1299 averaging these internally to determine bias?  If so, is a connection necessary on the bias pin(s)?  

    Else, I could record the earlobes and average them in EEGLab.  

    This should be simple, as Averaged Earlobes or Averaged Mastoids as reference are the gold standard in ERP research.  Reviewers will most likely reject the study otherwise.  I know this is possible based on the ADS1299 datasheet, but I am confused how to implement it.  I think I am overlooking something very obvious.  Perhaps AJ knows the answer @pushtheworld ?

    Kind Regards,

    Ryan
  • iWoodiWood
    @ryans,

    Did you ever find a solution for this?

    Thanks,
    IW
  • ryansryans
    I wish I did.  I am about to conduct a study, and afraid journal reviewers will criticize or reject my methods if Cyton is wired according to tutorial.

     Can anyone help us?

  • wjcroftwjcroft Mount Shasta, CA
    Ryan, hi. As I mentioned in July, it looks easiest to just use an extra channel for the other earlobe. 
  • ryansryans
    edited September 2017

    Dear William,


    I could use a channel to record the other earlobe, but then I believe I would need record both earlobes in order average them together offline as reference.  Now the methods are getting messy, as I would still need to send something to SRB while recording.


    I wish I could figure out how to achieve this inside the ADS1299.  According to the datasheet, it is possible to average two electrodes (i.e. earlobes) for reference.  This is then adheres to the methods of other investigators in my area of ERP research.  


    I am afraid reviewers may hassle me over a single earlobe as a reference citing possible artificial asymmetry between left and right hemispheres.  Am I possibly overlooking a feature of the ADS1299 that would validate my methods of using a single earlobe as reference??  


    Additionally,  I would be hassled by reviewers if I linked them physically with a Y-cable, which creates a zero-resistance electrical bridge between the hemispheres, which may distort the distribution of voltage over the scalp and reduce hemispheric differences.


    If I can not sort this out soon, I may need to use AFz as reference simply to satisfy journal reviewers.


    Kind Regards,


    Ryan

  • pushtheworldpushtheworld
    @ryans

    I have heard of double linked ear lobe and am about to start doing it for one of my experiements. Did you find an answer to your question?
  • ryansryans
    @pushtheworld

    Dear AJ,

    No, I could not solve how to implement "averaged earlobes as reference" on the firmware side of the ADS1299.  According to the data-sheet, it is certainly possible.  Can you have a look? 

    If all you wish to do is physically link earlobes as reference, then you can follow William's advice of a Y-cable.  However, physically linking the wires from these two electrodes creates a zero-resistance electrical bridge between the hemispheres, which may distort the distribution of voltage over the scalp and reduce hemi- spheric differences. This also invalidates the use of source localization techniques.  Most Neuroscience journals frown on this method.

    Kind Regards,

    Ryan
  • wjcroftwjcroft Mount Shasta, CA
    edited October 2017
    Ryan, thanks.

    On the other hand, many neurofeedback clinicians doing QEEG (and source localization) use the zero resistance linked ears technique. And they publish their own journals, such as ISNR Neuroregulation, Journal of Neurotherapy, Neuroconnections, etc.


    Also see,




    Regards,

    William

  • pushtheworldpushtheworld
    The ADS1299 Is a highly complex beast! Check out the Python raspberry pi ads1299 library by PushTheWorld for how to set up the ads1299 in this linked ear lobe configuration.

    @wjcroft awesome awesome link shares!! Thanks!
  • wjcroftwjcroft Mount Shasta, CA
    AJ, tell us more about the Raspberry Pi and ADS1299. Is the Pi connecting through wifi, dongle, or are you saying the pi has a shield with an ADS1299? It looks like the latter,


    Where does one get the shields?

  • wjcroftwjcroft Mount Shasta, CA
    Can the Pi address multiple, stacked shields?
  • pushtheworldpushtheworld
    edited October 2017
    @wjcroft

    This version cannot be stacked but the next one will be, I didn't breakout the bias on this version for multiple. In fact it's got a different design for the bias drive.

    It's an 1.1in by 1.3in and is a blessing to work with in size! It is powered by a single 3V3 power rail and only has an ADS1299.

    Do you want one? I can send a alpha/beta to you.

    The pi can address multiple for sure. Max sample rate for 8 channel is 2kSPS
  • naznaz Italia
    Hi @pushtheworld

    Is it possible to have one of your ADS1299 board to plug into the Raspberry Pi?
    Can you also share the schema?

    I would like to investigate the Raspberry Pi capability provided by your library.


    Thanks



Sign In or Register to comment.

© Vanilla Keystone Theme 2026