SRB and references

bmarkusbmarkus
in Cyton
Hi Fellow EEG Geeks,

I could use some help getting my hardware setting right for the references.

My professor wants to use A1 as a reference for all sensors on the left side, and A2 for all sensors on the right. A1 and A2 will be on earlobes, not on the bone. I have Bias on Cz. 

I've been trying to use SRB1 and SRB2 for A1 and A2. All the sensors on SRB1 give good results, but on SRB2 they all come out with the same signal..

How to connect A1 and A2 instead? And hardware settings?

Is it possible to take a lead from A1/2 and split it to all P pins? 

Another option would be to record A1 and A2 and then reference in software, creating a custom widget. What would be the hardware settings for this?

I hope someone can help out. If I get it to work, it will be used in a study to see about brain to brain synchronization during different activities. (studies have already shown that brains do sync when in close proximity and engaged in same activity)  

Cheers,
Bernard

Comments

  • wjcroftwjcroft Mount Shasta, CA
    edited April 2019
    Bernard, hi.

    The default Cyton settings, bus all the IN*P pins (IN1P, IN2P, etc. the pins farthest from the motherboard) and present that as SRB2, the SRB pin closest to the motherboard. Thus the IN*N pins are your channels and SRB2 is the reference for those channels.

    My impression of the solution for your unusual reference hardware requirement, is to do the following. Assuming IN1N through IN4N are you left channels, connect A1 to SRB2 as usual. IN5N through IN8N are your right channels.

    Now using some type of splicer mechanism, (such as cutting and splicing Adafruit jumpers),


    Splice together 4 females and 1 male. Connect the females to IN5P through IN8P. The male now connects to A2.

    With the Cyton GUI Settings panel,


    (Section 4e.) Turn off SRB2 on channels 5 through 8. Leave the SRB1 alone (as 'no').

    I think that should do it.

    Regards,

    William


  • BillhBillh
    The usual thing to do is record A1 and A2 and then, in software, vector subtract these from the scalp leads you are getting the EEG from. You connect them by clipping or sticking a wire lead to the ears, A1 left ear and A2 right ear. Paste with gold cup electrodes held perhaps with paper bandage tape can work, and clips can be bought for this. 

    You can also put the leads on the head at the small bony bare skin patches right behind each ear-- technically doing that is using what is called M1 and M2, not A1 and A2, but it should make little difference generally.


  • wjcroftwjcroft Mount Shasta, CA
    edited April 2019
    Bill, thanks.

    Bernard, Bill's suggestion works if you have enough channels to record A1 and A2 as normal channels, not references. If you are short on channels and need 4 left and 4 right channels, then the hardware hack I suggested should work.

    Though as everyone is mentioning, software re-referencing is more and more common, but is not provided directly by the GUI.


    Regards,

  • bmarkusbmarkus
    Thanks William and Bill.

    I'll be trying both of your suggestions today.

    First by splicing to the P-pins to demonstrate to Prof that the setup works, and in mean time I'll find a programmer to adjust widget for software re-referencing.

    I'm using 16 channels and Prof has already approved to sacrifice 2 channels for reference recording.

    I would like to record without using SRB or P-pins, and then reference in software, what would be the hardware settings for this? Have to change input type?

  • wjcroftwjcroft Mount Shasta, CA
    Before re-referencing, you do need to have a reference for hardware. That is how differential amplifiers work. Using SRB2 is the simplest, as it is the default. Not sure what placement to recommend, perhaps on the central axis. Cz might be OK with your Bias at AFz, AFz is the location used for Ground / Bias on Electro-caps.
  • bmarkusbmarkus
    Ok, I'll put SRB2 on Cz, and Bias on Fz (that's available on my cap).

    I think my understanding of SRB was wrong. I've worked with Biosemi before, where the software shows the channels as follows:
    Fp1-A1
    Fp2-A2
    F3-A1
    F4-A2
    Etc
    The A channels were subtracted from the other channels to reduce noise. I assumed that the same thing happened inside the ADS1299 through the SRB channels. Is that incorrect? Are SRB channels there just to complete the circuit?
  • wjcroftwjcroft Mount Shasta, CA
    All differential amplifiers have TWO poles, plus and minus. In operation the amp is subtracting the two. This cancels common mode noise (largely). The SRB2 just busses together all the plus pins.

    I have not used Biosemi, but it also has a bussing system. Two of them, separate A1 and A2 busses.
  • wjcroftwjcroft Mount Shasta, CA
    Another older technique used in QEEG assessments and neurofeedback, is called "linked ears", where the A1 and A2 are connected together (in hardware or software), and that becomes the reference for the scalp electrodes. This gives a more balanced "off the scalp" reference point, without the need for mathematical re-referencing. This linked ears technique predated re-referencing algorithms.
  • bmarkusbmarkus
    Yes, my Professor uses linked ears, but of course, he uses his own version. He doesn't actually link them. A1 for the left side, A2 for the right side, and for the central sensors he averages them in software.

    We look at alpha coherence between left and right. The measurements are more natural if A1 and A2 are not linked with an actual wire.
  • wjcroftwjcroft Mount Shasta, CA
    Biosemi referencing,

  • bmarkusbmarkus
    Hi William,

    I tried your solution, splicing the P pins to A1 and A2. P1,3,5,7 and on A1 and P2,4,6,8 and on A2.
    SRB1 and SRB2 are switched off in hardware settings. Works like a charm!


    I also read the Biosemi referencing link that you sent, good to know that their system is technically different there.

    Thanks for your help, much appreciate it!
  • wjcroftwjcroft Mount Shasta, CA
    Bernard, thanks.

    Cool 3D printed enclosure.

    re: switching off SRB2 for all channels

    Actually you could have saved some cabling by just using SRB2 for the reference on half of your channels. The other have could be separately wired, with SRB2 switched off.

    William

  • bmarkusbmarkus
    At first I tried with half the channels on SRB2. It worked of course. But for final version I think that all wires should be created equal :) 



  • ElliotMElliotM United States
    Great, I was wondering about linked ear referencing without sacrificing a port on my Ganglion for A2. I'll try the spliced jumper cable approach. 
  • wjcroftwjcroft Mount Shasta, CA
    Simplest linked ear wiring, which is used in many neurofeedback situations -- is to use a Y cable. The cable is made of the same Adafruit male / female jumpers mentioned above. Spice one female and two males. Plug the female onto your reference header pin. Plug the left and right ear clips/cups onto the two male connectors of the Y.
Sign In or Register to comment.

© Vanilla Keystone Theme 2026