24-Channel Ultracortex "Mark IV" Setup ?

edited February 15 in Build-it-yourself

I would like to add more channels to my Ultracortex "Mark IV" EEG headset to go beyond the 16-channel limit. Right now, I'm using it with 8 channels. Can I buy a Cyton + Daisy biosensing board (which has 16 channels) and modify the STL file for the Ultracortex "Mark IV" EEG headset so it can handle 24 channels? I already own a Cyton biosensing board with 8 channels. I believe that if I connect both the Cyton + Daisy board and my existing Cyton board to the modified Ultracortex "Mark IV" EEG headset, I'd get 24 channels in total. Would this work? Are there any major technical challenges that I'm not anticipating?
Thank you!

Comments

  • wjcroftwjcroft Mount Shasta, CA

    Hi Nicholas,

    If you email to (contact at openbci.com), which is customer support, they can sell you a Daisy board to go with your original Cyton. That will bring you up to 16 channels. However it is not currently possible to get 24 channels with the Cyton architecture. This is because ALL channels must be sampled at the exact same moment, under control of the microprocessor on the Cyton board. This cannot be done with TWO Cyton boards. Does that make sense?

    Customer support can also sell a set of 8 additional electrode units and wires, to go with the Daisy module.

    As you can see with the image on the Ultracortex Shop page,

    https://shop.openbci.com/products/ultracortex-mark-iv

    There are already 35 site locations. So you have much flexibility to position electrodes anywhere on the head.

    Regards, William

  • edited February 15

    I'm still not sure why we can't synchronize two Cyton boards. What if I create a synchronization module with a clock that sends out a timing signal? This signal could go to both Cyton boards at the same time, allowing them to start sampling data from all channels simultaneously. If latency is an issue, then I could implement a check. The two boards would communicate to each other: if the other board has not already received a timing signal then that board won't begin sampling. Another option, although it isn't ideal, would be to use both Cyton boards at the same time without a synchronization module and then in post-processing align the data they collect. For this to work, each piece of data would need to have an exact timestamps: this could be supplied by an external clock if these timestamps aren't already generated by the boards themselves.
    Best,
    Nicholas

  • wjcroftwjcroft Mount Shasta, CA

    Feel free to experiment as you like. But a Rube Goldberg setup like you envision would be problematic and a heck of a lot of work. Massive firmware mods would be required. The OpenBCI GUI has no capability to merge two data streams. 'Timestamps' as you suggest do not exist. The current timestamps recorded in the log files are assigned as the Brainflow / GUI receives the packets on the laptop, they are not generated in the firmware.

    Is there some reason that 24 channels is of significant value to you, that 16 channels is insufficient? Yes, commercial / clinical neurofeedback QEEG systems use 19 channels, but they also cost around $10K when all software and hardware is accounted. There are other vendors of EEG / BCI equipment that offer higher channel counts. The OpenBCI team is envisioning a potential V2 Cyton system that may be expandable. But they are focused on the Galea project currently and Cyton work is on the back burner.

    I'll mention two other non-OpenBCI systems that may offer alternatives, but support for both of these is extremely minimal: they are essentially one man projects:

    At the CrowdSupply website, look into the two projects, HackEEG or FreeEEG32.

    Regards, William

  • edited February 15

    I'd like to use an EEG headset to communicate a control signal or error-related potential to a robot. As well, I want to use it to provide observations for a partially observable Markov decision process. So, having 24 channels is not itself significant to me. I only want the option to expand the channel count if 16 or 8 channels prove insufficient for my objectives.
    I'll look into HackEEG and FreeEEG32. Thank you

  • wjcroftwjcroft Mount Shasta, CA

    If you are trying to use some type of BCI to control a robot, look at the free open source cVEP system called MindAffect. It is very versatile and only requires a few channels. 'More' channels is not always 'better'. QEEG mapping uses 19 channels so that they can do both surface EEG and 3D source localization inside the skull.

    https://www.google.com/search?as_q=mindaffect&as_sitesearch=openbci.com

    https://mindaffect-bci.readthedocs.io/en/latest/

  • wjcroftwjcroft Mount Shasta, CA

  • Thank you! You have been extremely helpful, and this seems extremely useful. I'm glad you can run MindAffect's software on Raspberry Pi since I'd like to somehow use ROS along with this software if they're compatible with one another.

  • wjcroftwjcroft Mount Shasta, CA

    Pay particular attention to their suggested headset / band, which uses water dampened sponge electrodes. Because the cVEP brain signals are fast, complex and relatively weak, 'wet' electrode systems are much preferred to completely dry. You could conceivably weave absorbant material into the Ultracortex comb pins, but probably better to follow the MindAffect tips.

  • I see. The documentation MindAffect provides isn't clear on what kind of electrodes they use. Do you happen to know if it would be fine if I were to use OpenBCI's gold cup electrodes to make the water-based electrodes that their headset calls for?

  • wjcroftwjcroft Mount Shasta, CA

    Paste based electrodes such as the cups work fine. So would an electrode cap using gel. The big advantage of their suggested electrode band is that there is no 'mess' to cleanup as with paste or gel. Only water is used in the sponge inserts. They do describe the electrode setup here,

    https://mindaffect-bci.readthedocs.io/en/latest/printing_guide.html

    There is a long history of no-cleanup electrodes. I did a page on this some time ago,

    https://sites.google.com/site/biofeedbackpages/velcro-saline-sensors/commercial-saline-sensors

    The Greentek saline system is listed in the Shop:

    https://shop.openbci.com/products/gelfree-bci-cap-kit

  • edited February 18

    In the documentation you provided I don't see anywhere where MindAffect mentions where to buy the electrodes they use. They only mention that that they are TSMi electrodes. I looked on the TSMi website and they don't have a product page specifically for water-based TSMi electrodes. I emailed MindAffect for a part-number or web-link where I can buy the electrodes they used; I also emailed TSMi inquiring if they sell water-based TSMi electrodes apart from any other product.

    I looked at the page you did some time ago. On brain-trainer.com I found these electrodes listed:

    https://brain-trainer.com/product/reusable-silver-eeg-cup-electrode/.

    Could I use this alongside some sponges (https://brain-trainer.com/product/sponges-for-tc26/) and a custom 3D print to make my own water-based electrodes?

    Alternatively, do you think placing a sponge (https://brain-trainer.com/product/sponges-for-tc26/) onto the dry comb electrodes provided by OpenBCI would be sufficient to make a water-based electrode, or would need to thread a sponge through the prongs of this comb for the best signal to noise ratio?

  • wjcroftwjcroft Mount Shasta, CA

    I'm not sure you are understanding the Brain-Trainer cap. It is a saline based system, used with ordinary cup electrodes and their sponges dampened in saline (salt water) solution. The Greentek is another example.

    MindAffect never sold their headband. You'll just have to do with something you piece together. But any wet electrodes will work, paste, gel, saline system. The idea about adding sponges or absorbent cloth to the 5mm combs may work. You may encounter issues with the sponges falling out when donning the headset. Possibly some type of hold-on clip could help.

    I don't think you have to buy sponges from a specific place, almost any type that works would do. The TC26 may be a good compromise.

    https://brain-trainer.com/shop/products/caps-sensors-supplies/

    The advantage of any wet based system is that the skin impedance is much lower than with passive dry electrodes. This boosts signal amplitude.

Sign In or Register to comment.