ThinkPulse active electrodes, Q&A

13

Comments

  • antoanto bang

    Thank You @wjcroft for Replying . I tried the Active Electrode without Gel Pad which means Directly connected on Subject , its not Picking Any signals from Subject that's Why I tried With Gel Pad (I thought may the Contact is not good Enough) just to make sure my Circuit Connection are good , And I tried the Reference electrode as Active electrode too , still its doesn't pick signals. could you share with me any kind of testing document or video with Subject , it will helpful to me for testing my KIT

  • Hi @anto,
    althouth the ThinkPulse Starter Kit for OpenBCI is meant to be used with OpenBCI Cyton/Daisy boards, the sensors can be used with other amps providing the setting is similar and the right power supply and the right connections are provided. Hard to tell what is going wrong from just what you said but I would say that most likely the input voltage range of your amp is too narrow which results in saturated (flat) signal. Polymer dry electrode tend to create a higher DC offset, especially if you use a different material for reference. Try using ThinkPulse for reference as well (not for ground since it has to be passive, you can use a gel pad) or try to reduce the input gain if this is possible. Also, the +/-2.5V should be understood with respect to the amp ground (and not a +/-2.5V coming from a separate power source like an OpenBCI Cyton powered by batteries). Please let us know how it goes. Best. Julien

  • evolution_encodedevolution_encoded East Coast US
    edited July 2023

    ThinkPulse.... need extension cables

    Hello all,

    I just upgraded to a Cyton with the ThinkPulse starter kit. I will not be using all 8 channels at once, I will only be using 2 to 4 channels at a time. My current Ganglion rig has a velcro headband that I made, and I think I want to use a similar velcro set up for the ThinkPulse electrodes.

    But... I just discovered that I'd prefer not to wear the Cyton and I would need extension cabling to do that. Can I get these extensions anywhere?

    Thanks
    Jim

  • wjcroftwjcroft Mount Shasta, CA

    Hi @evolution_encoded,

    I merged your question about extension cables to this Q&A thread. @julienConscious, the ThinkPulse creator, may have suggestions.

    William

  • Thank you @wjcroft
    @julienConscious - hoping to find a solution to the above

  • Hi Jim @evolution_encoded,
    as you know, ThinkPulse sensors use the standard Dupont connectors to be compatible with OpenBCI Cyton board. Each sensor has 3 wires: power +, power - and signal. You can just use standard jumper wires with male/female Dupont connectors (very easy to find) as extensions. That should work OK providing the total length including the extension is under 1m approximately (and that you use tape to secure the connection between wires). This being said, long wires might increase the sensitivity to EM noise, especially in the extension part that will not be shielded. Also, not wearing the Cyton on your head will increase the movement artefacts. If you only use 2 to 4 channels, remember to deactivate the channels you do not use.
    Best,
    Julien

  • Hi @julienConscious thank you for the info. I absolutely understand your recommendations. I will try to use the wires as-is and attach the unit to the headband.

    But if it proves uncomfortable, I will try adding 50 cm jumpers similar to this: amazon dupont jumpers
    ... and if that is too noisy, I can braid the wires which will help somewhat. But my use case is sitting still - no movement, so that should not be too much of an issue.

    Thanks again

  • wjcroftwjcroft Mount Shasta, CA

    Adafruit has MANY different size choices: https://www.adafruit.com/category/306

  • Thanks @wjcroft ... if the short wires don't do the trick I will look at adafruits offerings. I prefer to give a good company like that some business.

  • Hi @julienConscious - I have been using the thinkpulse electrodes and have some questions.

    Sometimes (I am not quite sure of the cause) an electrode will lose signal and go to zero on an oscilloscope type display. Lifting/reseating the electrode on the skin will show some "ringing" on the display, then go back to zero wtihin about 3 seconds. I try moistening the skin and reseating the electrode -- sometimes this works, but sometimes it doesn't.

    I am guessing that skin moisture is an important aspect with this. Is there a guide or suggestions to keep electrode signals steady for ThinkPulse?

  • wjcroftwjcroft Mount Shasta, CA

    Hi @evolution_encoded and @julienConscious,

    EE, we try to keep all ThinkPulse questions on this thread to make it easy for all ThinkPulse users to benefit from Julien's answers. I merged your new thread above:

    "Sometimes (I am not quite sure of the cause) an electrode will lose signal and go to zero"

    into this Q&A thread.

    Thanks everyone.

  • Once again, thanks William, I didn't quite realize it.

    @julienConscious - looking forward to your reply above. I am finding the signal loss confusing and frustrating. Hoping I just missed something and the solution is simple.

  • Hi Jim @evolution_encoded,
    thanks William @wjcroft for keeping the threads organized and consistent.
    If we dismiss any defective sensor and/or wires or connection problems, the most likely reason inmy opinion is saturation. I do not know exactly what your montage is but active sensors and analog to digital converters have input voltage ranges that you need to respect. For example, the ADS1299 has a +/- 4.5V input voltage range at gain 1 but only a +/- 0.1875V at gain 24 I believe. If your raw input signal including DC offset goes over these values, the output signal is "railed" or saturated hence constant: if you apply any high pass filter on a constant value, you get zero. Same input range constraints applies to active sensors that cannot output any voltage outside of their power range. The main reason for going over the voltage input range is a high DC offset which is linked to impedance, quality of contact (especially bias/ref), spatial configuration of sensors and consistency of impedance between sensors among other things. Also, you have to be careful if you use two different electronic systems. I see you use an oscilloscope to monitor and maybe create an input signal on one hand and potentially a Cyton board on the other hand to power up the ThinkPulse sensors: if they do not share the same ground, the "small" input voltage values seen with respect to the oscilloscope ground might be much larger values with respect to the Cyton. So in short, I would suggest to try lowering your gain, make sure you use a common ground across systems and make sure the contact/impedance is good. Hope that helps.
    Best, Julien

  • Hi @julienConscious, sorry in advance for the long post.

    I see that I was sloppy in the way I tried to explain the issue. I will try to clarify the current output, and hopefully you can suggest a next step to try. (I should have said "trace" instead of oscilloscope - I do not have a standalone oscilloscope. I was using the software BrainBay with the "Oscilloscope" trace widget. My apologies for that.)

    For the results I collected below, I used the cyton, with no electrode extensions, one ThinkPulse comb electrode (channel 1) at Pz, one TP flat electrode (channel 2) at ~FzP, and the OpenBCI earclip electrodes to SRB and Bias at earlobes. For the TP electrodes, I double checked that the signal was going to the correct cyton pins, and the connections to the power adapter were: red - upper pin, black - lower pin.
    I prepped my skin by lightly moistening it in the areas for all of the electrode contacts to both TP electrodes and the earlobe clips. There was no excess or overlapping wetted areas that could cause a short.
    Everything shown in this post was on one computer, in one sitting, not removing the electrodes during testing. But, there were different "sessions" as I switched software during the testing.
    Special note: my hair is very short where the comb electrode contacts, and I ensure the teeth are snug against the skin.

    BrainBay: this is the software that I usually use. I use it every day with my Ganglion for meditation feedback.
    The signal ranges for both channels are in the range of -115000 to -185000 uV. When I moistened the skin freshly, they might start in the range of -70000 uV. Obviously way off. (I missed this at first, because my configuration automatically re-centers the signal to zero.
    Unfortunately, I am not sure how to implement Cyton resistance testing in BrainBay, else I would have included it.
    I tried using the Brainflow OpenBCI Cyton driver, as well as the regular OpenBCI Cyton driver. They both show the same results.
    Typical BrainBay signal uV below:

    OpenBCI GUI: I show the results from OpenBCI GUI below. It's mixed.
    Trial 1: Cyton impedance on channel 1 & 2. Impedances are not great but they are pretty good.

    Trial 1: Cyton signal on ch 1 & 2. Pretty good

    However I strongly suspect that the signal trace in OpenBCI GUI is hiding the same type of "re-centering" algorithm that I have done in BrainBay. The signal in the image above is comparable to the signal in my BrainBay configuration while running normally, until it gets saturated.

    Trial 2: Cyton resistances and signal on ch 1 & 2. I tested resistance, then stopped that test, and immediately began a data stream. Channel 1 is saturated, yet the measured impedance just before was fantastic at 3 kOhm. Also notice the railed/not-railed/near-railed conditions in the traces in OpenBCI GUI.

    As an extra bit of info, using electrode extensions gives similar results to the above. I really hope to be able to use these electrodes. Please let me know of any further tests you'd like me to run / info to collect, and I will post results here.

    Jim

  • Hi all - update to the above. In digging through forum posts, I see that the default gain on the Cyton is a likely culprit. The default gain is 24x. Using the SDK commands found here, I modified the gain on ch 1 & 2 down to 8x, which seems to work fine.
    x1040110Xx2040110X
    I am able to run the Cyton pretty much continuously for over half an hour without the saturation I was seeing before. However, the signal is still pretty far from zero. Below, on dry skin: the comb electrode over Pz is at a fairly steady -106000 uV, and the forehead flat electrode is at a steady -71000 uV.

  • Hi Jim,
    that was the answer I wanted to post. With Gain=24, you are limited to +/-187 500 microV voltage input range. Good job in finding out. You have fairly normal DC offset values especially with using earclips. We recommend Gain=8 on the tutorial to avoid any saturation problem. It's pretty easy to change the Gain on OpenBCI GUI but when using another processing software like BrainBay or OpenVibe, the command is less obvious and sometimes not correctly implemented.
    Best, Julien

  • Hi @julianConcious
    I am using ThinkPulse Electrodes with Cython board and Ultracortex Mark IV cap. The data was record on OpenBCI GUI ver5.10
    I am trying to measure the brain activation from the visual cortex so I use channels, P3, P4, PO3, PO4, POz, O1, Oz and O2
    I'm facing a problem with rail and high amplitude signal on all eight channels as shown in the following screenshot.

    I followed the steps mentioned in the document but the problem still persists.
    Could you help me with this?!
    Recording screen

  • wjcroftwjcroft Mount Shasta, CA

    @goragodp, hi. Did you read the comment by Julien just above your post? He suggests setting the gain to 8, as mentioned in the tutorial. The default for Cyton is gain 24, which is too high.

    You misspelled Julien's username, it is @julienConscious.

  • hi @wjcroft
    My bad! @julienConscious.

    He suggests setting the gain to 8, as mentioned in the tutorial. The default for Cyton is gain 24, which is too high.<

    Yes, I try different level of gain from 1x to 8x but amplitude is still high.
    On the lower level (1x - 4x), I achieved ~120uVrms but suddenly the amplitude rose to >6000 uVrms and all channel became rail.
    I also tried circular motion to adjust the cap position and dig electrode's pong deeper into the hairs but I still face the same problem, high amplitude follow by rail.

  • wjcroftwjcroft Mount Shasta, CA

    Is it possible that you may have high EMF electromagnetic fields in your local area? These are generated by such items as: extension cords, power supplies, wall warts, conduits in (walls, ceilings, floors), wireless equipment, cell towers, metallic furniture, etc. First suggestion would be to position yourself in a new area away from these sources.

    Your screen image you shared above, shows very strong pulsing, around 10 Hz, which is usually a sign that you are picking up spurious EMF radiation. This is visible in both the time series and the FFT.

  • @wjcroft
    That may be the case! I tested it in the laboratory room full of extension cords, and all tables were made of metallic material.
    How far should I be from these radiation source?

  • wjcroftwjcroft Mount Shasta, CA
    edited November 2023

    Mains EMF (50 or 60Hz AC sine waves) drops off relatively quickly. Most is gone after a meter or so. However 10 Hz spikes you show is NOT mains, but likely from some equipment nearby. So I have no idea of the field dimensions for your 10 Hz signal. Try another room with minimal equipment.

  • @wjcroft
    Thank you for suggestion, I will try the difference place and see if the problem solve!

  • Hi Everyone,
    sounds like good suggestions indeed. You can also try to deactivate all sensors except pre frontal Fp1 / Fp2 with flat electrodes on the forehead (they are the easiest to get clean and also you can check blinking). Make sure you get a good clean signal with those 2 before progressively adding the other channels: one (very) bad channel can affect all the other ones. I recommend also to double check your bias and ref ear clip sensors: they affect all channels and so it is very important that they have a good contact and that they are correctly connected. Make sure you have the right settings in hardware settings . Keep up posted. Best, Julien

  • The ThinkPulse active electrodes come with two comb tips.
    There are short tips, already installed, and some longer tips, which I'd like to use instead.
    However, there's no "obvious" way to remove the short tips and I found to guide online for how to change them.
    I'd rather not damage the electrode tips since no replacement got shipped, so I'm curious if anyone has insight into how to change the tips or links to a guide (I've tried looking through the docs in vain)

  • Hi George3d6,
    I understand you initially bought the starter kit (that comes with the short prongs, and also flat electrode design for forehead/bare skin already installed into the joint ball sensor casing), and that you also bought the replacement kit with the longer prongs.
    To change the type of electrode design, you "just" have to unscrew the electrode that is installed and screw in the new one. Note that what I refer to as "the electrode" is actually made of 2 parts: the black polymer tip (flexible, spiky or flat), surrounded by a thin plastic ring (with the thread for screwing) that can be black or grey or potentially another color. I suggest you do not try to separate the plastic ring from the polymer tip: it is not necessary (since every electrode is provided already mounted with its plastic ring) and it is easy to tear them while doing so. So when you try to unscrew an electrode, you actually need to get a grab onto the plastic ring.
    It can happen that the electrodes get stuck somehow especially if you have not changed them for a while.
    If unscrewing with bare hands is not possible, you have two options:

    • instead of unscrewing, you can try to unblock by gently lifting up vertically the ring from the casing with your nail or a plastic tool (like a guitar pick), working your way around. It should come loose and then you should be able to finish unscrewing it manually
      OR

    • you can use a pair of pliers but be very gentle, applying the minimum pressure to get a grip on the ring and then give a start at unscrewing, then finishing manually should be easy. It's delicate but it's not going to get damaged if you operate carefully.
      Best,
      Julien

  • gently lifting up vertically the ring from the casing

    What is "the ring" and what is "the casting" ?

    Here are the replacements I have:

    and here are the electrodes as they are:


    There's no particular "screw" structure that I see as being "un screwable". Is there not a vide somewhere of this process being done ? As it stands they seem glued in place

  • Hi George3d6,
    thanks for the pictures that will help clarify. The ring is the rigid plastic part that touches your thumb in the first pic whereas the casing is the joint ball rigid plastic part that you hold in the second pic:

    These parts are not glued together. The normal method to remove the electrode is to press on it from the top and turn. The sensor you show seems really tighly screwed but there is a way to insert a very thin tool (ideally a broad and flat tool like a spatula, try avoiding sharp pointy tools that can be dangerous for you if they slip and could also damage the sensor) between the ring and the casing to unlock it.
    Best,
    Julien

  • Ok, that makes it clear, seems to work but rings are pretty stuck.
    A further question I have is if there's any guide as to how to properly figure out wiring issues/bugs ?
    As of recently, I've been experiencing some strange behaviors where my readings seem to have reduced to just noise (e.g. no alpha waves or occipital changes when I close/open my eyes) + strange patterns where 1-8 will suddenly go from "railed" to "not railed" and, seldom, 9 through 16 will follow the same pattern.
    I suspect this might have more to do with connection errors between the borads and/or powering the active electrodes but it's quite unclear how to go about debugging such issues.

    Would the electrodes send no signal if improperly supplied or just lower amp signal ?

  • This is the kind of pattern I'm getting, for reference, which is why I'm thinking the issue might be power related:

Sign In or Register to comment.