Expected Impedance range with Ultracortex + Cyton [resolved]

Alexander_ZuestAlexander_Zuest Switzerland
edited April 2020 in Electrodes

I just assembled my Mark 4 (which I ended up printing in ABS as one piece and glue the inserts in and turned out quite well). Now I'm having some trouble with the signal quality. I've seen some recent posts with similar problems, jet couldn’t solve my issues. My channels are FPz, FZ, PZ, C3, CP1, CZ, CP2 and C4. While measuring the impedance in the OpenBCIGui I got 150K,5150K,74K,2600K,4700K,3500K4550K and 2600K which are all a lot more than the okay range of 40-50K. I measured all the values separately. Screenshot of the signal with Eye-Blinks and jaw muscle artefacts see below. Fixing the cables in place helped a lot improving signal quality and at least the artefacts are clearly visible, but the impedances seems a bit concerningly high to me. I’m not jet very experienced with EEG, it might be obvious what to do. Also is it right to connect one ear clip to BIAS and the other to SRB (on cyton)?

Does anyone know what I could do to fix this issue? I have rather dense hair but it's not to long atm. And my Hardware is brand new.

Aside that, is there a way of changing the preset electrode configuration to the one mentioned above?
Any suggestions are appreciated!

Best wishes
Alex

Comments

  • wjcroftwjcroft Mount Shasta, CA

    Alex, can you mention how your reference lead is attached? Since you are seeing high impedance on basically all channels, that suggests an issue with a common shared lead.

    If you are using ear clip, try swapping with the other ear clip. (To eliminate possibility of a damaged wire or contacts.) Also check the surface of the ear clip metal, it should be a dull grey, not 'silvery' looking. Silver chloride electrodes are typically first plated with silver, then chlorided.

    You can also use an alcohol swab to remove skin oils from the reference lead connection.

    Another reference connection point that is sometimes used is the "mastoid bone area" behind the ear. This is usually done with a pre-gelled adhesive electrode pad, as there is no hair in that area.

    Figure C.4
    https://www.ynic.york.ac.uk/guides/megoperators/2010-01A/html/apcs03.html

  • I have the two ear clips attached to the bottom pins of Bians and SRB, their surface is completely grey, so the coating is still intact. In Addition, I checked their connection to the board.
    Unfortunately, I have no additional ear clips at my disposal, so changing them is not an option. Visual inspection of the ear clip cable did not imply any damage. Afterward I've redone the complete fixation of the cabling to reduce the movement as much as possible. All this did not change impedances during my next measurement.
    Also checked some cables in terms of resistance, but everything seems well.
    I can try to clean all electrodes with alcohol and then apply a conductive spray. I still don’t see any reason for the bad values.
    Below u can see my hardware and the cable connections.

  • The range of the impedance remains essentialy the same. Yet the signal did some funny things link being railed at the start of the recording on most channels then turning okay, not stoping the impedance measurements when told to do and being near railed only after the measurement. This could be corrected by restarting the GUI so it might be a software issue.

  • retiututretiutut Louisiana, USA

    The range of the impedance remains essentialy the same. Yet the signal did some funny things link being railed at the start of the recording on most channels then turning okay, not stoping the impedance measurements when told to do and being near railed only after the measurement. This could be corrected by restarting the GUI so it might be a software issue.

    Channels being railed is likely due to physical setup/connections. To help @Alexander_Zuest, I can get connected to my Cyton+Daisy in the next day or so, and share what it looks like for me with impedance values.

  • retiututretiutut Louisiana, USA
    edited April 2020

    @Alexander_Zuest I think you are connecting the earclips correctly per the Docs.

    Can you post a screenshot of the time series with and without impedance check on all channels? See below as an example.

    Here are some screenshots using same hardware (Mark IV + Cyton + Dongle). In the last photo, you can see the impedance ranging from ~150kOhm to ~6000kOhm. I can confirm these numbers are an accurate reflection of how well each electrode is connected to my head. Channels 5 and 6 are having the most problems because the hair is thick and I need to put long-hair electrodes in these positions.



  • I tried to adjust all electrodes for a better fit, jet it seems I was not very successful. As one can see channel 2 and 6 (Fz and Cz) are very noisy. But in terms of impedance even channel 1 and 3 produce very high values. Channel 1 is Fpz on my forehead, i prepared the skin with alcohol. So in theary that should produce a quite good signal...
    Here are the screenshots

    thx!

  • retiututretiutut Louisiana, USA
    edited April 2020

    @Alexander_Zuest Thanks for sharing!

    Channel 1 looks great, but still has higher impedance than mine (~320 > ~160). I should mention that I did not clean my forehead with alcohol before putting on the Mark IV.

    Channels 2 and 6 have poorest connections, identified using impedance check image. This is also reflected in the TimeSeries in the first image.

    @wjcroft Next steps for Alexander?

  • wjcroftwjcroft Mount Shasta, CA

    Richard, thanks for your measurements.

    It might be helpful to have some mention in the Ultracortex docs about nominal ranges of expected impedances. Your measurements would suggest say ~200K at forehead area where flat electrodes give better skin contact. And in the range of ~900K (best case) for areas with hair.

    Alex's initial post stated: "... While measuring the impedance in the OpenBCI Gui I got 150K, 5150K, 74K, 2600K, 4700K, 3500K, 4550K and 2600K which are all a lot more than the okay range of 40-50K...."

    I'm not sure where he got the impression that "okay range of 40-50K". Perhaps you are referring to the EGI paper where they are using a saline (salt solution) liquid electrolyte.

    https://openbci.com/forum/index.php?p=/discussion/2206/what-is-the-relationship-between-high-impedance-on-eeg-channels-and-50-60hz-noise

    http://wwe.eeginfo.com/research/ElectrodeImpedance.pdf [EGI impedance paper]

    Both the EGI and the Greentek cap are again saline systems, where this range of impedance is possible because of the liquid electrolyte is interfacing with the skin (an optimal condition). With completely dry electrodes, one cannot get readings this low (40K).

    If you notice the calculations in the EGI paper, they were done some years ago, with EEG amplifiers that had "input impedances" (at the amp front end), of only perhaps 200 megohms. The lower the input impedance of the amp, the more important is the skin impedance. As they show in that paper with the 200Mohm amps, then 40Kohm skin impedance resulted in no distortion. BUT, in fact in the conclusion they state: "...Although we did not test the signal quality for scalp-electrode impedances above 40 Kohm in this study, the engineering analysis shows that scalp-electrode impedances up to 200 Kohm still allow for accurate (< 0.1% error) signal acquisition..."

    The ADS1299 chip used in the Cyton, has an input impedance of 1 GigaOhm. FIVE times higher than the 200Mohm in the paper. This might imply that (if quantities scale linearly), a skin impedance of 200K * 5 = 1 Mohm, still results in accurate (< 0.1% error) acquisition. And extrapolating further, if a 0.2% error is acceptable, then 2Mohm skin impedance is still valid. For 0.4% error, 4Mohm skin impedance would work. I would venture to say that 4 or 5 megohm impedances are not anything to worry about, with this dry contact system.

    So my general statement is, don't be concerned so much about skin impedance with the dry electrodes, since the amplifier input impedance is so high. But since Richard is able to get impedances in the 200K to 900K range (on most of his electrodes), values substantially higher than this might indicate that there is room for improvement.

    I do know that in the QEEG field with Electro-cap, they advise NEVER to use fancy hair products before a session. The hair should be shampooed, but no conditioners or oils or any other materials should be on the scalp before recording the EEG with the cap.

    Regards, William

  • wjcroftwjcroft Mount Shasta, CA

    The main reason that the impedance measurement capability is present with OpenBCI, is that some research / academic / clinical environments using electrode caps and gel or saline electrodes -- have requirements for the impedance to be below certain levels. Generally dry contact (passive electrode) headsets do not have these same stringent impedance stipulations.

  • wjcroftwjcroft Mount Shasta, CA

    The other comment I would have for Alex, is that his images show a lot of 50 Hz mains noise. As you can see, channels with higher impedance are more subject to the mains noise.

    You should experiment with your positioning of power cords, adapters, extension cords, wall and floor hidden conduits, etc. -- all of these objects generate a mains EMF field, that extends for several feet from the object. So distance is one way to get out of the range of the EMF mains field. If you have other equipment that you can turn off, or experiment with positioning, that may reduce the 50 Hz noise as well.

    Even though impedance is not so critical for signal accuracy, the higher it is, the more prone to mains noise effects you will see. So the obvious solution is to try adjustments to reduce this. Then the higher impedance channels won't show so much coupling of mains.

  • @wjcroft @retiutat Thank you so much for your support!
    I will take your recommendations into account during my trials.

    I got the 40-50K Ohm from here: https://openbci.com/forum/index.php?p=/discussion/comment/13135#Comment_13135
    I think it is also stated elsewhere.

    Alex

  • retiututretiutut Louisiana, USA

    @Alexander_Zuest

    You should experiment with your positioning of power cords, adapters, extension cords, wall and floor hidden conduits, etc.

    I use a ~1 meter USB extension cord with the Dongle to move it away from all of the metal/objects on my desk.

  • wjcroftwjcroft Mount Shasta, CA

    Alex, thanks, sorry again for the 40K impedance confusion factor. As mentioned, this figure comes from the EGI paper. And we have widely cited the EGI work, because it underlines the importance (improved performance) of the 1 gigaohm input impedance at the amplifier front end.

    There are not many mentions of actual expected impedance numbers for dry electrodes occurring here on the forum. I corrected my statement on that link you gave. However, curiously that same link shows quite LOW impedances measured with the OpenBCI headband kit (using similar FRI dry silver chloride electrodes).

    Perhaps because with that setup it is possible to really crank down on the pressure used by tightening the velcro band beyond what is possible with the headset spring loaded contacts. I also believe the headband kit uses a completely FLAT forehead electrode, whereas the Ultracortex forehead uses a 'cone' shape, which has less contact surface area. Not sure what he was using for his rear headband electrodes, the kit comes with both flat and comb, so he was probably using comb. Although it could have been the LONG combs, versus what ships with the headset, which are short combs.

    Regards, William

  • yundayunda Singapore
    edited September 2020

    Hi, I'm having large impedances on my openBCI Cyton+Daisy board. These are dry comb electrodes, used with the Ultracortex Mark IV and openBCI Cyton+Daisy. The combs are brand new, I removed them and screwed them back on and these were the values I was getting. I observed that channels at the top of the scalp have much higher impedances than channels above the eyebrow or at the back of the head.

    Can I check with others who have been using the OpenBCI Cyton+Daisy with dry comb electrodes for the Ultracortex Mark IV - what were the impedance values you were getting? Were you able to pick out P300 ERPs on auditory or visual paradigms?

  • wjcroftwjcroft Mount Shasta, CA
    edited September 2020

    @yunda, hi. I merged your new thread into this existing one covering the same issues.

    Please look at the previous comments. In particular, this one,

    https://openbci.com/forum/index.php?p=/discussion/comment/13328/#Comment_13328

    William

  • TL730TL730 Beijing, China

    In some research and application scenarios, the impedance of EEG signal acquisition is less than 5K Ω. Do you have any good ideas? How to reduce the impedance of dry electrode and wet electrode?

  • wjcroftwjcroft Mount Shasta, CA
    edited September 19

    @TL730, hi.

    I merged your new thread, into this existing thread on expected Cyton dry electrode impedances. Please see previous comments. In particular, this comment and referenced pdf: https://openbci.com/forum/index.php?p=/discussion/comment/13328/#Comment_13328

    William

Sign In or Register to comment.