Differences between the Cyton and the Ganglion

lagidigulagidigu
edited March 2017 in Hardware
Hey there!

I read the FAQ and the technical specifications about the Ganglion as well as the Cyton, and there are a couple of things that I am not quite sure about:

- The Cyton resolution is 32 bit per channel, am I right? What about the ganglion, I couldn't find anything on that...
- Are there any other differences next to the channel count and the temporal resolution (250 vs 200 hz sampling) between the cyton and the ganglion? 
- What would then be the difference in simply purchasing 2 Ganglions instead of one cyton next to the differences mentioned above?

Thanks a lot in advance for any help :D

Regards,

Luc

Comments

  • wjcroftwjcroft Mount Shasta, CA
    No, both are 24 bits samples.
    Cyton uses a serial COM port and FTDI based dongle. Ganglion a Bluetooth Low Energy link; interfacing at low levels to Ganglion is more complex because of the BLE compression and packing, etc.
    Cyton is expandable to 16 channels (and possibly more with the wifi shield being developed; which will also support faster sample rates.)

  • lagidigulagidigu
    To what degree would the signal quality be worse on the ganglion than on the cyton then? I don't want to make the mistake of purchasing a board that has an insufficient quality for the project that I want to work on... 

    I'm really bad with the technical aspects of the hardware :/


    Thanks a lot in advance! :D

    Cheers
    Luc
  • wjcroftwjcroft Mount Shasta, CA
    I did not say anything about the comparison between Cyton and Ganglion signal quality. There is a paper for Cyton, but not yet for Ganglion. Ganglion should be very close to Cyton in terms of signal quality, but not yet any published measure of this.


  • ls870825ls870825 San Diego
    edited March 2017
    I am debating whether to buy Ganglion or Cyton development board? Apart from the difference in the number of channels, does the signal quality (SNR) differs between Ganglion and Cython. I noticed they are using different analog front-end, though with similar specs.
  • wjcroftwjcroft Mount Shasta, CA
    Ls, hi.

    Merged your new thread "Ganglion vs Cyton" into this existing thread.

    Regards,

    William

  • newdendritenewdendrite USA
    @wjcroft, could you confirm my understanding (below) of the resolution difference between the Cyton and Ganglion?

    While the Cyton and Ganglion both sample a 24-bit word, the Ganglion compresses that data and transmits only 18-bits.  During worse case conditions, the Ganglion will lose 6 bits of resolution.  In many circumstances, this does not occur and the data transmitted will be lossless, but if the signal contains large and fast amplitude changes there will be a loss of resolution.

    Thanks,
    Michael
  • wjcroftwjcroft Mount Shasta, CA
    Mentioning Joel @biomurph and AJ @pushtheworld.

    In the 'old' days, EEG amps only had 16 or even 10 bits of resolution. At least on the Cyton, the low order bit represents .02 microvolts of EEG amplitude. So the low order bits are for all practical purposes just can be considered noise. Some older amps using smaller resolutions had the low order bit represent 1 microvolt. This is still adequate for most applications. 

    Scale factor on the amps can of course be adjusted programmatically.

    AJ or Joel will have to comment on the lossless quality of Ganglion.

  • newdendritenewdendrite USA
    @wjcroft,

    Yep, I have several of those EEG amps from the old days.  I'm seeing a minimum noise floor of about 1.4 microvolts rms from the Ganglion here, measured in a faraday cage with all inputs shorted to ground.  There are also some differences in noise floor among channels.

    Nowadays, there's a lot of interest in gamma in the research lit, and gamma amplitude is typically very low, often only a few microvolts, so noise figures are more important than they were in the old days.

    Best regards,
    Michael
  • wjcroftwjcroft Mount Shasta, CA
    Michael, 

    Gamma is more reliably detected as a synchrony or coherence measure. Otherwise gamma can be easily confused with scalp EMG.




    Regards,

    William

  • newdendritenewdendrite USA
    William,

    Agreed.

    But use of synchrony measures doesn't mitigate the need for a low noise channel.  Wideband noise degrades precision estimation of phase in coherence calculations.  It also reduces reliability of most other synchrony algorithms.

    Take a look at synchrony algorithms sometime.  Nearly all of them are calculated over a fixed time interval, and the introduction of this interval increases latency of the feedback signal, which is a problem for neurofeedback.

    Best regards,
    Michael

  • wjcroftwjcroft Mount Shasta, CA
    Here's the Cyton noise graph,


    I'm surprised Ganglion is showing this much noise 1.4 uV. Hope Joel @biomurph comments. There is a Ganglion graph here, but it looks much different than Chip's Cyton graphs,


  • wjcroftwjcroft Mount Shasta, CA
    A related post on the Community section regarding Ganglion noise figures and front end filtering,


Sign In or Register to comment.

© Vanilla Keystone Theme 2025