Processing Cyton Infra Low signals

compujohnnycompujohnny
edited June 2018 in Software
I am building an Infra Low Neurofeedback design with Bioera and Cyton, similar to Cygnet Advanced Video player, I am sending the values to a custom HTML player that displays a video with a frame that resizes with the signal, but the signal I am getting is a bit confusing,

When I create a low pass filter at 0.1 hz, I get values ranging from 3000 - 4000, for example swings between 3300 - 3450 then back to 3100 - 3500 ...etc,

When I go down to 0.001 hz (1 mhz) the signal amplitude is much lower (I have to set the pipe to float to get a value) starting from 0.00001 then increasing gradually over a few minutes to the single digit number then double digits by 7 minutes into the session

When I set the filter to lower frequency 0.0001 (0.1 mhz) it takes even longer to get to single digits and double digits, and so on for lower frequencies

Currently at Cygnet they can go as low as 0.001 mhz, if I set the filter to that low I have to wait for maybe an hour to get a number that I can work with

But if I multiply the signal after filtering with a factor, it shows much faster but that factor can get very large as you go lower (for 1 mhz I multiply with 10,000, and for 0.1 mhz 10 Million ...etc to get the signal to single digits within 2 - 3 minutes)

and I am not sure if that's the right way to process the signal, any help or insight is greatly appreciated

Comments

  • wjcroftwjcroft Mount Shasta, CA
    edited June 2018
    The Othmer Method has a long history and is quite intricate; usually requiring training in the method and appropriate equipment. Have you experienced ILF yourself and do you know what you are attempting to emulate? 


    Various optimization bands have been known to achieve results, some faster than others.


    Mark Smith does an entirely different form, his term is ISF instead of ILF.


    Regards,

    William

  • compujohnnycompujohnny
    I have watched the On Demand Intro course and the Advanced Summit and I understand the method quite well, however I can't afford and don't have the qualifications required to buy the Neuroamp/Cygnet, so I decided to build my own system, I am a software developer and electronics hobbyist and I have read a lot about the subject of EEG and chose the OpenBCI Cyton because it's a Full band EEG device and doesn't have a High pass filter that cuts off the Infra low signal, also because it samples the signal in 24 bits so it can record from a fraction of a microvolt to thousands of microvolts,

    I also chose Bioera because it's the engine behind Cygnet so it's capable of achieving the technical aspects

    Regarding my question about the implementation of the Infra Low signal processing, do you have any idea why the signal at these low frequencies is looking like this? also is it acceptable to multiply it to get it to a usable number that I can pass to the Custom Video player to adjust the frame? or is there a better way?

    Is the signal really like this or am I doing something wrong in Bioera?

    The information mentioned in the Course and the Advanced Summits mention that the signal starts at 0 and increases slowly over a few minutes, and that it moves only in one direction as it takes days to finish a full cycle at these low frequencies, but if I leave the signal alone it doesn't reach single digits in minutes but it can take a lot of time, so is multiplying it by a factor correct? or is there a better way to process this signal? 
  • wjcroftwjcroft Mount Shasta, CA
    Since both the Othmer ILF Method and Mark Smith's ISF methods are proprietary protocols, we can never know the exact details of their implementations. One can only make hypotheses from the published research papers.

    The ILF / ISF signals are NOT unidirectional. You are probably thinking of the SCP Slow Cortical Potentials paradigm, which differs from ILF/ISF. SCP was the predecessor that inspired the Othmers to create ILF.

    Some of the protocols working in this domain DO have a sensitivity control, that is adjusted in session by the practitioner to keep the rate of change of feedback to appropriate levels. Some adjust sensitivity automatically, others manually. Some of the protocols are more sensitive to the ORF, optimal reward frequency, others less so.

    You may want to read more of Siegfried Othmer's / Mark Smith's papers for more details on the mechanisms behind this form of training, especially how it is not traditional band based neurofeedback. And how it works with bipolar montages. These placements are covered in Sue Othmer's Protocol Handbook.

    Regards,

    William

  • wjcroftwjcroft Mount Shasta, CA
    edited June 2018
    One key concept in all of these infra slow / low protocols, is that the incoming EEG filtered low band signal, is simultaneously tracked with a fairly long cycle moving average. The feedbacks are then based on how the realtime signal goes above or below that baseline moving average. The specifics of how this is done differs for each particular inventor / implementation.

    Some feedbacks are audio only, some using video. Some using audio to track the ILF movements, and video for inhibit feedback. If you look at the papers, inhibits are generally placed in the 1-40 hz range (using various rolloff schemes) to damp down more typical EEG / cognitive activity. Such that the session concentrates on the infra low segment, with "no efforting", just allowing the central nervous system to witness and learn from it's level of glial network activity, reflected in the ILF signal. The glial support network and neurons are in close synergistic relation. And one aspect of the glial voltage potential influences how sensitive are the neuronal synapses. By witnessing it's own level of activation, the system as a whole can achieve more resilience and self-regulation in the face of stressors.
  • compujohnnycompujohnny
    I don't mean that the signal is Uni directional, I mean that it's Uni-directional within the 30 min session as it takes hours to days to complete a full cycle, I have read the protocol guide and I have done/witnessed many sessions and I understand the concepts you're referring to

    I understand the inhibits concept, they are using multiple inhibit bands, and combining the results logically to inhibit any sudden increase in any EEG band

    My question is more related to the practical implementation, did you get a chance to try filtering the signal coming from Cyton at a Low pass 0.1 milli hz and 0.01 milli hz and 0.001 mhz? Are your results similar to what I am seeing? or am I doing something wrong? What are your suggestions for processing such signal, do you multiply it by a factor to scale it to a usable range quicker? or is this approach wrong and what are your suggestions?
  • compujohnnycompujohnny
    could it be a time constant (time response) of the filter since the frequency is very low, the time response of the filter is (1/2πf) could reach hundreds to thousands of seconds, is this correct?
  • BillhBillh
    Compujohnny,

    Are you taking a low frequency band with a high and low, or just everything below your lowpass filter? 
    Do you have a bottom number for your signal band?





  • compujohnnycompujohnny
    edited June 2018
    No it's a low pass filter so everything below 0.01 mHz for example
  • compujohnnycompujohnny
    edited June 2018
    ok I think I understand the reason, the time response of the filter at such low frequencies is very large reaching tens and hundreds of thousands of seconds, that is why the signal amplitude is increasing very slowly over hours, so to make it reach usable levels for Neurofeedback in a few minutes the only way I found is to amplitude scale the signal before filtering so it's very large and can reach single digit numbers faster at such low frequencies,

    Is there a better approach to handle this problem? or is this approach acceptable? amplitude scaling doesn't attenuate the signal and doesn't change it's time properties
  • BillhBillh
    Why not also use a highpass filter set a a very low frequency as well as your lowpass filter? It is a frequency band after all.

  • compujohnnycompujohnny
    I am not sure what an additional high pass filter will accomplish? going lower in frequency for Infra Low Neurofeedback purposes means limiting higher speed signals that may cause arousal symptoms such as anxiety ...etc and sticking to slower signals that are not agitating to that particular individual, so for that purpose I think just a low pass filter will accomplish the task

    Can you elaborate why you think a high pass filter is needed?
  • BillhBillh
    I would remove the problem of slow voltage drift over time in the signal. Such a drift may be purely due to electrochemical changes at the electrode surface and does not generally reflect a true brain voltage change. So why not mask it off with a filter?


  • BillhBillh
    I read your question again, and it occurs to me that you might be unsure of what a low-pass versus a high-pass filter does. It may actually be the opposite of what some might think :).  

    Wikipedia: "A high-pass filter (HPF) is an electronic filter that passes signals with a frequency higher than a certain cutoff frequency and attenuates signals with frequencies lower than the cutoff frequency. The amount of attenuation for each frequency depends on the filter design."

    You are already cutting out the higher frequencies with your lowPASS filter. The highPASS filter frequency would be set well below the lowpass filter frequency, to kill the really slow voltage drift out and hopefully return your baseline voltage to 0 at the start.

    <---------------------  lowpass frequency setting ----------------------->

    wave frequencies you keep are here

    <-------------- highpass frequency setting ------------------------->

    ... where highpass frequency setting < kept wave frequencies < lowpass frequency setting.

    You can get those same really slow (< 0.05 Hz) pseudo-delta-waves just by attaching the electrodes to the backside and using your existing filter, after all :)


  • BillhBillh
    Here is an example of a study which used a 0.1 Hz lowpass and a 0.001 Hz highpass filter (actually, implemented as a bandpass filter):


    Look at the red line in Figure 1, a one hour recording.


  • compujohnnycompujohnny
    edited June 2018
    You may be correct about the high pass modification, they do mention in the Othmer method research that they use both a high pass and low pass filter and consider the training frequency the middle frequency between both cutoff frequencies, so to train at 0.01 mhz I set a band pass at 0.015 mhz max frequency and 0.005 mhz min frequency

    Is that correct? and what do you suggest the filter order and ripple be? the defaults of Bioera are 4 for the order and 0.5 for the ripple
  • BillhBillh
    I would expect that would work.
  • compujohnnycompujohnny

    I found out that a low pass filter leads to a better training effect than a band pass filter, maybe because the signal has more information in it, so I reverted back to a low pass filter

Sign In or Register to comment.

© Vanilla Keystone Theme 2026