3 ms·

I'm intrigued by the electrodes they claim to be having 4 times better SNR than conventional electrodes. AFAIK there has been numerous attempts at dry electrode

I'm intrigued by the electrodes they claim to be having 4 times better SNR than conventional electrodes. AFAIK there has been numerous attempts at dry electrodes (electrodes w/o gel for impedance control), but due to higher impedance, SNR has always been worse than wet electrodes as a result. The main advantage of dry electrodes has been faster prep time & convenience at a cost of signal fidelity. I'd be interested in technical details, but information on their website doesn't really let on much. Wonder how they actually compare to market leaders in research grade electrodes (e.g. BrainProducts, Neuroscan). The founder's previous research seems to be focused on topics related to consciousness and sleep. I was hoping if he published articles about electrode materials, but can't find any. Maybe it's a trade secret.. Anyone with more info on this? Also, their electrodes seem to be mostly focused on the occipital area. I get the impression they'd be focused more on visual components rather than cognition and decision making (e.g. P300).

emilprotalinski 7y ago
I don't know the technical details but I can tell you the CEO would not disclose the material with me. He did confirm that it's proprietary and that the team worked with chemists to make it. Full disclosure: I'm the article author.
- charliemil4 7y ago
  Super cool and well played making it difficult on the duck hunt... Question: In the duck hunt video, I don’t see a noticeable “flicker” that’s normally associated with visual bci inputs... were the flapping of the wings the stimulus <> cortex hertz sync?? (Or did they even say??) If so, dang clever. If there’s no hertz sync, that’s flippin incredible. Can’t wait for the completely autonomous inputs, without the visual flickery required.
  - tastroder 7y ago
    Although I'm doubtful about most of the claims this product makes, there's quite a bit of research into visual stimuli without that flicker. This 2018 paper for example: https://www.nature.com/articles/s41598-018-24008-8 https://www.nature.com/articles/s41598-018-24008-8 "Highly Interactive Brain–Computer Interface Based on Flicker-Free Steady-State Motion Visual Evoked Potential" (Han et al.) Although that uses a real EEG headset, the article doesn't really convince me that this one will perform that much better than the existing products in the consumer space. Did anybody find publications that evaluate this one in particular?
    - charliemil4 7y ago
      Solid — thank you. To me the visual modulation (shape and contrast) is a clever ux work around. Definitely great for a glasses first environment
skummetmaelk 7y ago
It doesn't even matter what the SNR is. EEG is an extremely low bandwidth signal that measures average activity of BILLIONS of neurons. It's not possible to do anything significant with it. These guys are just capitalizing on hype and they will be able to build some toy demos, but never anything really useful.
- ganzuul 7y ago
  Neurons work in a coordinated manner. Experiments with monkeys shows you can get by with a lot of undersampling. But orders of magnitude still seems to matter.
- bobowzki 7y ago
  As an anesthesiologist working with EEG to monitor anesthesia, thank you for this comment. I wouldn't say you can't do anything significant, it's a bit like listening on a computer with a stethoscope. You can certainly deduce something from fans and psu noise etc, but it's VERY crude.
- phreeza 7y ago
  I wouldn't discard the bandwidth of EEG so quickly. Clinical EEG has low bandwidth (<100 Hz), but in principle you can get up to the kHz range, enough to pick up individual neuron activity. Of course you will pick up mix of many neurons at any given electrode, but if you could ramp up the electrode density, it would make some degree of unmixing possible.

emilprotalinski 7y ago

I don't know the technical details but I can tell you the CEO would not disclose the material with me. He did confirm that it's proprietary and that the team worked with chemists to make it. Full disclosure: I'm the article author.

charliemil4 7y ago
Super cool and well played making it difficult on the duck hunt... Question: In the duck hunt video, I don’t see a noticeable “flicker” that’s normally associated with visual bci inputs... were the flapping of the wings the stimulus <> cortex hertz sync?? (Or did they even say??) If so, dang clever. If there’s no hertz sync, that’s flippin incredible. Can’t wait for the completely autonomous inputs, without the visual flickery required.
- tastroder 7y ago
  Although I'm doubtful about most of the claims this product makes, there's quite a bit of research into visual stimuli without that flicker. This 2018 paper for example: https://www.nature.com/articles/s41598-018-24008-8 https://www.nature.com/articles/s41598-018-24008-8 "Highly Interactive Brain–Computer Interface Based on Flicker-Free Steady-State Motion Visual Evoked Potential" (Han et al.) Although that uses a real EEG headset, the article doesn't really convince me that this one will perform that much better than the existing products in the consumer space. Did anybody find publications that evaluate this one in particular?
  - charliemil4 7y ago
    Solid — thank you. To me the visual modulation (shape and contrast) is a clever ux work around. Definitely great for a glasses first environment

skummetmaelk 7y ago

It doesn't even matter what the SNR is. EEG is an extremely low bandwidth signal that measures average activity of BILLIONS of neurons. It's not possible to do anything significant with it. These guys are just capitalizing on hype and they will be able to build some toy demos, but never anything really useful.

ganzuul 7y ago
Neurons work in a coordinated manner. Experiments with monkeys shows you can get by with a lot of undersampling. But orders of magnitude still seems to matter.
bobowzki 7y ago
As an anesthesiologist working with EEG to monitor anesthesia, thank you for this comment. I wouldn't say you can't do anything significant, it's a bit like listening on a computer with a stethoscope. You can certainly deduce something from fans and psu noise etc, but it's VERY crude.
phreeza 7y ago
I wouldn't discard the bandwidth of EEG so quickly. Clinical EEG has low bandwidth (<100 Hz), but in principle you can get up to the kHz range, enough to pick up individual neuron activity. Of course you will pick up mix of many neurons at any given electrode, but if you could ramp up the electrode density, it would make some degree of unmixing possible.