4 ms·

What's the technical problem in measuring a differential signal at a 1000 samples/sec, plus removing noise and amplifying it ? It's a common analog electronics

by petra 8y ago

What's the technical problem in measuring a differential signal at a 1000 samples/sec, plus removing noise and amplifying it ? It's a common analog electronics problem. And if it's not technically complicated, why does nobody do this ?

lettergram 8y ago
It's a very difficult problem, first lets just look at data throughput: 1000 samples/sec, 4 bytes per sample, 32 nodes - that's 128Kb / second. Not crazy by any means. However, you then have to process that... That's doing an FFT + binning (in the case of brain waves). All doable at probably 60 FPS (where a frame is the prior 1 - 3 seconds). Still not a problem, then try to do error correction! You have 32 sensors, what's real, what's an artifact? That's when things slow way down typically. This type of analysis can be done after the fact and you essentially guess at errors. Removing known issues or more likely from a research perspective they just mark those sections of the data and remove them. Here's where what we built comes in: http://synaptitude.me/blog/using-computer-vision-to-improve-eeg-signals/ http://synaptitude.me/blog/using-computer-vision-to-improve-... The above is all about post processing, however, I managed to improve the method and write a proof-of-concept for a real-time application that removes artifacts in EEG signals, BUT leaves the underlying signal (AKA you don't need to throw away the whole sample). This can function to capture error corrected brain waves at 50 Hz with 512 samples per second + a 50hz video stream. Probably could do more. This is extremely technically complicated and probably took a year or so to develop. This enables real-time stream (50Kb / second) to our hosted server and real-time applications: http://synaptitude.me/blog/view-brainwave-data-thinksuite-io/ http://synaptitude.me/blog/view-brainwave-data-thinksuite-io... Unfortunately, I couldn't find someone willing to support the work further, and even though we had maybe 30 people vouching $300 each. That wasn't enough to continue the work. Most of this work concluded in early 2017. For reference, I may apply for an SBIR grant coming up to fund additional work on this (my wife started graduate school, utilizing EEGs, so it aligns pretty well). Honestly, if anyone is interested in this please let me know. Its an area of interest, just couldn't justify the expense or time investment (now working on this: https://hnprofile.com/ https://hnprofile.com/)
- bakul 8y ago
  http://web.mit.edu/6.111/www/f2017/handouts/FFTtutorial121102.pdf http://web.mit.edu/6.111/www/f2017/handouts/FFTtutorial12110... shows an FFT implementation using a Cyclone EP2C20 FPGA. It can do 50ksps, 1K points, 12bit precision, 100 µsec/1K complex FFT @ 50Khz clock. May be a Zynq can be used to build a standalone device? I wonder if the EEG signals can tell if your attention wanders. That may help improve meditation. Have you looked at/played with the Muse or Muse2 EEG headband?
  - lettergram 8y ago
    That's exactly what we were doing (without an FPGA), you need at least 8 sensors to properly localize attention though (based on our experiments). Things like the Muse I highly doubt work effectively
    - Erlich_Bachman 8y ago
      Perhaps you should try it instead of "doubting" without any basis. Based on my experience Muse is a great product and it delivers exactly what it says. It's hard to see any other EEG product competing with it any time soon (assuming similar cost usage style. Of course something like what "40 years of zen" are doing will be much more powerful).
    - bakul 8y ago
      Did your experiment show a measurable signal difference between stable vs wandering attention? Mindfulness supposedly helps with executive dysfunction (either due to frontal lobe brain injury or ADHD). It would be an interesting experiment to see if feedback via such a headband can help improve focus.
    - shaklee3 8y ago
      Coming from the RF world, those numbers honestly seem trivial for processing. Is it just a money issue or is nobody willing to build something like you did?
      - 8y ago
nabla9 8y ago
Oversampling removes only uncorrelated noise. Sampling 200 - 250 times times per second per electrode is enough for basic gamma. Gamma is problematic to measure because: * Power of the EEG signal decreases with frequency. Skull and skin muffles gamma frequencies effectively. * Gamma overlaps perfectly with muscle activity (~20–300 Hz). It's easy to measure muscle artifacts instead of gamma. Meditation relaxes muscles. Unless you are able to recognize and remove these artifacts, you are actually measuring how meditation relaxes muscles. * Typical research in gamma EEG setting records thousands stimuli-response epochs to find the signal by averaging over these epochs. Biofeedback applications need to get the signal from single epoch. The latter is much harder problem and generally just not done. If you could drill a hole into the skull and place electrodes inside, things would be much easier.

by petra 8y ago

lettergram 8y ago

It's a very difficult problem, first lets just look at data throughput: 1000 samples/sec, 4 bytes per sample, 32 nodes - that's 128Kb / second. Not crazy by any means. However, you then have to process that... That's doing an FFT + binning (in the case of brain waves). All doable at probably 60 FPS (where a frame is the prior 1 - 3 seconds). Still not a problem, then try to do error correction! You have 32 sensors, what's real, what's an artifact? That's when things slow way down typically. This type of analysis can be done after the fact and you essentially guess at errors. Removing known issues or more likely from a research perspective they just mark those sections of the data and remove them. Here's where what we built comes in: http://synaptitude.me/blog/using-computer-vision-to-improve-eeg-signals/ http://synaptitude.me/blog/using-computer-vision-to-improve-... The above is all about post processing, however, I managed to improve the method and write a proof-of-concept for a real-time application that removes artifacts in EEG signals, BUT leaves the underlying signal (AKA you don't need to throw away the whole sample). This can function to capture error corrected brain waves at 50 Hz with 512 samples per second + a 50hz video stream. Probably could do more. This is extremely technically complicated and probably took a year or so to develop. This enables real-time stream (50Kb / second) to our hosted server and real-time applications: http://synaptitude.me/blog/view-brainwave-data-thinksuite-io/ http://synaptitude.me/blog/view-brainwave-data-thinksuite-io... Unfortunately, I couldn't find someone willing to support the work further, and even though we had maybe 30 people vouching $300 each. That wasn't enough to continue the work. Most of this work concluded in early 2017. For reference, I may apply for an SBIR grant coming up to fund additional work on this (my wife started graduate school, utilizing EEGs, so it aligns pretty well). Honestly, if anyone is interested in this please let me know. Its an area of interest, just couldn't justify the expense or time investment (now working on this: https://hnprofile.com/ https://hnprofile.com/)

bakul 8y ago
http://web.mit.edu/6.111/www/f2017/handouts/FFTtutorial121102.pdf http://web.mit.edu/6.111/www/f2017/handouts/FFTtutorial12110... shows an FFT implementation using a Cyclone EP2C20 FPGA. It can do 50ksps, 1K points, 12bit precision, 100 µsec/1K complex FFT @ 50Khz clock. May be a Zynq can be used to build a standalone device? I wonder if the EEG signals can tell if your attention wanders. That may help improve meditation. Have you looked at/played with the Muse or Muse2 EEG headband?
- lettergram 8y ago
  That's exactly what we were doing (without an FPGA), you need at least 8 sensors to properly localize attention though (based on our experiments). Things like the Muse I highly doubt work effectively
  - Erlich_Bachman 8y ago
    Perhaps you should try it instead of "doubting" without any basis. Based on my experience Muse is a great product and it delivers exactly what it says. It's hard to see any other EEG product competing with it any time soon (assuming similar cost usage style. Of course something like what "40 years of zen" are doing will be much more powerful).
  - bakul 8y ago
    Did your experiment show a measurable signal difference between stable vs wandering attention? Mindfulness supposedly helps with executive dysfunction (either due to frontal lobe brain injury or ADHD). It would be an interesting experiment to see if feedback via such a headband can help improve focus.
  - shaklee3 8y ago
    Coming from the RF world, those numbers honestly seem trivial for processing. Is it just a money issue or is nobody willing to build something like you did?
    - 8y ago

nabla9 8y ago

Oversampling removes only uncorrelated noise. Sampling 200 - 250 times times per second per electrode is enough for basic gamma. Gamma is problematic to measure because: * Power of the EEG signal decreases with frequency. Skull and skin muffles gamma frequencies effectively. * Gamma overlaps perfectly with muscle activity (~20–300 Hz). It's easy to measure muscle artifacts instead of gamma. Meditation relaxes muscles. Unless you are able to recognize and remove these artifacts, you are actually measuring how meditation relaxes muscles. * Typical research in gamma EEG setting records thousands stimuli-response epochs to find the signal by averaging over these epochs. Biofeedback applications need to get the signal from single epoch. The latter is much harder problem and generally just not done. If you could drill a hole into the skull and place electrodes inside, things would be much easier.