4 ms·

Brain-computer interfaces are futuristic sounding, but you're still pretty limited to 1D or 2D control. I can see why they have fallback to EMG. Without someth

Brain-computer interfaces are futuristic sounding, but you're still pretty limited to 1D or 2D control. I can see why they have fallback to EMG. Without something invasive like implanted trans-cranial electrodes or non-portable like fMRI, you're never going to get anything past unreliable 1D control with the noise you get from trying to capture highly-attenuated microvolt potentials from the scalp. I built a brainwave-controlled Pong game using EEG for fun for my university project back in 2012: https://www.youtube.com/watch?v=HtJhbqliJYY https://www.youtube.com/watch?v=HtJhbqliJYY

stinos 8y ago
Without something invasive like implanted trans-cranial electrodes or non-portable like fMRI, you're never going to get anything past unreliable The time resolution of any existing fMRI implementation would be way to low for driving anythin mechanical reliably. The implanted electrodes though are being used already (lookup 'Obama handshake robot') and further researched. Very, very interesting stuff. Also depending on the implantation location one can theoretically get around the problem of EEG where people first have to learn to control the robot by randomly trying some thoughts to see what they do (because EEG just picks up everything): e.g. implant in the higher level motor action planning areas directly (well, if functional in the patient) is a better starting point.
- alexgmcm 8y ago
  The electrodes also have the benefit that you can stimulate the neurones as well potentially creating a full feedback loop. It would be interesting to see if the brain could adapt to a new input and learn how to interpret it. Rat experiments seem promising: http://www.sciencemag.org/news/2015/10/brain-implant-lets-rats-see-infrared-light http://www.sciencemag.org/news/2015/10/brain-implant-lets-ra...
  - YvetteBrooks 8y ago
    The delay between inputs of reading every neuron, reading and actually performing a task will be the problem to solve until it opens up totally new and fastest growing sphere for the science.
    - stinos 8y ago
      Imo there are more pertinent problems than delay. When the delays are sufficiently small, say < 10mSec between neural signal and initiation of movement of the robot, which is already possible to achieve now, you get something which is really usable already and feels close enough to actually making a decision and moving your own hand for instance. On the other hand there's longevity of electrodes, making hardware smaller etc. And then, probably biggest thing for now, the software side: developing (possibly self-learning) accurate algorithms for decoding of the signals (both single-neuron spikes + low frequency spectrum of the surrounding area) is just hard. Also because a lot of things about the brain are in essence still rather unknown it's sometimes a bit of programming in the dark.

stinos 8y ago

Without something invasive like implanted trans-cranial electrodes or non-portable like fMRI, you're never going to get anything past unreliable The time resolution of any existing fMRI implementation would be way to low for driving anythin mechanical reliably. The implanted electrodes though are being used already (lookup 'Obama handshake robot') and further researched. Very, very interesting stuff. Also depending on the implantation location one can theoretically get around the problem of EEG where people first have to learn to control the robot by randomly trying some thoughts to see what they do (because EEG just picks up everything): e.g. implant in the higher level motor action planning areas directly (well, if functional in the patient) is a better starting point.

alexgmcm 8y ago
The electrodes also have the benefit that you can stimulate the neurones as well potentially creating a full feedback loop. It would be interesting to see if the brain could adapt to a new input and learn how to interpret it. Rat experiments seem promising: http://www.sciencemag.org/news/2015/10/brain-implant-lets-rats-see-infrared-light http://www.sciencemag.org/news/2015/10/brain-implant-lets-ra...
- YvetteBrooks 8y ago
  The delay between inputs of reading every neuron, reading and actually performing a task will be the problem to solve until it opens up totally new and fastest growing sphere for the science.
  - stinos 8y ago
    Imo there are more pertinent problems than delay. When the delays are sufficiently small, say < 10mSec between neural signal and initiation of movement of the robot, which is already possible to achieve now, you get something which is really usable already and feels close enough to actually making a decision and moving your own hand for instance. On the other hand there's longevity of electrodes, making hardware smaller etc. And then, probably biggest thing for now, the software side: developing (possibly self-learning) accurate algorithms for decoding of the signals (both single-neuron spikes + low frequency spectrum of the surrounding area) is just hard. Also because a lot of things about the brain are in essence still rather unknown it's sometimes a bit of programming in the dark.