Slashdot Mirror
The Internet Meets the Neural Net
orangesquid writes "OpenEEG is a system for getting data from your brain to your computer. Recently, work was resumed on scEEG, a soundcard-based system which may one day make home EEG systems very cheap (they currently cost a few hundred US$ to put together; there are, though, some potential cheaper alternatives). But, what research is being done into getting data from your computer to your brain? There have been some systems that inject optical signals into your eyes, but, what about direct neural interfacing? It seems EMF and light are one option; playing with neurotransmitters may be another. What do /.'ers foresee coming in this field? What research have you seen being done? Particularly, is any of this to the point where homemade, low-cost systems are feasible? Where can I find out how to inject signals into my head? Combining this with openEEG might lead to some exciting new levels of Internet addiction."
Another issue would be the following. Do we really want something to interact directly with our brain? Imagine if you will the world of tomorrow, where we do have direct nueral nets. Now imagine the equivalant of a computer virus but going through said nueral net. Instead of formatting a C:\, we have something that erases short term memory, or long term memory, or induces schizophrenia. I remember an old outer limits episode where society was connected to a nueral net. Any data you requested was instantly put into your mind, such as a book like Moby Dick. Eventually the computer that ran the system had a bug that started demanding more information input from the people. Stupid things like the number of grains of sand on the beach, etc. This eventually drove the population insane, even with the failsafes they built into the unit. The moral of the story is don't hook anything up direectly to the brain!!
Whatever. As someone who has some experience with neurobiology, it is very unlikely you can actually get anything useful from EEG signals. They are totally useless for exploring the brain's functions. Sorry to burst your bubble, but even microelectrode arrays implanted directly into the brain don't provide anything too interesting. At best, you might see a different firing pattern based on external stimulus, and even that is rare.
Early tests with monkeys indicate great progress in controlling a robotic arm with a monkey.
... the picture on the screen changes position according to the monkey's mind. A robotic arm is also controlled, and the screen and arm basically do the same things as a result of the monkey "wanting" it to. ... and that's just a monkey!
The monkey is in a room looking at a screen, and thinks
Google it: robot arm monkey brain
is available here. one of the most promising techniques seems to be self-assembling nanowires and sensors running through the blood vessels of the brain. lots of facinating reading.
What Future?
Yeah, shortcutting around the 20Hz filter seems a little silly. I did a fair amount of work with this EEG business not too long ago, creating an ad-hoc sensor network, indegrating ADC directly on the electrode. (Available here) These ADCs cost less than 20 bucks a channel, (total, with COM and everything) and in even modest quantites could be made for much much cheaper. I think there is simulaneously too much excitement over the possibilites of EEG, and too little work done to further the technology for both medical and nonmedical purposes.
Now, getting data into the brain, that would be easier. We have two main, "high bandwidth" conduits for input; the eyes and the ears. First off - look up "brainwave stimulation", "light and sound", etc - here's a few links:
Hack Canada's Brain-Wave Machine
Futuremind Light & Sound
Neural Signals, Inc
There are other projects out there as well - just google, and you will find them.
Also - look into "Neurophone" and "Voice to Skull" technologies - these use two systems: ultrasound and microwave. Of the two, microwave seems to offer direct neuron stimulation. Basically, on both systems, a carrier wave is set up and voice is FM modulated on top of the carrier wave. The signal is beamed to the subjects head. In the ultrasound version, the skull filters out the carrier wave, leaving the original signal, and bone conduction allows the subject to "hear" the original sound. In the microwave system, the brain itself does the filtering, and the brain then reconstructs the sound. Both systems suffer from major drawbacks in sound quality. Both versions are patented. NASA at one time was interested in the research. Basically, to the subject, it sounds like voices are speaking in their heads - and in the case of microwaves being used as the transmission medium (the research originally started when radar and microwave technicians reported hearing "clicking" type noise whenever they worked on live equipment), it makes you wonder about wearing tinfoil hats (hmmm). I know that the ultrasound version has recently been used as a testbed for "beaming" custom music or advertisements to people on an individual basis - I know /. stories have reported on this in the past (heck, you will find my comments in them on voice to skull).
Anyhow - once you have a couple of ways to get data into the eyes and ears (and/or vestibular system) - and note, a good quality HMD could be used as a light/sound device - you then can play. I can see using the sound part to play music, and underneath the music have the sound binaural beat doing the brain-wave thing (basically, what you do is inject two different audio signals into the ears - say the left at 30 Hz, and the right at 36 Hz - which will yield a "beat tone" of around 6 Hz, which will make you drowsy, etc). Get the sound going, and sync up the eyes in a similar matter, to the sound. Maybe monitor (via IR leds and a camera) the eyes, see what they do, and if you can tell when you are in the meditative state - then alter the sound and/or visuals to force something different (say, ramp slowly from 10 Hz to 6 Hz - then hold at 6, then ramp quickly up to 7-8 Hz, injecting crazy patterns into the eyes - if using an HMD, maybe something like a visualization hack).
Another thing or possibility would be the idea of computer controlled or directed lucid dreaming, via a brain-wave system - imagine donning the goggles and headphones, lying back, listening to a relaxing audio CD as the computer drops you down to a lucid dreaming state, then starts putting suggestions into your ears and eyes, suggesting and guiding a lucid dream (perhaps the computer could also monitor breathing rate, skin conductivity, etc - to help control the "dream")...
Fascinating thoughts and ideas...
Reason is the Path to God - Anon
They are totally useless for exploring the brain's functions
Perhaps you won't be doing cutting edge neurology research with this kind of EEG interface to your computer but it's far from useless. Basic analysis of the spectrum of the signal is not so hard. The dominant frequencies correlate to states of consciousness such as relaxation or concentration.
It's fun. It's fascinating to watch your own brain in action. It can even be potentially useful as a biofeedback tool.
Stop worrying about the risks of nuclear power and start worrying about the risks of not using nuclear power.
Beta waves (about 15 Hz to 40 Hz) are associated with concentration and calculation. Alpha waves (about 8 to 11 Hz) are associated with calmness, creativity, and being "in the zone". Theta waves (about 5 to 7 Hz) are associated with hypnogogic states: drowsiness and the state of just waking up or falling asleep. Delta waves (about 1 to 3 Hz) are deep sleep waves.
To address your initial point, ADD and related disorders may be connected to a deficit of alpha brain waves. Although the increased beta activity leads to concentration in small bursts, a lack of calmness leads to great distractibility. ADD can be helped by techniques that enhance one's ability to generate alpha waves. Alpha waves are constantly being generated by the thalamus but can be swamped by other brain activity by time they arrive at the cortex.
All of these brainwaves can be trained for with neurofeedback training. I have received training at the Biocybernaut Institute and I can recommend it as quite remarkable on several levels.
Not surprisingly, caffeine stimulates beta activity and suppresses alpha and theta. Meditation increases alpha ability and advanced meditators get into increased theta while maintaining wakefulness.
Beware of entrainment systems that channel the brain into fixed frequencies with blinking lights or beat frequencies. Proper brainwave training is completely driven by the trainee's own brainwave patterns which are rewarded with positive feedback.
The game of Go (Igo, Weiqi, Baduk) has the simplest concept and the deepest play.
Whatever. As someone who has some experience with neurobiology, it is very unlikely you can actually get anything useful from EEG signals. They are totally useless for exploring the brain's functions. Sorry to burst your bubble, but even microelectrode arrays implanted directly into the brain don't provide anything too interesting. At best, you might see a different firing pattern based on external stimulus, and even that is rare.
Actually, this is an area where there has been a lot of progress in recent years. Check out the literature on "hippocampal place cells" for some very interesting results with microelectrode arrays.