Brain–computer interfaces

A place to discuss the science of computers and programs, from algorithms to computability.

Formal proofs preferred.

Moderators: phlip, Moderators General, Prelates

somebody already took it
Posts: 310
Joined: Wed Jul 01, 2009 3:03 am UTC

Brain–computer interfaces

Postby somebody already took it » Fri Jul 23, 2010 5:50 pm UTC

I've been hearing a lot about new developments in brain–computer interfaces (BCIs) recently. Most notably:
http://en.wikipedia.org/wiki/Brain_Gate
http://www.youtube.com/watch?v=fVhggGSjXVg

Neither of these is perfect by any means. I've head the Brain Implants carry a risk of infection (and I imagine there are many other health risk as well), and the headset seems rather finicky. But nonetheless, BCIs seem like something that could become quite practical in the near future. So, I was wondering what kind of ideas you all have about what we can potentially use them for. Some of the applications I've been pondering are in graphics/design software. For example, what if it becomes possible to alter an image in photoshop (or its equivalent) just by imagining what it would look like with different lighting or with an occluding object removed. Or further, what if we can "paint" scenes just by imagining in increasing detail what they look like. And for 3d animation, rather than taking time to repeatedly pose a character or set up a motion capture lab, it might be possible to just imagine being the character (a lot of the work with robotic prostheses suggests that this is quite feasible).
Also, could we use BCIs for writing code somehow? What would that be like?

fr00t
Posts: 113
Joined: Wed Jul 15, 2009 11:06 am UTC

Re: Brain–computer interfaces

Postby fr00t » Fri Jul 23, 2010 6:12 pm UTC

I don't think this technology is quite anything like what you want it to be yet. The BCI can measure brainwaves - but that's easy. It can in no way decipher them in any meaningful way, so it's up to the person (or monkey) using the BCI to learn how to create certain patterns of brainwaves to control a robotic arm, but that's orders of magnitude more simple than transmitting any sort of decent quality image. I don't think you could just think up an apple and have the BCI render it on screen.. maybe an 8 bit apple.

Also, something about emacs already having a command to read brainwaves.

somebody already took it
Posts: 310
Joined: Wed Jul 01, 2009 3:03 am UTC

Re: Brain–computer interfaces

Postby somebody already took it » Fri Jul 23, 2010 6:45 pm UTC

fr00t wrote:The BCI can measure brainwaves - but that's easy. It can in no way decipher them in any meaningful way, so it's up to the person (or monkey) using the BCI to learn how to create certain patterns of brainwaves to control a robotic arm

Actually, I think you are incorrect about that. In the case of the monkey with a robotic arm researchers actually are decoding the brain signals of the XYZ location it should move to.
See: http://www.youtube.com/watch?v=eg4BfRH4neE#t=20m55s
Although, that is currently only being done with a BCI that relies on a brain implant.

fr00t wrote:I don't think you could just think up an apple and have the BCI render it on screen.. maybe an 8 bit apple.

Probably not, but what I'm imagining is something where the user can gradually refine an image using visual feedback. So in the example of the apple, they might start off with a blob of red that becomes more and more apple-like as they imagine more and more fine grained details on the image they are presently seeing.

fr00t
Posts: 113
Joined: Wed Jul 15, 2009 11:06 am UTC

Re: Brain–computer interfaces

Postby fr00t » Fri Jul 23, 2010 6:58 pm UTC

Ah, ok. But what exactly was the monkey trying to do?

I see what you're saying though. Utilize the brain like a very accurate appendage of sorts. I would make sense that we'd be able to control our brainwaves more precisely than our stubby fingers. It would probably require ridiculous amounts of practice to get to a point where you can control a computer better with your mind than a keyboard + mouse, but it probably would have much greater potential.

I still fantasize about a computer actually being able to interface with the brain more directly. I mean the brain really is like a computer. Input devices are the 5 senses, and output devices are our muscles. I shouldn't think it to be impossible for the pattern of electrical impulses entering and leaving the brain to be deciphered... though I suspect the internal workings may actually be.

somebody already took it
Posts: 310
Joined: Wed Jul 01, 2009 3:03 am UTC

Re: Brain–computer interfaces

Postby somebody already took it » Fri Jul 23, 2010 8:27 pm UTC

fr00t wrote:Ah, ok. But what exactly was the monkey trying to do?

I suppose the monkey was trying to eat the food...
I don't know a lot about neuroscience, and pretty much everything I know of the experiment comes from the video I just linked, so there are probably more in-depth answers to that question if you're willing to research it.

fr00t wrote:It would probably require ridiculous amounts of practice to get to a point where you can control a computer better with your mind than a keyboard + mouse, but it probably would have much greater potential.

I've been thinking, one of the really beneficial things about BCIs is that they can function in tandem with other types of interfaces, so they don't need to be able to replace a keyboard and mouse in order to be useful. One example of this would be using brainwaves to select options from tool bars, the way you might use a hot-key. So, in Photoshop that could mean changing the type of brush you are using just by thinking about it, or in a word processor that could mean writing in bold or italics by thinking about it. In fact, there is already a keystroke emulator for the Emotiv EPOC (the device from the TED talk) that does this.
Another instance where a BCI could be used along with a keyboard or mouse is having a ridiculously accurate auto complete feature that learns from your brainwaves while you type (as someone's brainwaves probably differ enough between when they type four and fourier that simply having 'fo' might be enough to predict the word).

User avatar
Thesh
Made to Fuck Dinosaurs
Posts: 6579
Joined: Tue Jan 12, 2010 1:55 am UTC
Location: Colorado

Re: Brain–computer interfaces

Postby Thesh » Sun Jul 25, 2010 9:11 am UTC

First thing that comes to mind is exoskeletons that not only increase strength (by knowing what movements your body wants to make, its own motor could complement your body in that regard), but provides ballistic, thermal, and NBC protection as well as hardpoints for weapons. Along the same lines, you could use it for fire control systems in vehicles. As you stated technology isn't quite there yet.
Summum ius, summa iniuria.

User avatar
WarDaft
Posts: 1583
Joined: Thu Jul 30, 2009 3:16 pm UTC

Re: Brain–computer interfaces

Postby WarDaft » Sun Jul 25, 2010 9:36 am UTC

The ultimate result of this is the potential for your brain to interface directly with with a virtual reality, having seemingly godlike powers over the universe you perceive.

And then we can have god wars.
All Shadow priest spells that deal Fire damage now appear green.
Big freaky cereal boxes of death.

kmatzen
Posts: 214
Joined: Thu Nov 15, 2007 2:55 pm UTC
Location: Ithaca, NY

Re: Brain–computer interfaces

Postby kmatzen » Sun Jul 25, 2010 9:20 pm UTC

For awhile now I've hoped that this sort of technology would progress to the point where we can have a more capable class of prosthetic devices. However, the surgery for this technology is invasive and even if the technology does get there, I doubt the insurance companies would be willing to spring for it when there are far cheaper alternatives available.

tuseroni
Posts: 124
Joined: Fri Jun 25, 2010 10:41 pm UTC

Re: Brain–computer interfaces

Postby tuseroni » Mon Jul 26, 2010 3:23 am UTC

i would volunteer for such a surgery...

the limit of brain-computer interfaces are ostensibly limitless....but currently quite limited. for instance we can use brain control to move a mouse...in that regard you can make an image by using a mouse....personally id like to make a game which uses your brain waves to create certain conditions, for instance you are presented with a glass that has water in it, the water shows a graphical representation of your brain waves (certain waves...alpha i think) and you need to calm the waves and make it flat (see why i clarified that it wasnt ALL brain waves) then as you move through difficulties you try to change how it moves, favouring increasing some activity, lowering others.

brain waves alone can only tell us so much, they tell what areas of the brain are most active for instance but they can only tell whats going on at the surface. most brainwaves are absorbed by the higher levels. thats why surgical chips get better results.

what im waiting for is a time where nano technology is so good they can release a series of nanites into your body, direct a feild at the area you want the chip, and the nanites will build it inside the skull...simple procedure with no need for surgery. then another feild is given that causes the nanites to break into core components thats get absorbed into the blood stream. when you need to repair or upgrade it nanites made for this purpose are injected.
Image

kmatzen
Posts: 214
Joined: Thu Nov 15, 2007 2:55 pm UTC
Location: Ithaca, NY

Re: Brain–computer interfaces

Postby kmatzen » Mon Jul 26, 2010 4:26 am UTC

tuseroni wrote:i would volunteer for such a surgery...


http://www.braingate2.org/clinicalTrials.asp
Clinical trials are being conducted. You must live within 2 hours of Boston and have limited to no ability to use both your hands due to cervical spinal cord injury, brainstem stroke, muscular dystrophy, or amyotrophic lateral sclerosis or other motor neuron diseases. When I first heard about this project, their test subject was a guy who was a quadriplegic and wanted to contribute to science. Unfortunately, he died from sepsis which may or may not have been related to the surgery.

User avatar
DorkRawk
Posts: 51
Joined: Thu Mar 08, 2007 6:50 am UTC
Location: Chicago
Contact:

Re: Brain–computer interfaces

Postby DorkRawk » Mon Jul 26, 2010 6:00 pm UTC

Does anyone know much about what's being done (and where) in academia for BCI research? I've seen a few papers from Penn State. I've thought about this as a possible area of research in the future but it's been hard to find what CS programs (I would be coming at this from a CS side more than a neurological side) have professors doing work in this field.

kmatzen
Posts: 214
Joined: Thu Nov 15, 2007 2:55 pm UTC
Location: Ithaca, NY

Re: Brain–computer interfaces

Postby kmatzen » Mon Jul 26, 2010 7:13 pm UTC

DorkRawk wrote:Does anyone know much about what's being done (and where) in academia for BCI research? I've seen a few papers from Penn State. I've thought about this as a possible area of research in the future but it's been hard to find what CS programs (I would be coming at this from a CS side more than a neurological side) have professors doing work in this field.


I would probably look for biomedical BCI groups and see if they have a CS faculty member in their group. At my previous school, if anything, this type of research was done by the signal processing and machine learning professors in electrical engineering. Feature selection and classification are probably the main areas of research in this topic for CS.

somebody already took it
Posts: 310
Joined: Wed Jul 01, 2009 3:03 am UTC

Re: Brain–computer interfaces

Postby somebody already took it » Mon Jul 26, 2010 7:56 pm UTC

Here's another possibility I've been thinking about: film, music and maybe literature that responds to people's brainwaves choose your own adventure style. I think this application could have a really powerful Forer effect associated with it, so people might respond more strongly to media just because they think is is tailored to them, regardless of whether or not it actually is.
Additionally, this could make watching/listening to/reading things much more social by having media that accounts for the aggregate response of its audience.

tuseroni
Posts: 124
Joined: Fri Jun 25, 2010 10:41 pm UTC

Re: Brain–computer interfaces

Postby tuseroni » Mon Jul 26, 2010 10:08 pm UTC

"brainwaves not sufficiently scared at that last scene...lets try THIS!"
Image

Akujin
Posts: 27
Joined: Sat Feb 28, 2009 4:38 pm UTC

Re: Brain–computer interfaces

Postby Akujin » Sun Aug 01, 2010 10:15 pm UTC

fr00t wrote:I still fantasize about a computer actually being able to interface with the brain more directly. I mean the brain really is like a computer. Input devices are the 5 senses, and output devices are our muscles. I shouldn't think it to be impossible for the pattern of electrical impulses entering and leaving the brain to be deciphered... though I suspect the internal workings may actually be.


Thankfully it probably is not nessessary to decipher the interface for some applications. I went a guest lecture at University where the guy had been doing experiments involving a direct connection between a computer and the main nerve stem in his arm. He was able to connect an ultrasonic tape-measure and have his brain assimilate the new sense fairly quickly. Initially the sensor output manifested as a physical senseation but after a while this faded and he could just sense how far away an object was from his head [where the sensor was mounted] whilst blind folded. It will be very interesting to see how far the brain can take this ability to integrate new data.

somebody already took it
Posts: 310
Joined: Wed Jul 01, 2009 3:03 am UTC

Re: Brain–computer interfaces

Postby somebody already took it » Mon Aug 09, 2010 1:15 am UTC

Akujin wrote:I went a guest lecture at University where the guy had been doing experiments involving a direct connection between a computer and the main nerve stem in his arm. He was able to connect an ultrasonic tape-measure and have his brain assimilate the new sense fairly quickly. Initially the sensor output manifested as a physical senseation but after a while this faded and he could just sense how far away an object was from his head [where the sensor was mounted] whilst blind folded. It will be very interesting to see how far the brain can take this ability to integrate new data.

This could be incredibly interesting, as it offers the possibility of actually expanding the capabilities of our brains. Perhaps it could eventually be possible to connect the sensory parts of our brains to calculators that use activity from other regions as input, so that just thinking 3080980 times 293 would result in sensing the number 902727140. But that's just a really simple example, any function could work. Just by thinking about searching for a topic on Wikipedia it could be possible to sense the content of its page, probably not all at once, but moving around to different sections, or searching for specific details would just require thinking about doing that.

tuseroni
Posts: 124
Joined: Fri Jun 25, 2010 10:41 pm UTC

Re: Brain–computer interfaces

Postby tuseroni » Mon Aug 09, 2010 3:32 am UTC

i think the issue with the sonic tape measure is that it was passive. a sensing organ that sent data in and the brain interpreted it.
but this leaves other possible sensory organs:
a gyroscope, and accelerometer, a GPS, a magnetometer. an ir camera, really anything which detects EMR on any wavelength.
input seems easier than output.
Image

User avatar
Josephine
Posts: 2142
Joined: Wed Apr 08, 2009 5:53 am UTC

Re: Brain–computer interfaces

Postby Josephine » Mon Aug 09, 2010 3:34 am UTC

I remember reading about something similar. It was an ankle cuff that had 8 electrodes around it. It put out a mild current from the electrode facing north. Artificial directional sense.
Belial wrote:Listen, what I'm saying is that he committed a felony with a zoo animal.

squareroot
Posts: 548
Joined: Tue Jan 12, 2010 1:04 am UTC
Contact:

Re: Brain–computer interfaces

Postby squareroot » Mon Aug 09, 2010 6:07 am UTC

tuseroni wrote:i think the issue with the sonic tape measure is that it was passive. a sensing organ that sent data in and the brain interpreted it.
but this leaves other possible sensory organs:
a gyroscope, and accelerometer, a GPS, a magnetometer. an ir camera, really anything which detects EMR on any wavelength.
input seems easier than output.


I agree, I'm because the brain has a neural network, which works great for learning to understand the data, whereas it has to teach itself how to send input with little-to-no feedback without you consciously checking it. In order to learn to use the calculator, first the brain would need to learn to send a pair of numbers and an operation accurately, before it could hope to understand the result. With our eyes, we constantly have input, so we learn that; once we have those, we can learn muscles quickly too.

I think we might need to train the calculator to interpret signals, Here's my idea: give it twenty, thirty nodes, and some basic neural networking software (but with potential for a complicated neural structure, if needed). It will have two buttons, one for "Correct", one for "Incorrect", in addition to a regular keypad. We can train it (and our brain) by thinking hard about a number, and then typing it in. We do that many times, and hopefully it will begin to recognize it. Then we try thinking about numbers, and it tries to figure it out. We press "Correct" or "Incorrect" to tell it if it got it right.

Interleaved with this process would be practices where the calculator send numbers to the brain; maybe counting, then powers of two for a while, then 1-3-2-4-3-5-4-6, etc... maybe with audio so we can sync the feeling with the actual number.

Hopefully, the first exercise would teach the calculator to understand us, and the second would teach us to understand the calculator. Similar processes for learning would be used for learning operations and functions. "Training games" could be played later on, where you think a number, and the calculator sends back the number plus one, or the number times two, etc.. Or the calculator sends a number, you send back the number plus one, and then it might beep one tone if it receives the correct answer, and a different tone if it was wrong. This would continue to increase both our and its ability to process and send signals correctly. Perhaps this would eventually allow us to be "human computers"! :D

P.S. - I'm guessing digits would need to be sent to and from the calculator one at a time - thinking "1435" wouldn't work, but "1","4","3","5", might. I can imagine that then, if you were dealing with some binary string, it might get strange signals if it received a combination of mixed binary and decimal representation, as your brain flicked between the two.
<signature content="" style="tag:html;" overused meta />
Good fucking job Will Yu, you found me - __ -

tuseroni
Posts: 124
Joined: Fri Jun 25, 2010 10:41 pm UTC

Re: Brain–computer interfaces

Postby tuseroni » Tue Aug 10, 2010 1:25 am UTC

personally i like the idea of using a GA evolved neural net as an interlink between a machine and...well...our evolved neural net. but of course something has to tell it no so it knows it fitness.
but i still think a calculator is a bit too complicated. numbers are such abstract thing as are the concepts of mathmatics.

however it may be possible for instance to use the above mechanism for moving a mechanical limb. place an evolved neural net between the inputs from the nerves and the output to the machine the fitness is determined by how close the arm gets to an arbitrary goal. then how close it get to some other goal, keep moving the goalpost til it it gets to the point of doing what you are telling it to do. your brain and the neural net will both be adapting to the new machine.
your brain has a fitness function built in, it gets rewards if it does what you want and punished if not. you may not even need a NN, there was an old experiment using FPGA's and a GA to evolve a circuit that responded to the word "go" to light up a light, which it did without a clock and with only 50 gates.
whichever it might make a nice middle man to bridge the gap between man and machine.
Image


Return to “Computer Science”

Who is online

Users browsing this forum: No registered users and 8 guests