Implants Allow the Blind to See

gihan_ripper writes "Neurosurgeon Kenneth Smith has performed a revolutionary operation on St Louis resident Cheri Robertson, connecting a camera directly to her optic nerve. The rig is in principle similar to Geordi La Forge's visor, albeit in very rudimentary form. At present, the 'image' consists of a number of white dots, as on an LED display. There are also governmental restrictions on this research, forcing Kenneth and his team to fly to Portugal to carry out the operation. If this technology takes off, the future will be bright for the sight-impaired."

Comments

Re:Uh?

[phoenix.bam!]

It sure is news of the patient isn't sent into a fit of spasms from a seizure every 45 minutes while the camera is activated.
You know, like what happened 10 years ago.

Re:Infrared?

[MustardMan]

To look at the world "like one big xray slide" you'd have to carry around a source of xrays and put them behind the subject, then use your xray sensitive eyes (good luck developing those) to detect the rays coming through the subject. It's not exactly like there are a buncha xrays flying through us all the time, ya know.

Re:Uh?

[Anonymous Coward]

FTFA:

"When I realized yes, I am going to be blind, I thought, I guess I'm going to learn to do things a little differently now," Robertson says. And she did. She traveled to Portugal to become the 16th person in the world to have special electrodes implanted in her brain. With the help of a device, she could see again!

While it seems to be a rare operation, the parent was right: this has been done before.

Making brain neurons light-sensitive

[cyberied]

Another strategy was just invented: if you lost your photoreceptors, just make the other neurons in the retina or brain sensitive to light. A group just managed this today, for the first time, in mice. Blind mice, who had been treated with viruses that cause the targeted cells to express light-activated channels, were able to regain transmission of information about the external world to cortex. This was recently reported in a blog, and in other media.

Not for every blind person

[shizzle]

Note that patients need to have had sight in the past for this device to work. The visual cortex doesn't develop in people that were born blind, so their brain doesn't know what to do with these inputs. (Like in the movie "At First Sight".)

Pretty cool nonetheless.

Re:Difficulties in the US

[blincoln]

I didn't quite understand from the article why this procedure was prevented in the US, aside from cost.

This is more or less the same technique that's been researched for decades - I saw a film (as opposed to videotape) of it in junior high when I was a kid.

There are a number of problems - as others have mentioned, it tends to cause seizures in its users. IIRC this is because the apparatus itself is fairly crude and overloads the part of the brain it's connected to. It also doesn't work very well - the resolution now is not a whole lot better than back then.

Obviously an argument can be made that someone who loses their sight may consider any visual ability valuable enough to outweigh the risks, but in this case I think the FDA is right. This particular technology is not mature enough to allow as a commercial product. There are others in development that IMO are more promising.

Only useful for people who once had sight

[CorporalKlinger]

It's important to note that due to the way the human brain develops synaptic connections in the visual cortex, only humans who had sight from birth to some age beyond 3 to 5 years of age will benefit a great deal from such a procedure. While people who are blind from birth due to cataracts or other conditions obtain some visual perception when the cataract is later removed, most never develop the neural connections that allow them to identify what they're seeing. Everything from navigating around desks in a well-lit classroom to differentiating a face from a table, a television, a light bulb, or an automobile is all but impossible if the visual cortex doesn't develop properly in response to normal visual stimulus from birth. Sight is useless without the ability to percieve what one is really seeing. So while this is incredibly impressive and promising for people who had sight but lost it, don't expect that this will be a cure-all to allow people with all types of blindness to see again.

Re:Not optic nerve.

[HermanAB]

The optical nerve goes to the back of the brain.

Re:Wow

[daddyrief]

good ol' W. and his buddies -- aka, the Religious Right.

Obviously it was illegal before Bush/the current Congress. You can hardly say that the Right isn't responsible however, especially during this current administration, for keeping these types of research from coming to light. The technology and opportunities are here now, too bad 'religion' and 'morals' aren't synonymous with 'progress'.

And yes, I want to blame Bush and his whole cabinet for the Iraq war. It isn't entirely his fault, but to blame him is asinine. He stepped out of his bounds, fabricated stories and evidence, lied to millions, and killed thousands of people. Seems like reason enough for blame to me.

Re:Difficulties in the US

[Anonymous Coward]

I believe that this stems from certain governmental regulations and restrictions on medical research, notably FDA approval of implantable medical devices. It takes quite a bit of testing and analysis before that approval is given. If I remember correctly, the cited news articles states that a major complication in this regard was the possibility of infection at the point of entry through the skull into the brain.

Re:Wow

[mrhartwig]

Did you forget that it requires an act of Congress to declare war....

Please provide a reference for that act of Congress that declared a state of war to exist between the US & Iraq. Not the 2002 resolution that authorized force to enforce UN resolutions (http://en.wikipedia.org/wiki/Joint_Resolution_to_ Authorize_the_Use_of_United_States_Armed_Forces_Ag ainst_Iraq); the one that says "A state of war now exists between...."

Good luck.

Also see http://en.wikipedia.org/wiki/Declaration_of_war_by _the_United_States

it doesn't work like that

[Xerxes1729]

Uh, no. "Image sensors", like eyes, don't produce a signal that is fundamentally different from the signals produced by any other sensory organ. What matters is where in the location in the brain to which those signals are directed. Although I'm not certain, I'd guess that this is why the technique won't work on those who were never able to see - they never developed the necessary neural connections in the brain for vision.

Re:it doesn't work like that

[pla]

You should be able to send any type of signal to any part of the brain, and eventually the brain learns what that signal is about.

Yes and no...

For general-purpose processing, most parts of the brain can take over for other parts, (possibly) regaining almost full functionality over time.

With vision, however, you have the single biggest allocation of task-dedicated meat in your entire brain. Evolution has hard-wired the visual cortex for computational efficiency in dealing with a staggeringly large amount of input. For example, we do not actually see in 3d; we see 2x 2d, which our visual cortex manages to "tag" with depth information using clues from lighting, size, and seeing the same object from two angles. Interestingly enough, most of the processing removes the uninteresting parts of that information, as well as filtering out white noise in the signal - Nerves do not act just like little wires, and do a piss-poor job of accurately conducting a signal... You get a result more like the "telephone" game, though the signals average out to basically-accurate over time.

So, for someone who could never see, although this might let them, through conscious effort, interpret parts of their environment as meaning "about to walk into a wall", don't expect them to start driving. As the simplest reason, don't think of our eyes as an X-by-Y grid of sensors, but instead as a random jumble of X*Y sensors. They have no real order to them - Part of our very early development includes the visual cortex learning which "pixels" fall adjacent to which other pixels. And on top of that, our sensors don't even return raw light intensity data - They carry out a form of simple edge-detection and pass that on to the brain. So for a sense of what that would "feel" like to suddenly have the ability to see but never have developed that mapping, imagine trying to make sense of Times Square in terms of reading a list of boolean values that correspond to "edge" or "no edge".

A wired article related to this technology

[Rayston]

You might find this interesting as well.

http://www.wired.com/news/medtech/0,1286,59634,00. html

Summary is incorrect

[Illserve]

The device stimulates the brain directly, not the optic nerve. Stories like this have been kicked around the block for quite awhile.

Re:Lots of possible mods

[Anonymous Coward]

Friend, you have no idea how "tricky" it is. The retina itself does massive pre-processing of the signal: color detection, edge detection, and motion detection are all done inside the eye. What gets presented to the brain is not the raw "dots" of a visual image, it's more like it's been JPEG'd with a very, very customized and dynamically changing algorithem.

Also, there's a nasty trade-off in electrode size and stimulation area that severely limits this approach. As you make electrodes for implantation smaller, in order to stimulate a smaller area and only hit a few desired nerves instead of stimulating a whole region and making someone's face twitch or putting weird flavors in their mouth, their impedance gets higher and the current density increases. Eventually, the high current density causes the lovely salty body fluids to go through electrolysis, which you absolutely do not want happening around nerve tissue. And as the impedance gets higher, you may wind up needing quite large voltages to drive the current.

10 kOhm impedances are certainly not unheard of for good electrodes at some frequencies, and similar electrodes in cochlear implants have been noon to require more than a milli-Amp of current for a few users. +/- 15 Volts directly into your nervous system, anyone? You think accidentally touching a car battery is a surprise? Try 12 Volts DC straight to the nerves that are tied directly to your brain. It's not a nice thought.

So, safety aside, which I'm sure these doctors looked at extremely closely, the resolution of the resulting picture is going to be exceedingly poor due to the current spread. Unless they've implemented a new electrode technology, such as David Edell's work at MIT finally turning useful and getting a 10x10 grid in place by putting gold electrode plated silicon grid across the nerve by cutting the nerve and letting it re-attach through the grid, it's never going to work well.

Cochlear implants, which are a very similar technology at their core, work surprisingly well becasue the electrodes are in a bony channel, the cochlea, that helps constrain the current, and because the cochlea is laid out well for it. The deeper the electrode, the higher the frequency of sound that was detected there, and they're spread out over a distance large enough to reduce serious cross-talk from current spread. Implanting electrodes in the optic nerve doesn't have that stretched out area for stimulation: the spatial layout of the visual field is wrapped into a bundle a few millimeters across, and you have to get at that.

It's interesting work, but it's unlikely to ever work well enough to read until someone comes up with a really new and effective way of connecting electrodes to the nervious system.