How much RAM would you like with your brain, sir?

March 13, 2003 1:05 AM

World's first brain prosthesis revealed. Well, first hippocampus replacement at least. If this is not a dead end for science (which I doubt), I am gonna get my soul fully digitalized in 2020, then spreading it on the whole net with some new version of a code-red virus. :-)
posted by zerofoks (14 comments total)

I wrote about this article at my weblog. Basically, I think that it is pretty unlikely that it will work in this instance, although I do think that the research is worthwhile (we do recordings from rat brains in the lab I work at in Cambridge University).
posted by adrianhon at 2:10 AM on March 13, 2003

Absolutely incredible. This is nothing less than an artificial memory: The job of the hippocampus appears to be to "encode" experiences so they can be stored as long-term memories elsewhere in the brain. "If you lose your hippocampus you only lose the ability to store new memories," says Berger. That offers a relatively simple and safe way to test the device: if someone with the prosthesis regains the ability to store new memories, then it's safe to assume it works. Can you just imagine the potential of something like this - a cheap version might be able to extend the average working lifespan by years and stave off Alzheimers, while a really good version might even enhance your normal memory. Is this straight out of Snow Crash or what?
posted by Bletch at 2:53 AM on March 13, 2003

So there's hope for Mr. Bush?

(Sorry, there just didn't seem to be enough political content in this post...)
posted by jpburns at 4:47 AM on March 13, 2003

I thought this was the most compelling question - "If someone can't form new memories, then to what extent can they give consent to have this implant?"
posted by agregoli at 6:49 AM on March 13, 2003

There is a spectrum ranging from immediate sensory memory to long term -- presumably, those without a properly functioning hippocampus can remember long enough to briefly understand their situation and agree to the operation, but probably not long enough to finish a conversation. Think Memento.
posted by Tlogmer at 7:58 AM on March 13, 2003

Has anyone seen the movie "Memento"?...........

"Artificial hippocampus"...bizarre and creepy.....Where can I get one?
posted by troutfishing at 8:06 AM on March 13, 2003

Tlogmer - Ya beat me to it.
posted by troutfishing at 8:08 AM on March 13, 2003

We already allow people with damaged hippocampi to perform all sorts of other tasks, so I don't really see why consent would be that serious an issue. They still possess short-term memory (the hippocampus turns short-term memories into long-term ones), so you'd have to explain it to them in fifteen minutes or so, but it could be done.

My particular worry is how they're going to design a human version of this. Specifically, if they don't understand how the hippocampus works, how do they intend to scale a device designed to work on rats to the human level? It's a truism that human and rat brains are very different, and the only option that seems open to them is trial-and-error. I mean, I wish them well, but I wouldn't want to be the first person to get one.
posted by Pseudoephedrine at 8:38 AM on March 13, 2003

If this is not a dead end for science (which I doubt), I am gonna get my soul fully digitalized in 2020, then spreading it on the whole net with some new version of a code-red virus.

Lawnmower Man!!!
posted by me3dia at 8:48 AM on March 13, 2003

Lawnmower Man!!!

Yes, but with cuter and better actors.
posted by zerofoks at 9:14 AM on March 13, 2003

Slices of rat hippocampus were stimulated with electrical signals, millions of times over, until they could be sure which electrical input produces a corresponding output. Putting the information from various slices together gave the team a mathematical model of the entire hippocampus.

This, in itself, is fascinating. It means that the functioning of the hippocampus is entirely deterministic, doesn't it? That the same input is always going to produce the same output?

Would the rat receiving the implant wind up with its memory structure determined by the rat which was modeled? Even with some of its actual memories?

What does that imply for, you know, what makes us us?

*shivers*
posted by ook at 10:17 AM on March 13, 2003

Man, the only conditions under which I'd take an artificial brain is if I was dying anyway and it was the only hope I had of continuing to be conscious. It has not been proven to me that you would retain your personal identity if it was "downloaded" into a computer. In fact, quite the contrary, it seems like "you" would cease to exist.
posted by Hildago at 4:36 PM on March 13, 2003

Okay, I've gone back and read adrianhon's blog link now, and you should too. Yeah, you. Meanwhile, i have a couple questions for Adrian. (It's such a treat when a MeFite, y'know, actually knows something specific about the topic at hand.)

Setting aside for the moment the problem of modeling the whole hippocampus, accurate recording, the unknown chemical signals, preserving the actual neuronal structure, etc -- you know, the little details -- are neurones deterministic? I mean, if you're looking at some manageably small, arbitrary network of neurones, does the same electrical input always give you the same output? Or at least predictable output? Or is that just way too simple?

Question two: I get the impression from this article that the hippocampus, at least, is relatively self-contained -- that it has its own complex internal structure, but comparatively few connections out to the rest of the brain. Is that correct? And if so, are there lots of parts of the brain that are like that, or is most of it just one big interconnected mass?

I guess like most people who skipped their biology classes to spend extra time in the computer lab, I have this mental image of the brain as a bunch of little interconnected turing states churning away, and I'm curious just how far wrong that picture is...

And while you're at it, when did they stop spelling it just "neuron"?
posted by ook at 5:46 PM on March 13, 2003

IANAN(eurologist), but I was a psych major until I transferred out, so here's what I know about hippocampi.

What the hippocampi do is take short-term memories and encode them into long-term ones. The only reason we know this is because people who have their hippocampi (you've got two of them, one for each hemisphere) removed lose their ability to form new memories. What's the difference between a short-term memory and a long-term one in the brain? Your guess is as good as anyone's at this point.

It's not quite self-contained, as it is a conduit. Short-term memories go into one end of the system, and come out from the other end as long-term memories stored somewhere in (we think) the temporal lobe. Somehow, the amygdala and the limbic system (the part of the brain that deals with emotions) are involved. The frontal lobe might be involved too.

The hippocampi don't "create" memories in the strict sense though. That is, someone who's hippocampi are damaged is still able to remember things from the immediate past - usually the last three or four minutes, not the fifteen Leonard in Memento could - it just turns those short term records into ones that can be recalled later. Nor is the hippocampi involved in recollection. People who have no hippocampi can still remember everything up to the point when they stopped working. I recall watching a video of one very sad case (known as N.G., if memory serves, pardon the black humour) where a man, who is still alive, had been injured as a teenager back during the sixties. Every day he wakes up thinking that he's a fifteen year old boy still living with his parents, only to find out that he's almost fifty and living with himself, and that his parents are dead.

The thing to keep in mind for people like that, as compared to Leonard in Memento, is that Leonard's memory was very episodic - that is, every fifteen minutes he would forget what he was doing. In actual practice, there's a sort of continuity so long as they remain conscious - N.G. could maintain a conversation, sort of, because he remember that he'd been talking to you for the past three minutes or so, and though he couldn't remember who you were, or what you were doing, he'd learnt to live with it. Still, it was pretty sad.

Oh, and neuronal is the adjective, neuron is the noun.
posted by Pseudoephedrine at 6:44 PM on March 13, 2003