Darpa Wants Brain Implants That Record from 1M Neurons

http://spectrum.ieee.org/the-human-os/biomedical/devices/darpa-wants-brain-implants-that-record-from-1-million-neurons

by kungfudoi • 9 years ago

148 93 9 years ago

spectrum.ieee.org

If you were offered the chance to become as intelligent as Feynman but only if you turn over a copy of all your future thoughts to the private business that made this possible, would you take it? Something non-invasive, like a computer that images your brain while you sleep.

If we are machines, then it seems reasonable that within a few millennia we might have this capability. I'm trying to think up some scenarios where it might be reasonable to turn over your brain to a business.

The ability to become "immortal" would probably be enough for most people. Is your disembodied mind still you? Do you feel an attachment to the idea of preserving it? What about modifying it?

Where it gets really strange is if you think of your wife or husband making a backup of their brain, then running a simulator on that backup. I.e. giving it "life" by letting the simulated neurons fire. Does that count as thinking? What if you can see those thoughts? What if you can communicate? Would you love them just as much as you love your SO?

Probably useless questions, but I can't help but wonder.

(I've expanded my comment since it was posted; the "no"'s are in response to the first question.)

Also, https://www.youtube.com/watch?v=IFe9wiDfb0E

aquadrop3• 9 years ago

If you were offered the chance to become as intelligent as Feynman but only if you turn over a copy of all your future thoughts to the private business that made this possible, would you take it? Something non-invasive, like a computer that images your brain while you sleep.

No. If you were offered wings so you could fly, but in exchange you must spend rest of your life in a cage, would you take it?

bkmartin• 9 years ago

Would it then become important for that business to give these minds that would not make them feel trapped and caged? Wouldn't the quality of results be hampered if lots of suffering were to occur? Could the mind at some point become useless since it would be drowned in depressive, cyclical, and often illogical thought? That, too, would create a research point, but would be pretty inhumane and I would hope an ethics oversight panel would not allow such states to exist.

So many questions surround this... If the body is dead, but the brain is still active of reasonable thought and to that end communication, who decides it's fate? Does a mind deteriorate once there is no outside stimuli of the senses?

Not that I would agree to be in this state, but man there would be huge fields of research to do.

bognition• 9 years ago

Honestly it would depend on the size of the cage.

kirykl2• 9 years ago

People already essentially turn their brains over to businesses today, without such grand promises of intelligence or immortality.

Its actually maybe even a pretty accurate description of religion

emiliobumachar• 9 years ago

Religion does make a grand promise of immortality.

amelius• 9 years ago

If you were offered the chance to become as intelligent as Feynman but only if you turn over a copy of all your future thoughts to the private business that made this possible, would you take it?

"And, my students, if we equate both expressions, then, ... ahem, by the way, do you know how easy it is to get a brain extension from Acme Corp?"

jlebrech• 9 years ago

yes I would download all the knowledge I could into offline storage and only connect for updates.

EGreg• 9 years ago

A more interesting question is, what if you wake up one day as one of 10 copies of yourself?

Would you today be worried about preserving the life of all 10 clones (with memories)?

We are all afraid to die - but what if tomorrow there were clones with the memories you have today?

The question basically is posed here

http://existentialcomics.com/comic/1

booleandilemma2• 9 years ago

The movie The Prestige deals with the concept of self copies too.

https://en.wikipedia.org/wiki/The_Prestige_(film)

QuantumGravy3• 9 years ago

From the moment those copies exist, I'm not them and they're not me. We do not share subjective experiences. If you think we're equivalent and expendable, you'd best murder me in secret and pretend it never happened. You know, like incinerating me inside a teleporter or someth... oh crap, we're going there.

Yeah, I have a few issues with that comic.

The man was not a murderer.

This is one of a few logical leaps I can't accept. His past selves are already "dead". His future selves do not yet exist. If, for example, all my ancestors are dead and my offspring not yet conceived, how does my suicide harm them in any way? None will so much as feel the grief of losing me, but I'm supposed to feel guilty over "murdering" them? (Let's not follow up on what this would mean for topics such as abortion!)

If future and past selves are of such importance that we not murder them, how can present copies be valued any less? If destroying the original is not necessary, then that must surely be murder as well. It's worse, actually, because then it's not even hypothetical; "The Machine" would be intentionally and unnecessarily ending a life.

Why do I possess the subjective experience of me, not you, and not some other me separated by time or space? Barring notions of a soul, either those other instances are not real in some sense or else there's something phenomenologically special about each, which I can only attribute to being uniquely present at a particular point in space-time. If the teleported me is a clone, their existence ought to have nothing to do with me, thus I should not be murdered for their sake. However, if the teleported me instead can only exist as a function of my past-self ceasing to be, then that's little different than how I am from any moment to the next.

TL;DR: If it works like the movie version of "The Prestige" where one copy is murdered purely for convenience, I'm not stepping into that teleporter.

captainmuon5• 9 years ago

I think we should really fix society before we develop that kind of technology. We need a socio-cultural breaktrough of the same magnitude as the technological breakthrough. Something on par with the development of monotheism, enlightenment, capitalism, or democracy. A "communistic" utopia (for lack of a better word) without the flaws of its historical implementations. Then the question is moot. There is no private business, or indeed any other entity, government or private, that you would make such a deal with. There is no second party that wants to withhold technology, or steal your thoughts. Rather it will be "From each according to his ability, to each according to his needs" - you get the augmentation for free, and you contribute what you feel comfortable sharing back to society.

Often, utopia is portrayed as the result of a technolocial breakthrough, sometimes "if we could read each others thoughts there would be no misunderstanding and conflict". I believe it is the other way around - utopia is a precondition for singularity-type technology. At least, if we want it to be beneficial and not hellish.

Brain implants are easy, communism is going to be hard.

imron• 9 years ago

communism is going to be hard.

Not if everyone has brain implants!

sillysaurus33• 9 years ago

I agree. Once we develop the ability to upload and modify brains, it will be easy to morph them. We can extract the primal desire to defend yourself, to become angry, to want to own property. These emotions will be reserved for the founders.

It will take time to get there. Many will resist, so we have to keep our goal a secret. But people are gullible. With our morphogenic brain technology, we can give people experiences they've only dreamed of. Think of it. The high of heroin, with none of the downsides. The ability to know instantly when any of your loved ones are in distress. The ability to shape yourself and your children into any form you desire.

If we appeal to base emotions, if we appeal to their need to control, to own, to shape, then we can implement our plan.

AndrewKemendo5• 9 years ago

You've got it backwards. Our brains are what is holding society back.

We need a socio-cultural breaktrough of the same magnitude as the technological breakthrough

That's what transhumanism is. It's a restructuring of the human system around symbiosis with AI.

frenchie41112• 9 years ago

The other, and much more likely in my opinion, future is that developing this kind of technology will cause us to "fix" our society, or rather society will grow to support the new technology. It is unlikely that private business will relinquish it's control willingly, we must aim for a future that makes private business irrelevant (massive automation removing the need for humans to "pay" for anything)

eberkund• 9 years ago

"if we could read each others thoughts there would be no misunderstanding and conflict"

For that to be true you have to assume that everyone has enough empathy to come to agreement. It isn't really much different from the way things are now (without being able to read each other's thoughts).

k_lander• 9 years ago

i don't think any technology should be put on hold because society isn't perfect especially when said technology has the potential to improve society immensely.

synicalx• 9 years ago

Mass starvation and large-scale 'cleansings' will probably be a hard sell if the only upside is "Everyone can read your mind".

kingkawn• 9 years ago

All extrapolated technological symbolism that, when experienced as reality, will fail to stand up to the actual non-symbolic experience of existence.

petra• 9 years ago

I really wonder how valuable would augmented intelligence be , when AI will be so much smarter.

If it would be the same as physical strength today,maybe most people won't bother with that too much, and instead focus on other things,maybe more core to feeling well.

One of those things could be the experiences meditation practices talking about, enlightenment etc, and it seems that we won't need neuron level sensing to do that, but something more at the affordable fmri level or maybe eeg+localization, and those seem to be here much before intelligence augemntation.

So I wonder if we'll ever see intelligence augmentation.

morley• 9 years ago

We basically already have intelligence augmentation; it's just in the form of computers and smartphones.

Interestingly, despite widespread adoption of both, there's still some people "smarter" (i.e. better at certain tasks) than others because they can better apply the tools available.

mfoy_• 9 years ago

The Gentle Seduction is a short story that you might find interesting.

http://www.skyhunter.com/marcs/GentleSeduction.html

bmc7505• 9 years ago

What troubles me is not the permission to read, but to write. While potentially useful for a host of medical applications, this is a scientifically plausible proposal to create technology that will literally have the ability to control minds. While you might be comfortable with sharing intellectual property with a business entity, giving that same entity the ability to adjust internal motivation, or alter sensory input is a dangerous Faustian bargain. How much would you trade for free will?

lordnacho• 9 years ago

What troubles me is not the permission to read, but to write.

Argh, I can just imagine having extremely realistic advertisements beamed directly into my mind, in exchange for something like immediate translation.

It's gonna be really weird when those perfume ads start before Christmas.

propman• 9 years ago

Reminds me too much of Black Mirror

fosco• 9 years ago

makes me this of the original star trek which I have been rewatching recently. in the first episode [0] another race is capable of causing hallucinations and making humans experience anything they desire. later on you find out that the captain (not kirk in the very first episode) is found by kirk and now can only say yes or no through an version of Stephen Hawkings chair. Spock knowing that his brain is still fully operational decides through loyalty to Captain Pike to bring him back to this alien race so he can enjoy living in this fake world and not be in the handicapped state that he is. the episode closes with the Captain Kirk showing interest in wonder at the opportunity provided by this race.

I suspect given the opportunity many would take this. ( without knowing any of the consequences I tentatively would certainly take it in that situation)

[0] https://en.wikipedia.org/wiki/The_Cage_(Star_Trek:_The_Origi...

[1] https://en.wikipedia.org/wiki/The_Menagerie_(Star_Trek:_The_...

jacquesm• 9 years ago

The question you should ask is would Feynman have agreed to it?

Because that's the person who will have to put up with it. You can only ask such a question if you're willing to give the future recipient of the gift the option to back out and be sent back to your present state. And if that new individual isn't 'you' in a legal sense you might find that contract nul and void anyway.

Where it gets really strange is if you think of your wife or husband making a backup of their brain, then running a simulator on that backup.

There is a Black Mirror episode around that called 'Be right back'.

AgentME• 9 years ago

There is a Black Mirror episode around that called 'Be right back'.

Well, they were approximations based on people's social media presence, and the episode's conflict mainly came from the imperfection of the approximation. I think you get a very different set of cultural conflicts if the simulation is of an accurate copy of the mind (issues more like in the book Permutation City).

solotronics2• 9 years ago

what happens when someone copies a simulation of you and tortures it digitally to get all your passwords and information?

gene-h3• 9 years ago

Honestly, I think it's more worth focusing on the near term ethical problems of such technology rather than the far off ones. So let's say that this project does pan out, and three years down the road we determine a programmer with one of these neural interfaces installed can open 20 terminal windows at the same time. With a special set of emacs/vim bindings they can bang out code 100x faster than regular programmers. So now companies are sponsoring their programmers to get the surgery.

Sounds like a pretty good deal right? Well in three-five years time, surgery probably won't change much. Installing such devices is still going to be pretty risky. And even then, having the hardware in introduces its own medical risks. Our super-programmers can't ever go swimming because of the risks of the fiberoptic and power jack plug going through their skull getting infected. This is a problem current medical implants have, so probably one we will still have in a couple years time.

There are a whole host of other problems with this too. IE what do you do when the hardware in your head is obsolete? Can the company that paid to put it in fire you because you aren't keeping up, in part because you have obsolete hardware in your head?

[0]http://amputeeimplantdevices.com/what-are-the-risks/

beachbum8029• 9 years ago

Just make sure not a forget a very important phrase: ignorance is bliss.

fenwick67• 9 years ago

This this this... being intelligent will not make you happy.

I wouldn't take hyperintelligence for free.

abecedarius• 9 years ago

Deaf people today face a similar choice: getting a cochlear implant, you don't control the software (afaik). This is a significant part of why I still have no cochlear implant. If we want freedom of thought in the glorious transhuman future, well, that could evolve out of what's happening today, and we ought to value human autonomy more in today's systems.

DRM is another such arena: as I said in another thread the other day, do you want it enforced by your brain implant? No? Then think about how you want to allow the issue to get framed.

Ditto for compelled decryption.

jahabrewer• 9 years ago

when people read aloud or read silently to themselves the neural signal in the superior temporal gyrus can be used to reconstruct the words

How long until counter-interrogation techniques include not even thinking about what you want to conceal?

xxSparkleSxx• 9 years ago

This is a thought that's been floating around in my head for too long and probably doesn't exactly fit here but oh well.

Isn't language a bit of a proxy for what we are actually trying to communicate? As in the statement "Yesterday, I saw a yellow bus" the are millions of shades of yellow, different bus styles, and of course different times of day and different scenery in which you could have seen that bus.

Is it possible to convey that picture of the yellow bus (wherever and whenever you happened to see it) through a level of communication below typical "language." Is it possible to reconstruct thoughts that have not been put into explicit words? Additionally, if we were able to do this and do it in both ways (sending/receiving) would a Chinese speaker and English speaker be able to communicate without learning a new spoken language?

tim333• 9 years ago

People have thought about that stuff for a long time and it often comes up in fiction. Here's Spock doing it with an alien. https://youtu.be/65dzjOIkwwI?t=59s

It's interesting we may have a new way to communicate thoughts using tech coming up.

mywittyname2• 9 years ago

I think the limitation of our language is a representation of the limitation of our memory and/or thought processing. In college, I remember hearing about people whose language did not have a distinct names for colors, such as blue and green, who also could not distinguish between the two. Meaning, if one held up a blue crayon and a green crayon and asked one of these people if they were the same color, their response would be "yes" in disagreement with your typical American.

captainmuon5• 9 years ago

I would guess the internal representations are too different. Maybe some are inherited, but most develop over the course of your life.

You can think of it as MS Word vs. Libreoffice Writer. You don't get better interoperability by letting an app access the other app's memory, whcih is a crazy historically grown representation of memory-mapped structs.

And by accessing this private structure (and say memcpy'ing external data into it), you open yourself to lots of exploits, like buffer overflows. No way I'm letting somebody memcpy random information into my brain.

QuantumGravy3• 9 years ago

If everyone has similar memories to conceal, problem solved. Who's ready to sign up for Total Recall style memory implants?

mywittyname2• 9 years ago

Probably more like thought-lie training: intentionally mis-remembering the details the interrogator is trying to extract.

shallot_router• 9 years ago

Feels very Ministry of Love.

booleandilemma2• 9 years ago

"It's not a lie if you believe it" - George Constanza

accountyaccount• 9 years ago

or extreme meditative resistance; think nothing

mtgx• 9 years ago

How long until minds are scanned at the airport?

tectonic• 9 years ago

See also Elon Musk's Neuralink and the recent waitbutwhy post all about it: https://waitbutwhy.com/2017/04/neuralink.html

LeifCarrotson• 9 years ago

To pick out one interesting point from that article:

I always assumed Gutenberg had made some genius machine, but it turns out he just created a bunch of stamps of letters and punctuation and manually arranged them as the page of a book and then put ink on them and pressed a piece of paper onto the letters, and that was one book page. ... That's Gutenberg’s thing? A bunch of stamps?

Gutenberg's genius was figuring out that while it took a long time to cast and to set up the stamps - probably longer than just writing the page - he'd Automated his process and could print At Scale. (His other major contribution was a good screw press, and, more importantly, the metalurgy and a casting system for the letters).

sjg007• 9 years ago

Ok so this explains Musk's new startup then.

msimpson• 9 years ago

Not one Johnny Mnemonic joke? I'm proud of you guys.

lwansbrough• 9 years ago

This is hard, but perhaps even harder than you might think. I read an extremely long exploration of this concept a couple weeks ago. It's a layman explanation but it really does a great job of outlining the real complications of such a device. Here's the article: https://waitbutwhy.com/2017/04/neuralink.html

One takeaway that gives you the sense of scale: there are at least 40,000 neurons in 1mm^3 of the cerebral cortex. Those 40,000 neurons create around 20,000,000 synapses. You have to create a sensor that can capture a reasonable portion of those synapses to understand what's going on in that tiny section of the cortex.

criddell• 9 years ago

Modern CPUs have on the order of 10 billion transistors and we have a pretty good idea of how those work. In my (uneducated) opinion, it feels like current tech must be close to providing the tools and scale necessary to start figuring this stuff out, no?

apl• 9 years ago

We didn't have to reverse-engineer the CPU. Also, CPUs are rationally designed; we have zero guarantees that an evolved brain follows any principles at all. Every brain region may be highly specialised to a single algorithm or task with zero mechanistic overlap. Figuring out deep neural nets is a closer analogy, and at the moment we have very little intuition for how to do even that. Keep in mind that the brain is vastly more complex than even the most advanced ANNs.

On the technical side: The most advanced neural techniques allow for parallel measurement calcium signals in ~10,000 neurones. That's a long way off from complete observation!

swiley• 9 years ago

Most modern semiconductors are fairly flat (or are "3d" but it's the 90s autocad kind of "3d" where you have some number of layers.)

Semiconductors are really fairly simple in comparison.

hackpert• 9 years ago

More than anything, I am excited about the possibilities this opens up. A lot of fields that DARPA touches with its funding-filled hands turn into a technology that defines the future, from the internet to self-driving cars to GPS. Exciting times we live in indeed!

solotronics2• 9 years ago

defense technology for everyone! can't wait for my US NSA Facebook TM brain implant to keep me happy and productive.

laythea• 9 years ago

Is this not like shoving a single scrappy copper wire into an Ethernet port and expecting to be able to read, parse and process Ethernet data?

gene-h3• 9 years ago

My only worry with this is that it might be too focused on developing technology rather than doing basic science to make such things possible. Time and time again trying to develop the tech before the science is ready has not been very fruitful.

Current neural interfaces tend to kill neurons and eventually fail. Determining how to avoid these problems is still a very big question. This is incredibly important for making practical neural interfaces. However, the approaches developed might not necessarily be compatible with technologies developed here. For example, the manufacturing process might need to be completely different, or a flexible substrate might be required.

There has progress in solving these problems though: https://phys.org/news/2017-03-graphene-based-neural-probes-p...

devindotcom• 9 years ago

Yeah, I've been working on a BCI explainer and this is a critical point. The science isn't even close, either in hardware, software, or even basic brain knowledge.

Of course, it's kind of one of those situations where you need the knowledge to get the knowledge, but we'll break the cycle somehow!

gene-h3• 9 years ago

I wouldn't be too worried about the software, the human brain is pretty good at adapting to things. A great demonstration of this was an experiment where congenitally blind people were able to 'see' by sitting in a chair with an array of 400 vibrating pads.[0] There has also been some success on the software side with hippocampal prostheses[1]. All in all it's pretty simple stuff, we've been able to reverse engineer the transfer function for part of the hippocampus, so if that region is damaged we can emulate it with an implant and some matlab code.

I would say that this is the end of the beginning for BCIs.

[0]https://tcnl.bme.wisc.edu/laboratory/founder [1]https://en.wikipedia.org/wiki/Hippocampal_prosthesis#Recent_...

kafkaesq• 9 years ago

Which pretty soon will be implanted "optionally" in people released from mental hospitals, out on parole, convicted of sexual offense, etc. In the same way ankle bracelets are "optional" for inmates who fancy the idea of getting out of jail at some point.

fundabulousrIII• 9 years ago

No thanks. We were designed as autonomous nodes for the core purpose of survival. It would be nice to have unlimited information at the speed of light but informational veracity and trust would have to be factored in. Furthermore bidirection is pure IP theft.

Darpa Wants Brain Implants That Record from 1M Neurons

Related Stories