darb.ketyov.com

Caveat lector: This blog is where I try out new ideas. I will often be wrong, but that's the point.

Home | Personal | Entertainment | Professional | Publications | Blog

Search Archive

Loading...

16.7.12

Building a better RoboCop

This week marks the 25th(!) anniversary of RoboCop, one of the hallmarks of 1980s US cinema and hero of Detroit.

Detroit's soon-to-be RoboCop statue!
This past weekend was also my 15th (or so) year attending Comic-Con. In a row. Because I didn't get to hear about the new RoboCop remake starring Gary Oldman, Samuel L. Jackson, and Hugh Laurie, I decided to celebrate this nerdistry my own way: by talking way to much and overthinking a plate of beans.

This is me priming myself to thinking about wild ways of designing brain-computer interfaces (BCI) for my new research projects.

Thus, I've decided to take a modern neuroscientific look at RoboCop. Given the huge technological advances in the 25 years since Peter Weller, Miguel Ferrer, and Red Kurtwood Smith, I believe that RoboCop would be much cooler now.

So let's play with some probably very unethical human and cognitive enhancement possibilites, shall we?

As a quick aside: I just found out that Peter Weller is finishing his PhD in "the history of fifteenth-century Venetian art" at UCLA, with a "minor in Ancient Greek and Roman art". And he plays jazz trumpet in a bebop sextet. Whaaaat?! Awesome.

Now I'm not the first one to think about the neuroscientific implications of and possibilities of RoboCop. Apparently Ed Neumeier, writer of the original RoboCop"was grabbed as a consultant by the US Air Force for a program called 2025. They were trying to re-imagine their role in the future and they wanted me to foster imaginative thinking." Government guys, if you're listening, this isn't legit!


Look at this guy! Moral and ethical implications aside, how could you not want to be a bullet-proof, enhanced super cop!? But, cool as he is... he's a little dated. Here's a glimpse into the visual enhancements and HUD-like overlay implanted in RoboCop's nervous system:


I'm pretty sure that the Miguel Ferrer overlay is not part of the standard package, but the P39-green font is your only option! We've got processes such as "MEMORY.DAT" and "ROBO UTILS" that are very important pieces of information for your robo-visual inputs.

As a neuroscientist, how would I design RoboCop's enhancements... if money and was no object and the subject was totally willing?

Mind you, I'm a big proponent of research ethics, so I wouldn't actually do this even if the subject were willing. Seriously. Some of this stuff is dangerous and highly speculative. But this is my place to be wildly speculative, and this is a fun way for me to talk about where the current state-of-the art in BCI stands and maybe poke some holes in the places where our neuroscientific understanding is weakest.

Some of the biggest technological advancements in the last 25 years have come from the telecommunications industry. The first thing that popped into my mind was adding a GPS system into RoboCop and giving him a map overlay on top of his vision allowing for augmented reality interfaces.

Essentially this would be a beefed-up Google Glass.

Imagine integrating this system with something like Word Lens below:








Think about how many other things one could do with such a system: alternative computer-aided visual inputs such as UV, IR, etc? Done. Automatic object/threat detection with visual overlays to exogenously capture attention? Not a problem. Draw a green line over your visual field to highlight your GPS-based route to a destination? Easy.

We know that taxi drivers who have really learned the layout of their cities have bigger hippocampuses. So the obvious solution is to just inflate RoboCop's hippocampus to the size of a building! Best. Taxi driver. Ever. But really yucky oversized hippocampuses...

The great thing is that we can make his hippocampus beefier just by having him exercise, thus increasing his knowledge of maps! SCORE.

But seriously, given access to the nervous system you could put a quick interface between the external world and the eyes, ears, nose, and mouth to filter out noxious stimuli. And I'm not just talking about filtering out tear gas. A stun grenade, which is blindingly bright and deafeningly loud, would have no chance to enter the nervous system and stun RoboCop because it could be digitally filtered out before being passed along to the nervous system.

Kina like noise-canceling headphones but... more warlike.

Keeping with external enhancements, there's a great story out of Berkeley in 2011 where a paralyzed student was able to walk across the stage to accept his diploma via the use of a Berkeley-developed exoskeleton.

Okay, maybe what I want is less of a RoboCop and more of an Iron Man...


The point of RoboCop was to create the perfect symbiosis of the computing speed, accuracy, and durability of a computer with the flexibility and learning of the human brain. The problem was that RoboCop kept having flashbacks to his previous life with his family, as well as some traumatic memories of his near-death experience. These emotional responses were debilitating during critical moments of the film.

We know that there's a strong link between memory, emotion, attention, and other "executive functions" (which is the neuroscience term for all sorts of cognitive processes). One of the most classic studies showing the link between executive processes was the project by Stanford researchers Shiv and Fedorikhin that showed that if you gave people a hard working memory task (memorize a seven-digit number as opposed to a two-digit number) people were able to exert less cognitive control... meaning when presented with a choice for a snack they were more likely to eat the less-healthy chocolate cake instead of the more-healthy fruit cup.

In another study looking at fMRI activity and functional connectivity between brain regions--which may be a measure of how strongly two brain regions are "talking"--the researchers found that when they overloaded people's working memory, activity in the dlPFC--a brain region important for maintaining a memory over a short period of time--saturated while at the same time activity in the amygdala--a region important for "assigning" emotional context/content to experiences--increased.

Unpacking all that neuroscience: basically when subjects had to remember a longer string of letters, the strain on their working memory reduced their ability to make the more "rational" decision. It reduced their cognitive control. And as memory load increases, decisions may become less "rational" and more "emotional", insofar as such terms can be applied to differences in brain activity.

Nevertheless the relationship between memory and emotion is well documented. We know that more emotional events are better remembered, although stress reduces memory consolidation. So we might also want to be sure we can monitor RoboCop's stress levels and pump that dude full of beta blockers as needed.

How do we deal with RoboCop's intrusive memories though? How do we help his PTSD? Well there's some really amazing work being done in this realm right now and... well I can't do any justice to Jonah Lehrer's excellent summary, so just go read that.

For a total "what if" scenario such as this I was planning on avoiding pharmacological intervention ideas since drugs are so messy, non-specific, and are rife with side-effects, but the PKMzeta inhibitor idea discussed in detail in Lehrer's piece might be the only option with something as diffuse and dynamic as memory.

Lehrer correctly notes that, "the problem with eliminating pain, of course, is that pain is often educational. We learn from our regrets and mistakes..." but in this case we very specifically aren't going to worry about that! Convenient.

The flip side of this coin is memory enhancement.

Non-invasive brain stimulation techniques such as tDCS and TMS have been shown to improve the learning of motor skills, which is clearly relevant for a RoboCop. Imagine using tDCS or TMS to improve shooting accuracy or learning! There's even been a study looking at EEG markers for differences between expert and non-expert riflemen (riflepeople?) where they found that frontal EEG activity in the three seconds prior to the taking of the shot differed between the two groups.

Ostensibly then, if this marker is causal, manipulation of pre-shot activity in that brain area might improve accuracy! Or more likely, probably not! But hey, we've got a really willing subject with RoboCop, and this is an empirical question, so there's one easy way to find out. 

Since RoboCop's already got a big metal helmet anyway, we're just going to go ahead and put in some non-invasive brain stimulators to see if we can give him that extra "boost".

That said, given that we can remote-control rats using motor implants, we might even be able to give our RoboCop a manual override and have the cyborg-equivalent of a UAV.

Would that be an Unmanned Ground Man? UGM?

Although there's a lot of work regarding brain-stimulation and motor control, improving memory is another beast altogether.

There's a pretty amazing paper by Ted Berger and colleagues looking at how an artificial, digital hippocampus implanted in rats may actually improve memory encoding. Seriously, these researchers created a digital hippocampus that apes the activity of the real hippocampus and then implanted that into rats to see if that aped version performed the same basic functions.

Although that strikes me as some serious cargo cult type reasoning, amazingly, it seems to have worked!

Whoa.

Man, this post is getting to be unwieldy! Motor learning improvements, remote-controlled men, memory modification and enhancement, exoskeletons... it's just too much!

There may have to be a follow-up post exploring some other possibilities.

In the mean time, if you haven't seen RoboCop yet, check it out. Looks like Netflix even has it on Watch InstantlyThis movie's got it all! There's even a brief appearance by the amazing toxic-waste zombie! You know I loves me some zombies.


(Does that Netflix link count as advertising? If it does, please send royalty checks to that dude above for his amazing zombie skills.)

ResearchBlogging.org
Maguire EA, Gadian DG, Johnsrude IS, Good CD, Ashburner J, Frackowiak RS, & Frith CD (2000). Navigation-related structural change in the hippocampi of taxi drivers. Proceedings of the National Academy of Sciences of the United States of America, 97 (8), 4398-403 PMID: 10716738
Baba Shiv, & Alexander Fedorikhin (1999). Heart and Mind in Conflict: the Interplay of Affect and Cognition in Consumer Decision Making Journal of Consumer Research, 26 (3), 278-292 DOI: 10.1086/209563
Yun RJ, Krystal JH, & Mathalon DH (2010). Working memory overload: fronto-limbic interactions and effects on subsequent working memory function. Brain Imaging and Behavior, 4 (1), 96-108 PMID: 20503117
LaBar, Kevin S, & Cabeza, Roberto (2006). Cognitive neuroscience of emotional memory Nature Reviews Neuroscience, 54 (1), 233-264 DOI: 10.1038/nrn1825
Berger TW, Hampson RE, Song D, Goonawardena A, Marmarelis VZ, & Deadwyler SA (2011). A cortical neural prosthesis for restoring and enhancing memory. Journal of neural engineering, 8 (4) PMID: 21677369
Doppelmayr M, Finkenzeller T, & Sauseng P (2008). Frontal midline theta in the pre-shot phase of rifle shooting: differences between experts and novices. Neuropsychologia, 46 (5), 1463-7 PMID: 18280523
Talwar SK, Xu S, Hawley ES, Weiss SA, Moxon KA, & Chapin JK (2002). Rat navigation guided by remote control. Nature, 417 (6884), 37-8 PMID: 11986657