The Science of Avatar

by Stephen Baxter

  The amnio tank is an extension of a technology used on Earth, a kind of artificial womb used to grow replacement organs and limbs, cloned animals, and sometimes cloned humans. It contains a suspension fluid with carefully monitored nutrients, growth stimulants and other materials. Alternatively a patient with serious injuries or organ failures—Jake Sully, perhaps, if Quaritch had fulfilled his promise of a cure for spinal injury—can be placed in a “cellular rebuilder,” a modified amnio tank. With the patient in an induced coma the damaged tissues are rebuilt at a cellular level under the control of nanotechnology.

  In the specialised amnio tanks aboard Venture Star, nutrients and growth stimulants are supplied to the growing avatars, which are taken from childhood to young adulthood during the starship’s five-year flight. For human applications the fluid contains the salt balance of Earth’s oceans; for avatars a more alkaline solution like Pandora’s seas is used.

  The growing avatars we see in their amnio tanks are unconscious, but evidently alive. Later in the movie, whenever we see Jake come out of the link, his avatar body falls, unconscious again, yet it clearly stays alive until the next link. So even without the psionic link to its operator the avatar body maintains autonomic functions; its heart beats, its lungs take in air, its blood flows (or at least the avatar’s equivalent of these structures and functions work).

  This makes sense. Your mind has some conscious control over your body, at varying levels. You can will your hand to raise, and it rises; you can will yourself to run, and you run. But there is a whole set of neural subfunctions which translate your conscious command into the detailed operations required to fulfil that command. When running you don’t have to think about which leg to lift up next, let alone individually control the various muscle groups to achieve the lifting of that leg. All this interfacing is “downloaded” into the avatar. Driver Jake has to learn to use this interface as he trains with Neytiri; he says he has to “trust my body to know what to do.”

  In addition your body has a suite of infrastructure-type functions that operate beyond your conscious control entirely, and many beyond your awareness: they keep your heart beating, your blood circulating, your food being digested. There are even operations going on at the level of the individual cell; bone marrow cells keep producing new blood cells at a rate of millions a minute, every minute. None of this stops working when you forget to think about it, happily. Again all this (or its equivalent) is evidently downloaded into the avatar body.

  So the avatar body in the amnio tank is a living creature, with a set of necessary functions in place and operating, long before its first link. And it undergoes some basic “training” of its own during the star flight in support of these functions. In the scene when the new avatars, just off the Valkyrie, are inspected, avatar rider Norm Spellman notes that the bodies have undergone “proprioceptive sims” (that is, simulations) during the journey. As I noted in Chapter 25 proprioception, a sense of position, movement, locomotion, is surely especially important to a tree-climbing quasi-Na’vi like the avatar. And the sims, training these senses into the avatars in their tanks, have the added benefit of improving muscle tone, as Max Patel notes. The in-tank training is based on early, Earth-based experiences. Norm would have worked with his avatar when it was at the developmental stage of a child, playing games to develop motor control. (What fun that must have been!) The games are recorded and replayed to the growing avatar in its amnio tank as developmental exercises.

  Perhaps the most revealing bit of dialogue in these scenes is when the techs remark that Jake’s avatar shows no “truncal ataxia.” Ataxia telangiectasia, or Louis-Bar Syndrome, is a rare inherited neurodegenerative disease that affects motor control and weakens the immune system. The first signs of the illness are often detected when the patient is a toddler. “Truncal ataxia” means difficulty with body posture and movement; the child may have difficulty learning to walk. The reference to ataxia reminds us that to the techs the force-grown avatars are indeed children, only a few years old, and they worry about their development like any anxious parent. (All this is impressive scriptwriting, by the way. The creators took care to make the science detail convincing, and to reflect it in the dialogue.)

  But the avatars only exist at all because of some pretty advanced genetic engineering. In fact, as hybrids grown from a mix of human and Na’vi genetic material, the avatars may be the twenty-second century’s ultimate GMOs: genetically modified organisms.

  In our century a GMO (also known as a GEO, genetically engineered organism) is an organism whose genetic material has been purposefully altered through what is known as “recombinant DNA technology,” in which DNA molecules from different sources are used to create a new set of genes. This material is then implanted into the organism to give it new or modified genes. A “transgenic” GMO takes genes from different species, while a “cisgenic” GMO takes genes only from the organism’s own species. The genes are transferred using viruses, or mechanical means like syringes—techniques used in the genetic therapy I mentioned in Chapter 19.

  The first experimental GMO, produced in 1973, was an E. coli bacterium implanted with a salmonella gene. Today GMOs have wide applications, including medical research and the production of drugs. And GMO technology has become a multi-billion-dollar global industry through the use of GMOs in agriculture. Crop strains can be produced that, for example, naturally produce pesticidal proteins. By 2005 over eight million farmers throughout the world were using GMO crops. The controversy over the use of GMOs in agriculture derives partly from uncertainty about their long-term impact on humans, the food chain and indeed the biosphere as a whole, and also from their commercial nature; poorer populations may not derive the benefits of the new crops if they can’t afford the licence to use them. Certainly it is odd to think that we now share our world with life forms that are patent-protected for the benefit of companies like Monsanto of the U.S.

  Transgenic animals have also been produced. In 2009 the U.S. Food and Drug Administration gave its first approval to a human-intended drug produced from such an animal: a goat, from whose milk the drug can be extracted. Other animals have been produced for the purpose of biomedical research, and to produce human hormones such as insulin. In 2009 a Japanese firm announced the first transgenic primate, a marmoset. Among the most spectacular applications of all is the Enviropig (the name is trademarked) produced by scientists at a university in Ontario, Canada in 1999, which is said to produce less phosphorus in its manure than unmodified animals.

  From Enviropig to avatar! But they are both transgenic animals produced for a specific purpose.

  And there’s more to an avatar than simple gene-swapping.

  If it were an entirely terrestrial creation, say a mix of human with lemur traits, an avatar would still be an impressive enough application of genetic engineering—far beyond our capabilities today, but whose principles we can clearly grasp. But an avatar is more than this. Jake tells us that avatars are grown using a “mixture of human and native DNA.” A driver’s avatar is derived from his own DNA (or in Jake’s case, his identical twin brother’s). This is necessary to facilitate the synchronising of nervous systems between avatar and driver that makes the psionic link possible.

  But the Na’vi are alien creatures, from another star system entirely! How can their “DNA” be “mixed” with ours? Why should they even have “DNA?”

  Actually when Jake refers to “native DNA” he’s using the term more generally, to refer to “genetic material” rather than the specific molecule. (To be fair Jake is a Marine; his brother was the science guy…) The purpose of our DNA is to carry genetic information from one generation to the next, and then to use that information in the building of a new life form. “DNA” itself is a term for the specific molecule, deoxyribonucleic acid, that carries out that function for life on Earth. The Na’vi have a similar system but far from identical, based on a different biomolecular set and with a different logic to the coding. Their equivalent o
f DNA is called NVTranscriptase. (This is an example of how, while the Na’vi are similar to humans externally, they are quite different internally—as proven by the dissection of “specimens.”)

  The “mixing” of human and Na’vi genetic material to create a hybrid avatar is done at a logical level. Information from both coding systems is extracted into a computer store, mixed using a translation table, and then downloaded into a third biochemical substrate, the genetics of the avatar.

  The resulting hybrids are more Na’vi-like than human, though they have inherited some human features, such as smaller eyes, five-fingered hands. It remains to be seen whether the genetics will allow avatar-Jake and true-Na’vi Neytiri to have children…

  However it’s done, growing an avatar in a tank is one thing. Now we must consider an even harder step: linking Jake Sully’s consciousness to it.

  32

  HACKING THE BRAIN

  As a concrete example of the challenges involved in establishing a mental link between Jake and his avatar, let’s consider the scene in which avatar-Jake captures his great leonopteryx by falling down from the sky onto its back. As this is happening human-Jake is motionless in his tank. And yet Jake senses everything the avatar senses, and commands every aspect of its conscious movements. He feels the impact as the avatar lands on the creature’s back, feels the surge of acceleration as the indignant leonopteryx flies off.

  How could you make this work?

  To some extent Jake is like a player of a virtual reality (VR) system, with the “game” being Pandora as a whole. A virtual reality system feeds what is not real into our senses, well enough to enable us to believe that it is real—or at least well enough to suspend our disbelief.

  And in some aspects existing systems do this pretty effectively. A music system is a VR system for the ears, fooling you into imagining there’s a rock band or a symphony orchestra in the room with you. The best modern high-fidelity systems have reached such a level of detailed simulation that the ear can’t tell the difference from the reality. For sight, too, watching the movie Avatar itself in 3-D gives you a flavour of what’s possible in delivering a convincing simulation.

  So suppose you constructed an “avatar” like a high-tech robot, laden with cameras, microphones and other sensors. Jake meanwhile is in a wraparound suit with earphones, goggles and maybe with sense-stimulating plugs in his nose and mouth. He is in a motion-capture system of the type Quaritch uses to control his AMP suit, with the machine’s motions aping his own body’s gestures—or like the modern Wii game system. As the leonopteryx looms below the falling robot, you could imagine an all but perfect sensory simulation of the experience being relayed to Jake by all the little cameras and microphones and other sensors: he smells the leonopteryx’s leathery stink, an aroma simulated in some miniature chemical factory, and feels the rushing air of Pandora in his face, blown by tiny fans.

  But this is a simulation which would end in dismal disappointment as soon as the robot hit the back of the animal with a shuddering crash—and Jake felt nothing of the impact.

  Oh, you could provide human-Jake in his tank with some token jolt, like the little bumps you get in a fairground-ride flight simulator. But here we’ve reached the limit of modern VR technology. We don’t know any way to build systems external to the body to simulate the inner sense of the sharp deceleration that ends a fall, or indeed the acceleration that comes with a rocket launch, say. That’s why astronauts train for zero gravity by floating around in tanks of water, or in planes which make powered falls to provide the illusion of zero gravity for a few seconds: “Vomit Comets.”

  You can list plenty of other “inner” sensations Jake needs to experience fully the avatar’s reality. He could be made to feel the Pandoran fruit in his hand, he could taste the juice in the avatar’s mouth—but how could he be made to feel hungry, when the avatar is hungry?

  External VR systems of the kind we have today won’t be sufficient. Just as we see in the movie, it is necessary to hack into Jake’s brain to make this work.

  In the link room we see Jake, preparing to drive his avatar, lie down in a “psionic link unit.” This has an architecture that looks similar to a modern medical scanner, like a magnetic resonance imager. With this, Max Patel and Grace Augustine are able to extract a three-dimensional image of Jake’s brain, complete with ongoing neural activity.

  Then a data link is established between Jake’s brain and the avatar’s, as evidenced by similar-looking images in the scans. The techs speak of achieving “congruency,” as the brains are mapped one to the other. In mathematics, congruent triangles are the same shape and size; you could cut them out and overlay them exactly, though you might have to turn one over to do it. The word is also used in psychology to mean internal and external consistency of the mind. Ultimately “phase lock” is established between the two nervous systems.

  What is happening is that the technology is hacking into the input-output systems of Jake’s brain. When he’s outside the link unit, Jake’s brain is connected to his body by a set of neural connections. Sensory information comes flowing into the brain through these connections, and Jake’s commands for his body—lift that arm, jump from that banshee—flow out of his brain. What the link technology has to do is hack into this flow of data, and into the similar flow of data in and out of the avatar’s brain. Sensory input coming in from Jake’s own body must be ignored, and replaced with the data flowing from the avatar’s body. Similarly Jake’s motor-control commands must be diverted from his own body, and transmitted to the avatar body. And all this is done “non-invasively,” in the jargon; the scanning machine manages all this without the need to stick wires into Jake’s skull.

  This resolves the problem of inner sensation. It’s as if Jake’s brain has been physically implanted in the avatar’s body. Signals arising from the avatar’s inner proprioceptive senses of falling and then slamming to a halt aboard the leonopteryx are now sent direct to Jake’s brain, so that he “feels” the impact in a way he never could using an external suit.

  So that’s the principle. What about the practice? Is this feasible?

  Something like the avatar-link process has been studied in the context of “neuroinformatics.” “Mind uploading” is the process of scanning and mapping a biological brain in detail and transferring that data to a computer, or another machine. Clearly this is like half of an avatar link, with a link to a computer store rather than directly to another brain. And it is like the fate of Grace Augustine, when as her human body dies she passes through the “Eye of Eywa,” to become one with the Great Mother—that is, her consciousness is stored in Pandora’s great biological computer. (In this case Eywa was meant to be used as a temporary buffer; Grace’s mind was supposed to return through the Eye of Eywa and then enter her avatar body.)

  We have taken some baby steps towards this kind of technology today. In “neuroprosthetics” the nervous system is connected directly to some device. And through a “brain–computer interface” (BCI—a variant is BMI, for brain–machine interface) the brain itself is connected to a computer. Researches in the field began in earnest in the 1970s at the University of California, where the term BCI was first coined.

  The first neuroprosthetic applications have been medical, with the aim being the repair of damaged human sensory or motor functions. There have been some attempts to use this technology as an alternative way to treat spinal injuries, like Jake Sully’s. A non-profit consortium called the Walk Again Project has a five-year goal to help a quadriplegic paralysed by a spinal injury to walk again; the patient would use neuroprosthetic devices to control an exoskeleton, an interface reading control signals from the brain to pass to the hardware. The current leading BCI technology is called BrainGate, in which an array of microelectrodes is implanted in the primary motor centre of the brain. In 2008 researchers at the Pittsburgh Medical Center were able to show a monkey operating a robotic arm, with the relevant data being read from the animal’s brain with an
invasive implant.

  As for writing information to the brain, the most common neuroprosthetic device to date is the cochlear implant, in which deafness is alleviated by a device attached to the skull which directly stimulates the part of the cortex that controls hearing: “writing” a signal derived from auditory data to the appropriate part of the brain. There are also neuroprosthetic devices to restore vision, including retinal implants.

  To be able to achieve such feats, you have to be able to understand the brain’s coding of the data it uses: how the firing of a particular set of neurons in a particular way is related to a particular movement of the arm, say. But experiments are proceeding worldwide on reading and understanding motor-control signals, and much more subtle signals, involving mental states associated with language, for example. These are still-tentative steps to something like true mind-reading.

  The U.S. military is interested in this kind of technology; the defence research agency DARPA announced a research programme in March 2010. There are ethical concerns however about using such technologies to go beyond meeting clinical needs to enhancing human abilities beyond the natural limits.

  Most of these experiments involve invasive procedures, in which the patient’s head is literally invaded by bits of wire. Jake’s scanning is non-invasive—no wires. Is this possible? We do have non-invasive neuroimaging technologies. Techniques include electroencephalography (EEG), the reading of brain waves (which dates back to the 1920s), and magneto-encephalography (MEG) and functional magnetic resonance imaging (fMRI), which are capable of producing three-dimensional images of the brain’s electrical activity. The latter techniques exploit the fact that charged particles, such as those passing between neurons in a brain, give off radiation when moving in a strong magnetic field: signals that can be picked up and analysed. Resolution is a problem; the skull itself dampens signals and blurs the neurons’ signals. Progress is being made. A company called G.Tec, based in Austria, already has a non-invasive system that allows users to control avatars in Second Life. Non-invasiveness only adds to the technical hurdles involved in hacking into the brain.