Humanity Gets an Upgrade at the MIT Media Lab

Two years ago, IFTF’s Technology Horizons program explored the implications of what we called “the extended self”. The core hypothesis was that the body is becoming a platform for a whole range of technological augmentations. These deliberate enhancements run the gamut from mobile phones and social network software, to cochlear implants to restore lost hearing.

One of the really interesting implications of that research was the insight that we are increasingly seeing people leveraging therapeutic technologies to create super-human abilities. This is clearly the case in athletics, but you can also this at work with college students abusing drugs like Ritalin and Adderall to maintain focus during long study sessions. Michael Chorost, who wrote so eloquently about his relationship with his cochlear implant, forecasts that people will almost certainly exploit the potential of that technology to provide super-human hearing for healthy people. Even though only six people have had therapeutic retinal implants to date, it probably won’t be long before someone develops a retinal implant that gives its symbiant the ability to see X-rays or infrared.... if they haven't already.

It was with this in mind that I set out this Wednesday on the 7am shuttle to Boston to attend the Humans 2.0 symposium at the MIT Media Lab. Bearing in mind that Media Lab events are often a big, over-produced show designed to impress sponsors into coughing up another year's worth of funding, I had low expectations. As you can see form the archived video of the day, it was actually a pretty good event all things said.

John Hockenberry, the former NPR/NBC correspondent who is now in residence at the lab, emcee-ed the day and kept everyone awake with wise-cracks, innuendo and jokes. Since Hockenberry has been paralyzed from the waist down since a car accident in 1977 at age 19, it provided a great context for thinking about upgrades to the body.

Most of the presentations by Media Lab researchers were incremental updates on ongoing research, some of which has questionable relevance to the idea of human augmentation. (And nary a Q&A session was offered). If there was an over-riding or under-lying theme to the days presentations, I didn’t get it. But then, that’s what happens at a place like MIT - so many bright people with money inevitably pursue their own desires. The Media Lab especially is notorious for having as many “Centers” as it does professors.

Here are some thoughts about the various presentations.

The Highlight: Progress Towards Neural Interfaces

The highlight of the day, in terms of scientific interest, were two excellent presentations by non-MIT people. These presentations described current research on and the future potential of neural interfaces - technical devices that would allow the human nervous system to connect to electronic devices. Essentially, a network bridge between the human nervous system and digital electronic networks.

Douglas Smith of the University of Pennsylvania shared some results of his work, which focuses on quickly growing long strands of nerve cells in culture using a mechanical stretching device that mimics the tension that stimulates nerve growth in large, rapidly growing animals like whales. In his lab, researchers have been able to grow nerve strands as long as 10 cm at a rate of 1 cm per day. The cells in the strands self-organize into cohesive linear structures, and they have had success taking these strands, wrapping them in collagen and inserting them into rats that have had existing nerve segments removed. While they are not yet sure if the grafted segments will allow normal nervous system functioning to be restored (i.e. to undo the surgically created paralysis), they have proven that electrical current is passing along the nerve-graft boundary, indicating that new connections are being formed.

John Donoghue, who teaches at Brown but is also foudner of a company called Cyberkinetics, contined the discussion on neural interfaces. He began by likening the current state of research to where electronic Pacemakers were in the 1950s, and showed a slide of the first Pacemaker - a device about the size of a large dishwasher. He went on to describe the small, pill-sized sensor that his company has developed for implantation in the cerebral cortex in a region that controls movement of the arm - the BrainGate neural interface system. It has 100 tiny micro electrodes that sense action potential in motor nerves. It is being implanted in people paralyzed through various injuries and disorders. It allows people to "think" about moving their missing or paralyzed arm, and use that output to control a cursor on the screen. The next step will be to couple BrainGate with functional electrical stimulation systems that are already widely used (though with external siwtches) to bridge damage in nerves and the spinal cord (5 years or so).

The Show-Stopper: Biomechatronic Prosthetics

From a showmanship point of view, first prize goes to Hugh Herr, director of the newly formed MIT Biomechatronics Group, who demonstrated an active ankle prosthesis that is essentially a robot itself, actively powering his walk like a real human ankle does. In a dramatic unveiling, Prof. Herr lifted his own pants leg to reveal that he himself (a double amputee) was wearing his own invention, as well as conventional leg prosthesis. While the device does not use a neural interface - it achieves its great functionality largely by sensing orientation and acceleration and driving actuators to stabilize itself and deliver thrust - seeing this device makes you realize just how close we are to robotic prosthetic devices as an everyday medical technology.

Off-topic: Recording an Entire Life

Another highlight was a presentation by Deb Roy, the soft-spoken director of the Media Lab’s Cognitive Machines Group, who described the “ultra-dense observational analysis” he is performing to record every waking moment of his infant son’s development through a network of video cameras installed in their home. This work was recently the subject of a feature-length article in Wired but Prof. Roy showed several fascinating extracts of the massive data set (250 terabytes or 200,000 hours of video by the end of the 3-year collection phase). The best was a time-lapse audio clip of about 150 instances of his son saying the word ball, from first identifiable use to final mastery. And so while Roy is primarily interested in using the data to develop computers that are better at learning by modelling how children develop language skills, this technology has many applications - the most obvious being for memory augmentation. But while fascinating, it was hard to link Roy’s work directly to the human augmentation theme of the symposium. He explicitly is clear that while this technology could be used for memory augmentation, that’s not his goal - he’s trying to mine the data to build software that learns more like children do.