Michael Chorost on Cochlear Implants and Transhumanism

On August 31st, Michael Chorost, author of Rebuilt : How Becoming Part Computer Made Me More Human, spoke at the Institute. Below are my notes from the talk (which haven't been vetted by the author, so all caveats apply--- quotes are approximate, and you should assume that the overall shape of these notes reflects my attention and interpretation, not what Mike actually said or meant).

(from the IFTF flickr photoset)

Technorati Tags: communication, cyborg, digital/physical, IFTF, paloalto, Silicon Valley, work

Rebuilt is the first cyborg memoir, the story of how Chorost lost his hearing, got a cochlear implant, and learned to hear again-- this time with the world of sounds mediated by a set of microphones, microprocessors, and electrodes. A cochlear implant is really a couple devices: the implant that sits on the surface of the skull, with an thin wire running into his inner ear. At the end of the wire are six electrodes. Under normal circumstances, the cochlea transduces sound vibrations into electrical impulses that travel to the brain; many cases of deafness (like Chorost's) are caused by a failure in the cochlear, which leaves it unable to translate motion into electrical signals, but still able to send those signals to the brain. An implant bypasses all the ear's delicate sound transmission and transduction mechanisms, and feeds signals directly to the cochlear nerves.

Learning how to hear speech and music through "electrodes strobing on and off" is not small task. While there's plenty that a technician can do to calibrate and tune the implant, your brain has to learn to take this new input and translate it into something you can recognize. To do this, researchers have found that you need at least eight channels (a channel covers a range of frequencies) to figure out speech. (A normal ear, by contrast, has 3500 channels.)

Chorost reports that with the implant, environmental sound is better; speech is somewhat better (even though it sounds highly processed and filtered to someone with normal hearing, "it doesn't sound harsh to me... because I've learned to interpret it"). while music is much worse.

After the implant, the software is really important, because there's lots of unused capability in the hardware. (This is no accident: designers of cochlear implants know that software will be a lot easier to upgrade than hardware, so they make the implants as powerful as they can, regardless of what the software can do.) Michael's original software had 16 channels; the latest version delivers 121 channels. With this new version, "speech is harder to understand, but music is much better." No one really knows why; the software is too new, and it's only been out about 6 weeks. Learning to reinterpret sound after the upgrade is the subject of a forthcoming Wired article.

Scientists and medical researchers are currently working on bilateral implantation (i.e., putting implants in both ears, rather than just one); fully implantable devices that have nothing on the outside; use of pharmaceuticals to encourage nerve cells to grow onto the electrodes (which would lower system power requirements); and hair cell regeneration (this was done on guinea pigs about 6 months ago).

Chorost also talked a bit about the transhumanist literature. He describes himself as trying to stake out a middle ground between enthusiasts like Kevin Warwick and critics like Francis Fukuyama; he sees his work emphasizing the complexity of the human body, and the necessity of soft approaches such as training and social infrastructure to gain maximum benefit from whatever technology is developed. (I thought he was the first transhumanist author to really emphasize the "human" over the "trans.")

He's also more skeptical of claims that new technologies will transform humans. Almost all current technologies are used to restore senses or sensory capabilities, rather than extend or enhance. Restoration is the goal of most ordinary people, and is difficult enough: "I am skeptical of potential for enhancement via bionics," on the grounds that our natural sensory organs are fantastically sophisticated, and reproducing them-- or completely new things-- will be very hard to create. Further, in the pre-nanotech state of the art, "bionics is big and clunky: it works on the scale of millimeters. The body works at the level of nanometers. We're not even close" to matching the body's capabilities and scale.