Programmable nanoprocessors and neural prosthetics?

17 February 2011

A dream many of us over the years had involve having head computers of one kind or another implanted. Augmentations of our existing capabilities, replacements for damaged sectors, direct neural interface with other computers, encrypted partitions for carrying data, brand new functionality - you name it, chances are there's a geek out there who'd love to beta test it. One of the problems at the moment, however, is a distinct lack of space inside the cranium. When you get right down to it there isn't a whole lot of wiggle room inside your skull. Layering circuitry on the surface of the brain beneath the meninges might work up to a point but it's probable that things are going to shift around a bit, and at best you'll start seeing thin spots or even holes worn into the membranes (to say nothing of lacerations on your favorite brain). You could always squish wiring into the grey matter but you'd be disrupting many thousands of delicate neural connections if you did. If you knew ahead of time the particulars of someone's gyri and sulci it might be possible to fabricate circuitry custom-molded to fit into them. You could always excise bits and pieces of the host brain to make room for circuitry, but then you'd effectively be performing a fractional lobotomy on the subject.

Or, you could make the implants so small that you really wouldn't have to worry about how much space you had.

A research team consisting of scientists from Harvard University and the MITRE Corporation announced that they have just constructed the world's first programmable nanoscale CPU. 960 millionths of a meter on a side, the processor is comprised of a lattice of metallic and semiconducting wires that, when current is applied, functions as a network of logic gates. Right now they've got enough horsepower to perform basic computational operations, but you'd be amazed at what you could accomplish with some elementary arithmetic and a few choice logical operators. Even better, these nanoprocessors are more scalable than today's commodity CPUs (probably due to their comparative simplicity) so you can hook together as many as you care to and they'll function in concert. Oh, and did I neglect to mention that their data storage is non-volatile? Yes, whatever values you store in memory stay there even when the power is turned off.

If that development's not enough to flip your bits, the field of implantable neural prosthetics is beginning to take off. In 2003 we saw the development of the first artificial hippocampus, which serves to replace some of the functions of long-term memory storage, some emotions, mood, and other things (though it's not yet really being used anywhere outside of laboratories yet). Braingate, the first direct neural interface was unveiled in 2008. Braingate was designed to allow users to send commands to external computers to indirectly control portions of their bodies where communication was lost due to disease, damage, or injury. Just a few days ago it was announced that Dr. Eberhard Fetz of the University of Washington at Seattle is the recipient of a $1mus, three year Keck Foundation grant to continue his work on implantable neuroprosthetics. Dr. Fetz's work involves directly interfacing processors with living brain tissue; the self-contained computers are fully implantable and used to record and play back patterns of neural activity from one area of a brain into another. One of the practical results shown in his research thus far is that experimental primates are able to use the implants to bypass damaged or anesthetized nerves to move limbs more or less normally. Other aspects of his work involve inducing or manipulating neuroplasticity, the phenomenon of networks of cortical neurons reorganizing and reprogramming themselves to carry out new tasks.

Remember, everyone: You want the first x.y release of these implants after the first DefCon after they hit the market.