Musings of a Compulsive Thinker

More on the pet project

In the broadest sense, a cybernetic interface is one that provides a communication between man and machine. A PC keyboard falls into this category. When most people think of cybernetics, however, what usually comes to mind is actually a neuroelectric interface, which serves as a sensing and processing system that uses signals from the brain or peripheral nerves as control input for some piece of hardware and/or software.

What we'd really like for the purposes of this "plug-in prosthesis" is a two-way neuroelectric interface. Ideally, this will allow "outgoing" motor nerve signals to be sent to the prosthesis, and in addition the interface will translate "incoming" signals from sensors in the prosthesis and transmit them back up the sensory nerve pathways to the brain. This is not as "sci-fi" as it may sound.

An important goal in realizing a fully-functional, two-way neuroelectric interface is the first step in a bottom-up approach: the development of a functional two-way interface between an electromechanical "terminal" and individual neurons (nerve cells) or a neural network (e.g. an individual nerve bundle leading to one muscle). Some important issues that need to be addressed in the design of such an interface are as follows:
- Neuronal and/or network plasticity ("recovery") following stimulation
- Biocompatibility of interface materials (material degradation, toxicity; how the body responds to the material, i.e. does it cause irritation or inflammation in the surrounding tissues?)
- Insulatory and discriminatory behavior (it should prevent signal "slop" to unintended recipient neurons)
- Separation/isolation of sending and receiving pathways (it should prevent feedback and interface-generated noise)

Two separate groups approached this design problem. Stett et al. designed, constructed, and successfully tested a silicon micro-structure that allowed two-way communication with an individual neuron; Reiher et al. used a titanium-gold electrode-based interface to simultaneously stimulate a large area of a neural network. If you're so inclined, you can find the articles through the links below.

Stett et al.: "Two-way silicon-neuron interface by electrical induction." [PDF, 180 K]

Reiher et al.: "In vitro stimulation of neurons by a planar Ti-Au-electrode interface." [PDF, 153 K] (HTML version)

I realize that to many people this may sound like a bunch of technical mumbo-jumbo. If you don't understand all of it, well, the only thing you really need to glean from this is the fact that for the purposes of my little conceptual design exercise, I consider the neuroelectric interface problem to be "solved." However, if you have questions about any of this or simply think it's cool, feel free to comment. I'll be happy to explain things, discuss further, and/or otherwise "get my geek on."

I'm weird enough that I actually enjoy killing time by thinking about this kind of stuff.

--

Labels: engineering, science

posted by the alygator @ 4:00 PM