Applications - Quantitative Neurophysiology - page 111

10.6.1 Nerve Guides
Some of themost debilitating injuries effect only a small population of neurons. For example, quadriplegics
and paraplegics often have a gap in their spinal cord that prevents signals from reaching portions of the
body below the gap. If this gap could be spanned, all (or most) function would be returned. In the repair
of neural gaps, creating nerve guides or nerve channels have shown some promise. The idea is to provide
a channel that limits the direction that a nerve can grow. In this way two neurons (or nerve bundles) can
be forced to connect. Typically, this method will only work to span gaps less than 1 cm . To span larger
gaps a nerves can be grafted from some other part of the body. More recently, the material of the nerve
guide has been the focus of much research. In particular, the nerve guide can be made of a bio-degradable
substance which serves its purpose and then is naturally broken down by the body. Some new materials
can be created with embedded growth factors that will release slowly and stimulate both neuron and
capillary growth. Lastly, there is evidence to suggest that Hebb's “fire together wire together” idea applies
in the peripheral nervous system. Therefore, a continual stimulus down the nerve gap may promote a
better electrical connection.
10.6.2 A Biological Flight Simulator
A second application for tissue engineering is to create biological neural networks that will function
similarly to artificial neural networks. These networks of real neurons can take in inputs from some
external source, process the inputs and send out a meaningful output. Similar to an artificial neural
network, the connections between real neurons are plastic, meaning that they can adapt to identify
patterns in inputs.
Perhaps one of the most striking and creative examples of the capabilities of a biological network of
neurons is the work ofThomas DeMarse.Using circuit boards created from special materials, neurons were
grown and interfaced with recording and stimulating electrodes. The neurons, however, were randomly
spread on the circuit and allowed to self-assemble. During this self-assembly, they are being trained to fly
a flight simulation program in much the same way a neural network would be trained. After the neurons
connected, the network could successfully fly the flight simulator without external input. Even more
impressive was that given an unexpected input (e.g., a cross-wind, engine failure), the network adapted
the outputs to compensate. Growing neurons on a circuit board may have some interesting applications
in creating biologically based computers.
Growing an entire brain is far outside the realm of our current technology. The most significant
hurdle is that tissue engineering has not found an effective way to grow three-dimensional tissues. The
fundamental limitation is the transport of O 2 and glucose and removal of waste products. In the body,
these functions are performed by capillaries which are no more than 100 μm from each neuron. A first step
toward overcome this limitation is again in newmaterials (e.g., gels) which allow for vasculature to develop
along side the neural tissue.To complicate matters, neurons in a 3D structure will most likely require glial
cells to perform maintenance functions and possible participate in electrical impulse communication.
10.7 ARTIFICIAL INTELLIGENCE
Artificial Intelligence (AI) is an interdisciplinary field composed of computer scientists, mathematicians,
psychologists, philosophers and engineers. There are two primary goals of AI. First is to mimic the
function of the brain to do something useful (e.g., drive your car for you). Second is to build an artificial
intelligence so we can learn more about the way our own brain works. One fundamental stumbling
block has been to define what is meant by intelligence. To most, intelligence requires memory, pattern
recognition, the ability to generate an orderly sequence of reasoning (logic), the ability to plan ahead and