 |
1997-98
NEW MEXICO
HIGH SCHOOL
SUPERCOMPUTING
CHALLENGE
Interim Report |
The goal of our project is to accurately simulate nerve impulses in common neurons. By
simulating the neurons we hope to better understand neural activity and aid in the invention of
devises that simulate nerve impulses. Although there are four classes of neurons, all neurons
have five common components including the cell body, the dendrites, the axon, the axon hillock,
and the synapse. The cell body contains the nucleus and much of the metabolic machinery of
the cell and can vary in size and shape. The dendrites are usually numerous, short, threadlike
cytoplasmic extensions that receive stimuli from other cells. The axon is a long process that is
capable of rapidly conducting an electrochemical signal, the nerve impulse, over great distances,
reaching up to 2 meters in humans. Axons are also known as nerve fibers. The axon hillock is
the sight where the axon is joined to the cell body and is where the electrical firing known as
action potential usually occurs. Finally, the synapses are the junctions formed with other nerve
cells where the presynaptic terminal of one cell comes into contact with the postsynaptic
membrane of another. One of the fundamental activities that we must take into consideration
when simulating a nerve impulse is the electrical propagation of the impulse along neural network.
Sodium, Na-, and potassium, K+ ions play a key role in maintaining the nerve impulse. Nerve cells
maintain a gradient in both sodium and potassium. The sodium-potassium pump keeps a high
concentration of potassium inside the cell and a low concentration of sodium inside the cell. As
a nerve impulse moves along a neuron, sodium rushes into the cell and potassium rushes out.
This depolarizes the next adjacent area of the membrane, opening Na- ion channels to open,
allowing the process to be repeated.
In order to accurately simulate nerve impulses we will create a computer model containing
only as many of the components of actual neurons as are required. In our research we have
discovered a compartmental model that breaks neurons into compartments based on function.
A problem with some of the current models, is that they do not take into consideration the
length of time it takes a neuron to regain its normal chemical gradients so it can successfully
conduct another nerve impulse. We intend to break the problem into small parts that become
increasingly detailed and create a simpler model that yields comparable accuracy. Due to the
large scale of neural interaction the use of a supercomputer will be necessary. We are currently
working on determining what details of a neuron we must accounted for in our simulation. We
plan to build our program from a very simple neural model adding details until have created a
sufficiently accurate model. Using this method we can solve problems as they arise and will have
a completed project at each successive revision of the neural model. The program will be written
in C and will produce text output that will be displayed graphically using AVS (an advanced
visualization program).
Team Members:
Sponsoring Teacher(s):
Project Advisor(s):
New Mexico High School Supercomputing Challenge