We began our project with the ambition of designing a new and creative method of modeling the interactions of neurons. To accomplish our goal we began by learning about neuron structure, the way neurons connect to each other, and the way they effect each other.

Becoming acquainted with neuron structure is simple, because each neuron consists of only four major parts; dendrites, axon, cell body, and axon terminals. The dendrites receive stimuli, the axons conduct stimuli, the cell body contains the metabolic machinery, such as the nucleus, Golgi complexes, and mitochondria, and the axon terminals connect the neuron to other neurons or to other body cells. In the brain, an intriguing area that sparked our interest in modeling neurons, hundreds of neurons can be connected to one another by their axon terminals. The actual connection and interaction of neurons is more complicated and is explained in detail later.

To attain our objective we also gathered the most useful supplies available to us. We received permission to use the Blue Mountain Supercomputing Platform at Los Alamos National Laboratories. This enabled us to use parallel processing, and minimize the time it took to execute our program. Our project also presented us with the opportunity to learn to use "pixie," a utility that gathers information about different aspects of the program throughout its execution. As we began the actual programming, we wanted to create the framework of a program that could easily incorporate increasingly complicated factors. Using this technique we would have a flexible program that would give us a finished project with the successful addition of each new factor. Now we could avoid setting unattainable goals and ending up with an incomplete project.

As we continued programming we encountered a variety of situations that taught us a great deal about programming. These included I/O bound efficiency, the xpm and ppm graphics file formatting, and the actual multi-threaded parallelization of the program.

After we produced a program that modeled the interaction of neurons, we wanted to be able to see a visual demonstration of the interactions. As a finished product the program produces a file that contains all the information about each neuron throughout the program’s excection. We then created another program that converts this information to xpm or ppm graphics files. A program called MPEG Encode converted the graphics files to an Moving Pictures Experts Group (MPEG) movie.

As the final product of our endeavors we produced efficiency diagrams to measure the validity of using multiple processors. We predict that the efficiency of multiple processors and the accuracy of the model will increase as more complicated factors concerning the interactions of the neurons are incorporated into the program. Although there isn"t another program of its kind to use as a comparison, we have shown that our method of modeling the interactions of neurons is successful, and can easily be modified to fit changing conditions.