1 Introduction

The circuit activity in the circuits closest to the motoneurons is of great interest in determining the origin of motoneural stimulation. A considerable amount of data has been collected that describes the underlying circuitry between the motoneuronal and interneuronal populations that stimulate a few mammalian muscles.

The complexity of the system, the many different outputs of the system, and the complexity of the spinal cord make it very difficult to visualize the relationships between circuit activity and dynamics. The complexity of the model makes the movement seem unintuitive given the neural circuit activity. To gain a clearer understanding of the model, it is necessary to take a statistical or population approach. The population approach allows the observer to view how neurons interact with each other to form units of collective function. Simulation is helpful in understanding the spinal circuit activity behind movement.

MacGregor's SYSTM series FORTRAN programs, from which the numerical simulation was developed, produced streams of output data for each discrete time step in the simulation. The data was manually inspected for analysis. Later sorting and statistical tools were used to massage the data and produce additional measures. The process was highly labor intensive and time consuming.

This project eliminates many of the inefficiencies in the process and makes the model more readily understandable in two ways:

The advantage of using data visualization tools for analysis of the output of the simulation is that it allows for better perception of the relationships within large data sets. The human visual system is capable of perceiving complicated patterns, trends, and features in visual data in a more intuitive way than is possible in simple inspection of raw data.

A similar system has been previously implemented by Dr. K. R. Subramanian, Professor at the University of North Carolina at Charlotte. The system was implemented by Dr. Subramanian using the interactive visualization tool kit GEOMVIEW, available f rom The Geometry Center, University of Minnesota. GEOMVIEW supports simple and complex geometry from polyhedra to Bezier patches of arbitrary degree, including rational patches. GEOMVIEW allows for the addition of external user defined models. These modules are independent processes that communicate data through the use of standard UNIX pipes. The implementation was useful and offered an estimated eight fold decrease in the time needed to perform the analysis of each simulation. The main problem with the implementation was the issue of speed. GEOMVIEW can only generate the graphical output for a full twenty-two populations every ten to twelve seconds. It was apparent that a speed up in the graphical simulation was both necessary and possible.

Back to the Table of Contents
Next Section ->