| In this thesis, model based excitability analyses of the neuromuscular cell and network during resting and refractory statuses are carried out for the application purposes of stimulation design, electromyographic signal interpretation, and associative memory study.; Simplification on the Hodgkin-Huxley (H-H) model, a circuitry membrane model, is carried out in our study for neuronal threshold analysis and prediction during resting status and refractoriness. The results suggest that the thresholds explicitly predicted by our method are consistent with those by circuitry membrane model estimations; and that the new threshold prediction method is much simpler than the threshold estimation by the H-H model. The threshold prediction method based on the simplification of H-H model also is proved to be valid for predicting the threshold of cat sciatic nerve, by comparing the data obtained from experiments, estimated by the full circuitry membrane model, and predicted by our method.; For electromyographic (EMG) signal interpretation, a single channel linear system model incorporated with absolute refractoriness at the post-membrane of the neuromuscular junction is developed for the studies of motor unit firing in both time and frequency domains. When analyzing the effects of the absolute refractory period variation on the output signals, consistence between the theoretical and simulated results is found given the impulse stimuli with exponentially and normally distributed inter-pulse intervals. Given the gradually increased absolute refractory period, variations of the output mean firing rates and standard deviations are found for both exponential and normal cases in the motor unit firing statistic analyses. The results obtained in power density spectral analysis of the motor unit output suggest that the increased absolute refractoriness affects the overall shape of the output spectra.; The integrate-and-fire neuron model without considering the refractory period has been organized into a network, the bifurcating neural network (BNN), by Lee and Farhat for associative memory study in 2001. The absolute refractory period is introduced into the individual neuron model of the BNN to construct the refracted bifurcating neural network (RBNN) with an attempt to improve the convergent speed of the network in our study. The study results of the refractory effects on single neuron model activities suggest that absolute refractory period acts as another potential parameter controlling the arrival of the crisis point of single neuron, besides the amplitude of relaxation level. The comparison study between RBNN and BNN indicates that introduction of refractory period into BNN does not reduce the pattern recalling rates. (Abstract shortened by UMI.)... |
