A model for normal swallowing sounds generation based on wavelet analysis

In this paper a new model for swallowing sounds generation is proposed. The model consists of two systems: one that simulates the movements and activities of the muscles and the bones in the pharynx and the interactions between the bolus and the pharynx structure, followed by the second system representing the transfer function of the esophageal wall structure, tissue and skin beneath the microphone as the bolus travels through. Cepstrum analysis was used to estimate this transfer function. Based on the estimated transfer function, Symlet wavelet of order 8 was used to find the impulse train representing the output of the first part of the model. To have a general understanding whether such a model can be a representative of swallowing sound generation, its performance in describing differences in swallowing sounds characteristics for two bolus textures with different viscosities was evaluated. The results show the model may predict the changes in the swallowing sounds characteristics and the outcomes of the model comply with the common knowledge of the swallowing mechanism.