Higher order spectral analysis for excitation characterization in cough sounds.
Friend-K; Goldsmith-T; McKinney-W; Reynolds-J; Watkins-S; Frazer-D
Ann Biomed Eng 2001 Oct; 29(Suppl 1):S48
Higher order spectral analysis is used to separate the sound source excitation and transmission path during a cough using an autoregressive model base on third order statistics. Volunteer subjects awaiting pulmonary function testing at the West Virginia Hospital were recruited to perform a voluntary cough maneuver. Cough sound pressures were recorded by a perviously described system . Because the respiratory tract changes shape with respect to the excitation with time, it can be modeled as a time-varying filter, H(z), with fixed parameters over shrort time intervals. In order to determine H(z), the cough sound samples were processed through an all-zero digital filter created with an autoregressive routine based on third order cumulants. . The inverse filter was used to separate the excitation source producing the cough sound pressure wave. Coughs were recorded and analyzed from patients diagnosed with obstructive and restrictive lung disease, as well as healthy control subjects.
Models; Sampling; Lung-disease; Pulmonary-system-disorders; Pulmonary-congestion; Respiratory-system-disorders
Abstract; Conference/Symposia Proceedings
Annals of Biomedical Engineering