Four and six parameter models of forced random noise respiratory impedance in normals

Total respiratory impedance has been measured between 3-42 Hz by the forced random noise technique, in 15 subjects breathing either air or a helium oxygen mixture in three experimental conditions: at basal state, and then with a resistor or a tube added at the mouth. Impedance is modelled, either by a 4-parameter model (M4), derived from the series model (resistance, inertance, compliance) by making resistance a linear function of frequency; or by a 6-parameter model (M6) including a central compartment (airway resistance and gas inertance), and a tissue compartment (resistance, inertance and compliance in series) placed in parallel with alveolar gas compliance. The additive resistance is perfectly evaluated by both models, whereas the additive inertance is not accurately estimated by the model M6, the fitting of which combines the real and imaginary parts of impedance. Resistance extrapolated at zero frequency on the one hand, inertance of M4 and central inertance of M6 of the other, are highly correlated. However, changes in some parameters of both models according to the experimental conditions are difficult to explain on physiological grounds. We conclude that the model M6 cannot be easily and accurately identified over such a limited frequency range, at least in normals, while the model M4 yields a simplified description of impedance which may be sufficient for diagnostic purposes.