This ends our tour through the inverse modeling world of lung mechanics, but it is only the tour that has ended. The inverse modeling world itself is limitless in extent, and its known borders continue to be extended by the synergistic forces of advancing experimental methods and increasing computational power. Nevertheless, the landscape we have covered in this book already exhibits widely varying terrain, and so it is easy to lose sight of the fact that all the different regions of this landscape are part of a single inverse modeling world. A brief overview will help bring it all together.

We began by pointing out in Chapter 1 that inverse models of lung mechanics can be usefully grouped into a hierarchy of increasing complexity (Fig. 1.6). The first level of complexity is represented by the single-compartment linear model, consisting of an elastic tissue compartment served by a single flow-resistive airway (Fig. 12.1). This model, which comprises the standard conception of lung mechanics, was explored in Chapter 3 and is characterized by the two parameters lung resistance (RL) and lung elastance (EL). Evaluation of RL and EL can be achieved by applying multiple linear regression to measurements of the independent variables volume (V) and flow () together with the dependent variable transpulmonary pressure (Ptp). Confidence intervals about the estimated values of RL and EL can be determined according to classical regression theory under the assumptions that the noise in the data arises only in Ptp and that this noise is random, uncorrelated, and normally distributed.