Sensitivity analysis for evaluating nonlinear models of lung mechanics

Area of Science:

• * Computational Biology and Physiology
• * Biomedical Engineering
• * Respiratory System Modeling

Background:

• * Lung mechanics models often exhibit complex nonlinear behaviors.
• * Understanding parameter influence is crucial for accurate diagnosis and treatment.
• * Existing methods may not fully capture nonlinearities in lung tissue and airways.

Purpose of the Study:

• * To develop and validate a combined theoretical and numerical procedure for sensitivity analysis of nonlinear lung mechanics models.
• * To apply this procedure to a comprehensive nonlinear lung model incorporating viscoelastic tissues and airway inhomogeneities.
• * To assess parameter uniqueness and justify the inclusion of specific nonlinear mechanisms.

Main Methods:

• * Development of a theoretical and numerical framework for sensitivity analysis.
• * Application to a nonlinear lung model with viscoelastic tissues and airway resistance distribution.
• * Simultaneous system identification using time- and frequency-domain data.
• * Numerical approximation of sensitivity coefficients and parameter confidence regions.

Main Results:

• * Normalized sensitivity coefficients revealed the relative importance of model parameters.
• * Sensitivity analyses provided justification for nonlinear tissue properties in healthy and constricted lungs.
• * Both airway inhomogeneities and tissue nonlinearities were justified during bronchoconstriction.

Conclusions:

• * The developed sensitivity analysis tools are effective for nonlinear lung models.
• * The findings support the inclusion of specific nonlinearities for accurate lung mechanics modeling.
• * The methodology is generalizable to a broad range of nonlinear systems.