Multiscale Computational Modeling in Vascular Biology: From Molecular Mechanisms to Tissue-Level Structure and Function - Multiscale Computer Modeling in Biomechanics and Biomedical Engineering - page 226

Fig. 4 Sensitivity of the number of smooth muscle cells (SMC), value of hoop stress and
concentration of PDGF to changes of the m parameter in the chance of SMC proliferation
function. Note the interdependency between PDGF, hoop stress and number of SMC. Outputs are
able to obtain a new equilibrium to small changes of the m parameter
produced as the wall thickens (Fig. 4 ). Ensuring proper responses to simple per-
turbations, as illustrated by this example, can help identify appropriate bounds to
place on free parameters.
Finally, note that multi-cell ABM rules are often based on cell level in vitro
experiments, which can be well controlled but do not always have direct physi-
ologic relevance. CMMs are typically based on either in vitro or in vivo tissue-
level experiments, which can have considerable physiological relevance but often
little control of the inputs sensed by the cells (e.g., humoral in addition to
hemodynamic). Because each data set will be inherently limited, some advantage
can be gained by ensuring that models of the same processes based on very
different data should yield the same result [ 29 ]. The third step, therefore, is to
enforce congruency across two (or potentially more) scales through parameter