UC Berkeley

In medical conditions with disrupted cerebrovascular flow, including chronic alterations throughout aging or acute events such as ischemic stroke, cerebrovascular modeling can provide mechanistic and treatment insights. Reduced-order cerebrovascular modeling can be particularly useful for patient-specific or population-averaged cerebrovascular hemodynamics, where the governing flow dynamics include physically-driven non-linear constraints with active homeostatic regulation as well as highly variable and interconnected cerebral blood vessel networks. A coupled one-dimensional lumped parameter network (1D-LPN) cerebrovascular model is described with myogenic and CO2 reactivity autoregulation modeling, circle of willis anatomical variations, cortical collateral flow compensation, cerebrospinal fluid generation/reabsorption, and intracranial pressure dynamics modeling. For parameter tuning of the numerous cerebrovascular model variables, a Bayesian optimization procedure establishes crucial parameter values with the remaining model parameters assigned through vascular aging and cerebral autoregulation quantitative population-averaged reviews. A specific focus on cerebrovascular aging alterations, a major source of vascular variance besides patient-specific variations, explores the parameter tuning procedure in a young population. The coupled 1D-LPN cerebrovascular model design, age-related vascular critical review, cerebral autoregulation critical review, and model parameter tuning procedure, comprises a prototype vascular test bed with the capability to model cerebrovascular-specific medical conditions while accounting for age-related vascular alterations.