涩里番

On the Use of Computational Models to Improve Understanding of Cardiovascular System Dynamics

Mette Olufsen,聽North Carolina State University
Tuesday September 23, 12-1pm
Zoom Link:听
In Person: 550 Sherbrooke, Room 189

Abstract:听The incorporation of mathematical models in data analysis is essential for understanding physiological processes and developing better diagnosis and treatment strategies. The primary focus will be on the use of computational models for the study of cardiovascular system dynamics. Vascular function is typically assessed either from non-invasive images, including Echocardiograms, CT scans, or MRI, or time series measurements of heart rate, blood flow velocity, and blood pressure. We combine the former with fluid dynamics models predicting flow in vascular networks perfusing a given organ and the latter to examine how the system is regulated, e.g., in response to a head-up tilt test, the Valsalva maneuver, or active standing. Independent of the model type, the use of a mathematical/computational model in clinical practice requires that the model be calibrated and validated against data. Most models are defined by a system of equations, with parameters describing how different model components are connected. Calibrating the model to data involves determining parameters that identify the patient being studied. To do so, we use sensitivity analysis to identify how influential each parameter is, we use subset selection to determine what parameters can be estimated given the model and available data, and parameter inference to determine unique values for the influential parameters. Using this workflow, we can calibrate any of the models studied to data and subsequently perform in-silico simulations to examine how changes can impact observed dynamics. Example studies will explore how to distinguish sympathetic and parasympathetic blood pressure in patients with autonomic dysfunction, the use of micro-balloons in patients with pulmonary hypertension, and how to track patients after a heart transplant.