COMPUTATIONAL RESEARCH in BOSTON and BEYOND (CRIBB)

The autonomic nervous system (ANS) regulates bodily functions that are not under conscious control, such as heart rate, blood pressure, digestion, etc. The ANS integrates information from the body as a whole and its activity reflects perturbations caused by various disease processes. We aim to develop a real-time method to noninvasively estimate the activity in the two branches of the ANS for use in diagnostics or patient monitoring.

The current state of the art for cardiac ANS estimation falls under the topic of heart rate variability (HRV) or cardiovascular system identification (CSI). HRV and CSI have shown promise for diagnosing and tracking the progression of diseases such as hypertension, diabetic neuropathy, heart failure, sleep apnea, and others, as well as quantifying the consequences of lifestyle changes such as smoking, diet, and exercise. An opportunity exists to improve the existing methods to increase the time-resolution and provide more easily interpretable results.

To improve the existing ANS estimation methods, we are incorporating more physiologic detail into the model of the system, and plan to use this model to more thoroughly constrain the estimation problem linking heart beat times to ANS tone. This talk will cover the details of translating biological data into a mathematical model of the sino-atrial node cell (the pacemaker of the heart), with a focus on the compromises that must be made between capturing biological detail and ensuring computational feasibility and mathematical clarity. The model is realized as a system of nonlinear ordinary differential equations (ODEs); we also describe a preliminary implementation using a numerical ODE integrator in serial (ode15s in Matlab) versus parallel (CVode in Star-P).