Analysis of the Contribution of Cardiovascular Components to the Ballistocardiogram Signal Using a Mathematical Model

Mohamed Zaid1, Raul Invernizzi2, Lorenzo Sala3, Laurel Despins4, Mihail Popescu5, Marjorie Skubic6, James Keller6, Riccardo Sacco7, Marcela Szopos8, Virginia Huxley9, Giovanna Guidoboni10
1University of Maine, 2Departement of Civil Engineering, Politecnico di Milano, 3INRAE, MaIAGE, Universite Paris-Saclay, 4Sinclair School of Nursing, University of Missouri, 5Health Management and Informatics, University of Missouri, 6Electrical and Computer Engineering, University of Missouri, 7Department of Mathematics, Politecnico Di Milano, 8Université Paris Cité, CNRS, MAP5, 9Department of Medical Pharmacology and Physiology, University of Missouri, 10Electrical and Computer Engineering, University of Maine


Abstract

The ballistocardiogram (BCG) offers a more comprehensive assessment of cardiovascular performance than conventional monitoring techniques by capturing the mechanical and fluidynamical characteristics of the cardiovascular system as a whole. While having the great advantage to be acquired noninvasively, the use of BCG as a clinical monitoring and diagnostic tool is limited due to the lack of understanding the cardiovascular mechanisms responsible for changes in the signal. Computer-aided approaches can help provide a mechanistic and quantitative interpretation of BCG signals. To this purpose, Guidoboni et al. (IEEE TBME; 2019) proposed a closed-loop mathematical model based on the physical principles governing vascular physiology to provide a theoretical estimation and interpretation of BCG signals. Using the above-mentioned cardiovascular model as a reference, in this study we assess and analyze the contribution of different circulatory components, including ventricles, aorta, systemic circulation, cerebral circulation, and iliac arteries, to the BCG waveform. Starting from the model by Avanzolini et al (Int. J. Bio-Med. Comput.; 1988), we consider 5 sub-models with step-by-step increased levels of detail in the main arteries and the cerebral arteries (Table 1). In addition to the BCG, we will also analyze pressure and volume (PV) loops in the left ventricle simulated by each model, which is particularly relevant in determining cardiovascular performance and functions. Results show a considerable improvement in the BCG theoretical reconstruction with the addition of the aorta. The addition of the cerebral circulation component shows the smoothing of the pressure waveforms in all the aortic segments and yields the ventricular pressure closer to physiological values.