Machine learning estimation of an arterial pressure model using electrical impedance

Abstract

Cardiovascular System Diseases (CVD) are among the most common causes of death and illness in the world. Arterial pressures is a good indicator of CVD existence. However, measuring arterial pressure commonly involves either invasive techniques that require catheter insertion, or noninvasive oscillometric techniques that require inflating a cuff around the arm and don’t provide continuous information. Recently, new methods are being develop to provide continuous, reliable and comfortable measuring of arterial pressure. One promising technique involves using Electrical Impedance (EI) in a highly vascularized segment of the body (such as an arm) to estimate the arterial pressure in that segment. In this paper, we present an experimental setup which includes a gelatin model that emulates some physical and electrical properties of the forearm, an automated system to control pressure and measure EI in such model, and a computational method that makes use of regression algorithms in order to predict the pressure value based on the EI magnitude and phase values