Estimating Blood Pressure Measurements through an Electro-vascular-gram: A pilot study

The ability to accurately measure blood pressure (BP) is considered an important clinical tool in improving the health of patients with cardiovascular disease. In this study, we investigated the use of a cuff electrode as a potential means of predicting arterial BP. Acute experiments were conducted in urethane anesthetized male rats (n = 3), where electrical artifact signals - termed the electro-vascular-gram (EVG) - were measured using a cuff electrode surgically implanted around the right carotid artery. The left carotid artery was catheterized to directly measure BP (ground truth). The EVG signal was processed and trained using a convolutional neural network (CNN), which could predict systolic (SYS) and diastolic (DIA) pressure values (with an error < 5mmHg) in 45-98% and 52-97% of the testing datasets, respectively. Analysis of this preliminary data suggests that it is feasible to use the EVG signal to measure BP in a continuous manner.