Prediction of cardiovascular vital signs using laser speckle flow index signals from a wireless wearable device

Cardiovascular complications account for nearly half of US and global maternal deaths. Monitoring cardiac output (CO), systemic vascular resistance (SVR), and mean arterial pressure (MAP) would enable early detection of cardiac complications. A low-cost wearable laser speckle flow index (LSFI) sensor that captures pulsatile waveforms across the cardiac cycle was developed to accurately estimate CO, SVR, and MAP. Waveforms were collected from swine undergoing blood removal and human subjects undergoing exercise. From these waveforms, features were extracted and used to train multivariate regression models to predict CO, MAP, and SVR. In swine, average absolute errors of 2.13 L/min, 9.0 mmHg, and 464 dynes/sec/cm-5 were produced, respectively. The correlation coefficients for a regression between true and predicted values were r=0.79, r=0.88, and r=0.68, respectively. In human subjects CO and SVR were predicted and produced absolute errors and correlation coefficients of 1.52 L/min and r=0.80 and 278 dynes/sec/cm-5 and r=0.66, respectively.