Analysis of Pulse Transit Time Derived From Imaging Photoplethysmography and Microwave Sensor-Based Ballistocardiography

Mototaka Yoshioka, Souksakhone Bounyong; Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2020, pp. 286-287

We present a basic analysis of the pulse transit time derived from imaging photoplethysmography and ballistocardiography. The pulse transit time is considered as an indicator of blood pressure. For convenient estimation of the blood pressure, previous studies have used the time delay between electrocardiography and photoplethysmography using contact based sensors, which is known as the pulse arrival time. In this paper, we propose a noncontact system to measure the pulse transit time, which consists of microwave and image sensors. The microwave sensor allows ballistocardiography from tiny body movements generated by the human heartbeat using reflected wave signal from a subject's chest, and the image sensor enables imaging photoplethysmography of subject's face. By temporally synchronizing two noncontact sensors, the proposed system is able to provide an estimate of the pulse transit time remotely. We conducted experiments on 16 subjects (age range of 69 to 79 years old) with a supine posture. The correlation coefficient between the noncontact pulse transit time and systolic blood pressure was -0.64 (P<0.05). The pulse transit time had a better correlation with systolic blood pressure than the pulse arrival time, which was -0.20. This result indicates that the pre-ejection period influences the pulse arrival time. The pre-ejection period calculated from electrocardiography to ballistocardiography ranges from 54 to 130 ms owing to individual differences. This is an important finding for noncontact blood-pressure estimation in the future.