Discussion of Cuffless Blood Pressure Prediction Using Plethysmograph Based on a Longitudinal Experiment: Is the Individual Model Necessary?

Koshiro Kido,Ming Huang,Md Altaf-Ul-Amin,Zheng Chen,Masachika Takeuchi,Shigehiko Kanaya,Wei Chen,Toshiyo Tamura,Naoaki Ono

doi:10.3390/life12010011

Abstract

Using the Plethysmograph (PPG) signal to estimate blood pressure (BP) is attractive given the convenience and possibility of continuous measurement. However, due to the personal differences and the insufficiency of data, the dilemma between the accuracy for a small dataset and the robustness as a general method remains. To this end, we scrutinized the whole pipeline from the feature selection to regression model construction based on a one-month experiment with 11 subjects. By constructing the explanatory features consisting of five general PPG waveform features that do not require the identification of dicrotic notch and diastolic peak and the heart rate, three regression models, which are partial least square, local weighted partial least square, and Gaussian Process model, were built to reflect the underlying assumption about the nature of the fitting problem. By comparing the regression models, it can be confirmed that an individual Gaussian Process model attains the best results with 5.1 mmHg and 4.6 mmHg mean absolute error for SBP and DBP and 6.2 mmHg and 5.4 mmHg standard deviation for SBP and DBP. Moreover, the results of the individual models are significantly better than the generalized model built with the data of all subjects.

Highlights

Ambulatory blood pressure (BP) monitoring provides abundant cardiovascular information, and we have seen numerous studies focusing on replacing the conventional auscultatory/oscillometric measurement that requires the occlusion of arterial blood flow cuffless by using the cardiovascular biosignals with state-of-the-art machine learning methods.The cuffless methods can be roughly categorized into three groups
The first one is based on the pulse arrival time (PAT) [1,2], the second one is based on photoplethysmograph (PPG) signal, which has attracted more and more attention in recent years [3,4,5,6], and the third one is based on other methods [7,8]
Since the PPG signal depends on the blood volume in its optical path, which typically covers the arterial and venous capillaries [10], it can be related to the cardiovascular indices such as the blood oxygen saturation and arterial compliance [11,12]

Summary

Introduction

Ambulatory blood pressure (BP) monitoring provides abundant cardiovascular information, and we have seen numerous studies focusing on replacing the conventional auscultatory/oscillometric measurement that requires the occlusion of arterial blood flow cuffless by using the cardiovascular biosignals with state-of-the-art machine learning methods.The cuffless methods can be roughly categorized into three groups. The first one is based on the pulse arrival time (PAT) [1,2], the second one is based on photoplethysmograph (PPG) signal, which has attracted more and more attention in recent years [3,4,5,6], and the third one is based on other methods [7,8]. It is generally accepted that the PPG waveform comprises the arterial pulse wave from the left ventricle to the distal sites and the reflected wave from the sites of impedance mismatch [13]. With these physiological understandings, research interest has been shifting to BP estimation based on PPG signal recently. Chowdhury et al used an open dataset obtained in a hospital with 219 subjects and 657 measurements to extract the PPG waveform features and biometrics, which were inputted into the Gaussian Process

Methods

Results

Discussion

Conclusion