Font Size: a A A

Causal Inference Based Continuous Cuffless Blood Pressure Estimation

Posted on:2022-02-11Degree:MasterType:Thesis
Country:ChinaCandidate:L LiuFull Text:PDF
GTID:2530307079462254Subject:Biomedical engineering
Abstract/Summary:PDF Full Text Request
Blood pressure(BP)is one of the most important physiological indicators of human body.Cuffless continuous blood pressure estimation is of great significance for human health monitoring and early prevention,diagnosis and treatment of cardiovascular diseases.Enabled by wearable sensing,e.g.,photoplethysmography(PPG)and electrocardiography(ECG),study on cuffless blood pressure(BP)measurement with knowledge-driven and data-driven methods has become popular in recent years.However,causality has been overlooked in most of current studies,which can result in unsatisfactory performance.In this study,we aim to examine the feasibility of causal inference for cuffless continuous blood pressure estimation.First of all,relying on the constraint-based causal inference algorithm,this study proposed a "minority rule majority" causal inference strategy to infer the causal graph between wearable features and BP changes.We found new causal features,directly connected with BP in the causal graph,that could better track BP changes than pulse transit time(PTT)in frequency domain.Finally,there are two cuffless continuous blood pressure estimation models based on causal graph established:(1)inspired by Granger causality test,this study further established a blood pressure regression model by utilizing the time-lagged causal links identified from the causal network,with mean ± standard deviation of estimation error(ME ± SDE)being-0.35±4.2,0.59 ±2.26 mm Hg,respectively;(2)moreover,a causal cuffless continuous blood pressure estimation model based on spatio-temporal graph neural network was established in order to fully explore the topological structure information in causal graph and take advantage of the temporal dependence characteristics from continuous cardiac cycles,with mean standard deviation of estimation error being-0.5±3.94,0.40±2.41 for SBP and DBP respectively.To the best of our knowledge,this work is the first to study the causal inference for cuffless BP estimation,which can shed light on the mechanism,method and application of cuffless BP measurement.
Keywords/Search Tags:Causal Inference, Cuffless Continuous Blood Pressure Estimation, Pulse Transit Time, Time-lagged Causal Links, Spatio-temporal Graph Neural Network
PDF Full Text Request
Related items