Photoplethysmography-Based Blood Pressure Estimation Combining Filter-Wrapper Collaborated Feature Selection With LASSO-LSTM Model

Wang, Dingliang; Yang, Xiu; Liu, Xuenan; Mao, Shaolin; Li, Longwei; Wang, Wenjin

doi:10.1109/tim.2021.3109986

“…7–9 Photoplethysmography (PPG) sensors are non-invasive devices that measure variations in blood volume within the vascular system by detecting changes in light absorption or reflection. 10 They utilize light-based techniques to assess changes in the volume of blood vessels, primarily through the detection of variations in light absorption or reflection. This enables the estimation of parameters, such as pulse rate, blood flow, and oxygen saturation without the need for invasive procedures.…”

Section: Introductionmentioning

confidence: 99%

Improving accuracy of vascular access quality classification in hemodialysis patients using deep learning with K highest score feature selection

Julkaew,

Wongsirichot,

Damkliang

et al. 2024

J Int Med Res

0

View full text Add to dashboard Cite

Objective To develop and evaluate a novel feature selection technique, using photoplethysmography (PPG) sensors, for enhancing the performance of deep learning models in classifying vascular access quality in hemodialysis patients. Methods This cross-sectional study involved creating a novel feature selection method based on SelectKBest principles, specifically designed to optimize deep learning models for PPG sensor data, in hemodialysis patients. The method effectiveness was assessed by comparing the performance of multiple deep learning models using the feature selection approach versus complete feature set. The model with the highest accuracy was then trained and tested using a 70:30 approach, respectively, with the full dataset and the SelectKBest dataset. Performance results were compared using Student’s paired t-test. Results Data from 398 hemodialysis patients were included. The 1-dimensional convolutional neural network (CNN1D) displayed the highest accuracy among different models. Implementation of the SelectKBest-based feature selection technique resulted in a statistically significant improvement in the CNN1D model’s performance, achieving an accuracy of 92.05% (with feature selection) versus 90.79% (with full feature set). Conclusion These findings suggest that the newly developed feature selection approach might aid in accurately predicting vascular access quality in hemodialysis patients. This advancement may contribute to the development of reliable diagnostic tools for identifying vascular complications, such as stenosis, potentially improving patient outcomes and their quality of life.

show abstract

“…The addition of PRV reduced the root-meansquare error (RMSE) of systolic BP (SBP) from 6.32 to 5.71 mmHg for healthy subjects [37]. Some studies used timedependent information by employing the long-and short-term memory (LSTM) network [35,39,40], multi-stage feature extraction [22,41], or dynamic compliance [42]. These algorithms showed better performance compared to algorithms without dynamic features [23,24].…”

Section: Introductionmentioning

confidence: 99%

Contribution of Timescale Patterns to Photoplethysmography-based Blood Pressure Estimation

Xing

¹

,

Huang

²

,

Jiang³

et al. 2023

Preprint

0

View full text Add to dashboard Cite

Objective: Single-site photoplethysmography (PPG)-based blood pressure (BP) estimation has raised a lot of interest due to its compactness and low cost. However, this method relies on PPG morphological features, which are sensitive to noise and measurement conditions. The underlying physiological mechanism was also unclear at this moment. In this study, we propose to add timescale patterns of PPG to improve the BP estimation performance and clarify the underlying mechanism. Methods: In-silico simulation with a four-element Windkessel model showed that peripheral resistance and vascular compliance variation during a cardiac cycle correlated with PPG’s long- and short-term self-similarity, which significantly correlated with BP. A publicly available dataset was used to validate the simulation predictions using mutual information analysis and regression performance assessment. Results: The hemodynamic property of the cardiovascular system determines how fast PPG responds to a stimulus. High vascular compliance or peripheral resistance leads to prolonged and overlapped responses, which could be described by timescale patterns. Adding these patterns significantly increased the PPG-BP mutual information and improved the BP estimation performance. Compared to algorithms with morphological and biometric features, the mean absolute error (MAE) of calibrated systolic BP (SBP) reduced from 5.37mmHg to 4.51mmHg, while the MAE of calibration-free diastolic BP (DBP) reduced from 3.46mmHg to 2.81mmHg. The median intra-subject correlation between SBP/DBP estimation and ground truth increased from 0.63/0.34 to 0.80/0.68, which means the intrinsic BP fluctuation was better captured. Conclusion: Timescale patterns were vital to single-site PPG-based BP estimation. Understanding its physiological implication may help us design algorithms with clear interpretability and simplified structures.

show abstract

“…The addition of PRV reduced the root-meansquare error (RMSE) of systolic BP (SBP) from 6.32 to 5.71 mmHg for healthy subjects [37]. Some studies used timedependent information by employing the long-and short-term memory (LSTM) network [35,39,40], multi-stage feature extraction [22,41], or dynamic compliance [42]. These algorithms performed better than those without dynamic features [23,24].…”

Section: Introductionmentioning

confidence: 99%

Contribution of Timescale Patterns to Photoplethysmography-based Blood Pressure Estimation

Xing

¹

,

Huang

²

,

Jiang³

et al. 2023

Preprint

0

View full text Add to dashboard Cite

Objective: Single-site photoplethysmography (PPG)-based blood pressure (BP) estimation has raised a lot of interest due to its compactness and low cost. However, this method relies on PPG morphological features, which are sensitive to noise and measurement conditions. The underlying physiological mechanism was also unclear at this moment. In this study, we propose to add timescale patterns of PPG to improve the BP estimation performance and clarify the underlying mechanism. Methods: In-silico simulation with a four-element Windkessel model showed that peripheral resistance and vascular compliance variation during a cardiac cycle correlated with PPG’s long- and short-term self-similarity, which significantly correlated with BP. A publicly available dataset was used to validate the simulation predictions using mutual information analysis and regression performance assessment. Results: The hemodynamic property of the cardiovascular system determines how fast PPG responds to a stimulus. High vascular compliance or peripheral resistance leads to prolonged and overlapped responses, which could be described by timescale patterns. Adding these patterns significantly increased the PPG-BP mutual information and improved the BP estimation performance. Compared to algorithms with morphological and biometric features, the mean absolute error (MAE) of calibrated systolic BP (SBP) reduced from 5.37mmHg to 4.51mmHg, while the MAE of calibration-free diastolic BP (DBP) reduced from 3.46mmHg to 2.81mmHg. The median intra-subject correlation between SBP/DBP estimation and ground truth increased from 0.63/0.34 to 0.80/0.68, which means the intrinsic BP fluctuation was better captured. Conclusion: Timescale patterns were vital to single-site PPG-based BP estimation. Understanding its physiological implication may help us design algorithms with clear interpretability and simplified structures.

show abstract

Photoplethysmography-Based Blood Pressure Estimation Combining Filter-Wrapper Collaborated Feature Selection With LASSO-LSTM Model

Cited by 38 publications

References 28 publications

Improving accuracy of vascular access quality classification in hemodialysis patients using deep learning with K highest score feature selection

Improving accuracy of vascular access quality classification in hemodialysis patients using deep learning with K highest score feature selection

Contribution of Timescale Patterns to Photoplethysmography-based Blood Pressure Estimation

Contribution of Timescale Patterns to Photoplethysmography-based Blood Pressure Estimation

Contact Info

Product

Resources

About