A Multiplayer Perceptron Algorithm Based on Global Electrical Heterogeneity and Cardiodynamicsgram for Non-Invasive Detection of Coronary Microvascular Dysfunction

Xiaoye Zhao; Yinglan Gong; Jucheng Zhang; Haipeng Liu; Tianhai Huang; Jun Jiang; Ling Xia; Jiandong Mao

doi:10.1109/eei63073.2024.10696636

A Multiplayer Perceptron Algorithm Based on Global Electrical Heterogeneity and Cardiodynamicsgram for Non-Invasive Detection of Coronary Microvascular Dysfunction

Xiaoye Zhao, Yinglan Gong, Jucheng Zhang, Haipeng Liu, Tianhai Huang, Jun Jiang, Ling Xia, Jiandong Mao

Centre for Intelligent Healthcare

Research output: Chapter in Book/Report/Conference proceeding › Chapter › peer-review

8 Downloads (Pure)

Abstract

Coronary microvascular dysfunction (CMD) is a key etiology of myocardial ischemia. Currently, early and non- invasive detection of CMD is lack in clinical practice. We proposed a multiplayer perceptron (MLP) model based on global electrical heterogeneity (GEH)- and cardiodynamicsgram (CDG)-based features for non-invasive detection of CMD. 25 GEH parameters and CDG-based features were derived from vectocardiograms (VCGs) of 82 CMD patients and 107 healthy controls to train the support vector machine (SVM) and multiplayer perceptron (MLP) models. Seventeen features (14 GEH parameters, 3 CDG-based features) were finally selected, and the performance of the two models were compared. The proposed MLP model performs better than SVM and provides accuracy: 0.871, sensitivity: 0.925, specificity: 0.775, F1-score: 0.810, and AUC (area under curve): 0.849 in the testing dataset. The proposed MLP model could distinguish CMD from healthy controls, providing possibilities for a noninvasive, low-cost approach for early detection of CMD.

Original language	English
Title of host publication	2024 6th International Conference on Electronic Engineering and Informatics (EEI)
Publisher	IEEE
Pages	504-509
Number of pages	6
ISBN (Electronic)	979-8-3503-5359-4, 979-8-3503-5358-7
ISBN (Print)	979-8-3503-5360-0
DOIs	https://doi.org/10.1109/eei63073.2024.10696636
Publication status	E-pub ahead of print - 8 Oct 2024

Bibliographical note

This document is the author’s post-print version, incorporating any revisions agreed during the peer-review process. Some differences between the published version and this version may remain and you are advised to consult the published version if you wish to cite from it.

Keywords

Cardiodynamicsgram (CDG)
Global electrical heterogeneity (GEH)
Electrocardiography (ECG)
Multiplayer perceptron (MLP)
Coronary microvascular dysfunction (CMD)

Access to Document

10.1109/eei63073.2024.10696636

LIU2024VORAccepted author manuscript, 1.06 MB

Cite this

Zhao, X., Gong, Y., Zhang, J., Liu, H., Huang, T., Jiang, J., Xia, L., & Mao, J. (2024). A Multiplayer Perceptron Algorithm Based on Global Electrical Heterogeneity and Cardiodynamicsgram for Non-Invasive Detection of Coronary Microvascular Dysfunction. In 2024 6th International Conference on Electronic Engineering and Informatics (EEI) (pp. 504-509). IEEE. Advance online publication. https://doi.org/10.1109/eei63073.2024.10696636

A Multiplayer Perceptron Algorithm Based on Global Electrical Heterogeneity and Cardiodynamicsgram for Non-Invasive Detection of Coronary Microvascular Dysfunction. / Zhao, Xiaoye; Gong, Yinglan; Zhang, Jucheng et al.
2024 6th International Conference on Electronic Engineering and Informatics (EEI). IEEE, 2024. p. 504-509.

Research output: Chapter in Book/Report/Conference proceeding › Chapter › peer-review

Zhao, X, Gong, Y, Zhang, J, Liu, H, Huang, T, Jiang, J, Xia, L & Mao, J 2024, A Multiplayer Perceptron Algorithm Based on Global Electrical Heterogeneity and Cardiodynamicsgram for Non-Invasive Detection of Coronary Microvascular Dysfunction. in 2024 6th International Conference on Electronic Engineering and Informatics (EEI). IEEE, pp. 504-509. https://doi.org/10.1109/eei63073.2024.10696636

Zhao X, Gong Y, Zhang J, Liu H, Huang T, Jiang J et al. A Multiplayer Perceptron Algorithm Based on Global Electrical Heterogeneity and Cardiodynamicsgram for Non-Invasive Detection of Coronary Microvascular Dysfunction. In 2024 6th International Conference on Electronic Engineering and Informatics (EEI). IEEE. 2024. p. 504-509 Epub 2024 Oct 8. doi: 10.1109/eei63073.2024.10696636

@inbook{c2ad37f58fbb41a2a86a31cf0bb7edee,

title = "A Multiplayer Perceptron Algorithm Based on Global Electrical Heterogeneity and Cardiodynamicsgram for Non-Invasive Detection of Coronary Microvascular Dysfunction",

abstract = "Coronary microvascular dysfunction (CMD) is a key etiology of myocardial ischemia. Currently, early and non- invasive detection of CMD is lack in clinical practice. We proposed a multiplayer perceptron (MLP) model based on global electrical heterogeneity (GEH)- and cardiodynamicsgram (CDG)-based features for non-invasive detection of CMD. 25 GEH parameters and CDG-based features were derived from vectocardiograms (VCGs) of 82 CMD patients and 107 healthy controls to train the support vector machine (SVM) and multiplayer perceptron (MLP) models. Seventeen features (14 GEH parameters, 3 CDG-based features) were finally selected, and the performance of the two models were compared. The proposed MLP model performs better than SVM and provides accuracy: 0.871, sensitivity: 0.925, specificity: 0.775, F1-score: 0.810, and AUC (area under curve): 0.849 in the testing dataset. The proposed MLP model could distinguish CMD from healthy controls, providing possibilities for a noninvasive, low-cost approach for early detection of CMD.",

keywords = "Cardiodynamicsgram (CDG), Global electrical heterogeneity (GEH), Electrocardiography (ECG), Multiplayer perceptron (MLP), Coronary microvascular dysfunction (CMD)",

author = "Xiaoye Zhao and Yinglan Gong and Jucheng Zhang and Haipeng Liu and Tianhai Huang and Jun Jiang and Ling Xia and Jiandong Mao",

note = "This document is the author{\textquoteright}s post-print version, incorporating any revisions agreed during the peer-review process. Some differences between the published version and this version may remain and you are advised to consult the published version if you wish to cite from it. ",

year = "2024",

month = oct,

day = "8",

doi = "10.1109/eei63073.2024.10696636",

language = "English",

isbn = "979-8-3503-5360-0",

pages = "504--509",

booktitle = "2024 6th International Conference on Electronic Engineering and Informatics (EEI)",

publisher = "IEEE",

address = "United States",

}

TY - CHAP

T1 - A Multiplayer Perceptron Algorithm Based on Global Electrical Heterogeneity and Cardiodynamicsgram for Non-Invasive Detection of Coronary Microvascular Dysfunction

AU - Zhao, Xiaoye

AU - Gong, Yinglan

AU - Zhang, Jucheng

AU - Liu, Haipeng

AU - Huang, Tianhai

AU - Jiang, Jun

AU - Xia, Ling

AU - Mao, Jiandong

N1 - This document is the author’s post-print version, incorporating any revisions agreed during the peer-review process. Some differences between the published version and this version may remain and you are advised to consult the published version if you wish to cite from it.

PY - 2024/10/8

Y1 - 2024/10/8

N2 - Coronary microvascular dysfunction (CMD) is a key etiology of myocardial ischemia. Currently, early and non- invasive detection of CMD is lack in clinical practice. We proposed a multiplayer perceptron (MLP) model based on global electrical heterogeneity (GEH)- and cardiodynamicsgram (CDG)-based features for non-invasive detection of CMD. 25 GEH parameters and CDG-based features were derived from vectocardiograms (VCGs) of 82 CMD patients and 107 healthy controls to train the support vector machine (SVM) and multiplayer perceptron (MLP) models. Seventeen features (14 GEH parameters, 3 CDG-based features) were finally selected, and the performance of the two models were compared. The proposed MLP model performs better than SVM and provides accuracy: 0.871, sensitivity: 0.925, specificity: 0.775, F1-score: 0.810, and AUC (area under curve): 0.849 in the testing dataset. The proposed MLP model could distinguish CMD from healthy controls, providing possibilities for a noninvasive, low-cost approach for early detection of CMD.

AB - Coronary microvascular dysfunction (CMD) is a key etiology of myocardial ischemia. Currently, early and non- invasive detection of CMD is lack in clinical practice. We proposed a multiplayer perceptron (MLP) model based on global electrical heterogeneity (GEH)- and cardiodynamicsgram (CDG)-based features for non-invasive detection of CMD. 25 GEH parameters and CDG-based features were derived from vectocardiograms (VCGs) of 82 CMD patients and 107 healthy controls to train the support vector machine (SVM) and multiplayer perceptron (MLP) models. Seventeen features (14 GEH parameters, 3 CDG-based features) were finally selected, and the performance of the two models were compared. The proposed MLP model performs better than SVM and provides accuracy: 0.871, sensitivity: 0.925, specificity: 0.775, F1-score: 0.810, and AUC (area under curve): 0.849 in the testing dataset. The proposed MLP model could distinguish CMD from healthy controls, providing possibilities for a noninvasive, low-cost approach for early detection of CMD.

KW - Cardiodynamicsgram (CDG)

KW - Global electrical heterogeneity (GEH)

KW - Electrocardiography (ECG)

KW - Multiplayer perceptron (MLP)

KW - Coronary microvascular dysfunction (CMD)

UR - http://www.scopus.com/inward/record.url?scp=85207525205&partnerID=8YFLogxK

U2 - 10.1109/eei63073.2024.10696636

DO - 10.1109/eei63073.2024.10696636

M3 - Chapter

SN - 979-8-3503-5360-0

SP - 504

EP - 509

BT - 2024 6th International Conference on Electronic Engineering and Informatics (EEI)

PB - IEEE

ER -

A Multiplayer Perceptron Algorithm Based on Global Electrical Heterogeneity and Cardiodynamicsgram for Non-Invasive Detection of Coronary Microvascular Dysfunction

Abstract

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this