Atrial Fibrillation Monitoring Based on Noncontact Capacitive ECG Using an Integrated Microhumidity Fabric Electrode-Sheet Sensing Scheme

• Published in: IEEE transactions on instrumentation and measurement Vol. 72; pp. 1 - 11
• Main Authors: Xiao, Zhijun, Ma, Caiyun, Xing, Yantao, Yang, Chenxi, Hao, Menglong, Li, Jianqing, Liu, Chengyu
• Format: Journal Article
• Language: English
• Published: New York The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2023
• Subjects: Algorithms; Cardiac arrhythmia; Correlation coefficients; Electrocardiography; Electrodes; Fabric structures; Fibrillation; Monitoring; Signal quality; Skin; Support vector machines
• ISSN: 0018-9456; 1557-9662
• DOI: 10.1109/TIM.2023.3273648

This study designed a novel noncontact capacitive electrocardiogram (cECG) system for long-term atrial fibrillation (AF) monitoring. Unlike conventional ECG monitoring, the proposed system can record ECG signals without touching a subject’s skin, avoiding the problem of skin allergy by using wet electrodes. The main contributions of this study include the two aspects. First, a noncontact electrode based on a microhumidity fabric structure is proposed, which can improve the noncontact cECG signal quality. Second, based on the proposed electrode, the possibility of its application in long-term AF monitoring was investigated. The microhumidity fabric electrode can improve the microhumidity of the skin-electrode interface and the noncontact cECG signal quality. This humidity is low and can hardly be felt by the human body, which can also ensure comfort. The experimental results show that compared with the conventional noncontact fabric electrode, the skin-electrode impedance decreases from 104.1 ± 58.1 [Formula Omitted] (when using the electrode without the microhumidity structure) at 0.5 Hz to 2.8 ± 1.79 [Formula Omitted] (with the microhumidity structure). In the long-term cECG recording experiment, the system has high consistency with the reference equipment. The RR (R-wave to R-wave) interval correlation coefficient is 0.998, and the maximum error [Formula Omitted] ms. For the validation of AF monitoring, a designed AF analysis algorithm is integrated into the noncontact cECG system, including effective signal screening and AF recognition algorithm based on RR interval and support vector machine (SVM), with an average accuracy of 98.57% on the wearable database. The algorithm is verified on three healthy people and one paroxysmal AF (PAF) patient using the noncontact cECG system with an average accuracy of 98.12%, showing that our system has the potential for long-term AF monitoring.