Sengottuvel S, Devi S Shenbaga, Sasikala M, Satheesh Santhosh, Selvaraj Raja J
SQUIDs and Applications Section, Condensed Matter Physics Division, Materials Science Group, Indira Gandhi Centre for Atomic Research, Kalpakkam-603 102, India.
Department of Electronics and Communication Engineering, College of Engineering, Guindy, Anna University, Chennai 600 025, India.
Biomed Phys Eng Express. 2021 Mar 17;7(3). doi: 10.1088/2057-1976/abec17.
Magnetocardiograms (MCG) provide clinically useful diagnostic information in a variety of cardiac dysfunctions. Low frequency baseline drifts and high frequency noise are inevitably present in routine MCG even for those measured inside magnetically shielded rooms. These interferences sometimes exceed subtle cardiac features in MCG recorded on subjects with implanted devices like cardiac pacemakers; this makes interpretation of cardiac magnetic fields difficult. The present study proposes a correlation-based beat-by-beat approach and principal component analysis to eliminate drifts and high frequency noise respectively; the approach is suitable for denoising both single and multi-channel MCG data. The methodology is critically evaluated on simulated noisy measurements using a 37 channel MCG system, when objects such as implantable permanent pacemaker and stainless-steel wire are sequentially kept externally on the chests of five healthy subjects. By characterizing the noise introduced by each of these objects, the deterioration in the quality of MCG and its subsequent restoration by using the proposed method is assessed. The performance of the proposed method is also compared with other conventional denoising techniques namely, bandpass filters, wavelets and ensemble empirical mode decomposition. The proposed method not only exhibits least distortion, but also preserves the beat-by-beat dynamics of cardiac time series. The method has also been illustrated on actual MCG measurements on two subjects with implanted pacemaker which highlight the ability of the proposed method for denoising MCG in general and during extremely noisy measurement situations.
磁心动图（MCG）在多种心脏功能障碍中提供了具有临床实用价值的诊断信息。即使对于在磁屏蔽室内测量的常规MCG，低频基线漂移和高频噪声也不可避免地存在。在记录有植入式设备（如心脏起搏器）的受试者的MCG中，这些干扰有时会超过微弱的心脏特征；这使得对心脏磁场的解读变得困难。本研究提出了一种基于相关性的逐搏方法和主成分分析，分别用于消除漂移和高频噪声；该方法适用于对单通道和多通道MCG数据进行去噪。使用37通道MCG系统，在模拟噪声测量中，当将诸如植入式永久起搏器和不锈钢丝等物体依次放置在五名健康受试者的胸部外部时，对该方法进行了严格评估。通过表征这些物体各自引入的噪声，评估了MCG质量的恶化情况以及使用所提出的方法对其进行的后续恢复。还将所提出方法的性能与其他传统去噪技术，即带通滤波器、小波和总体经验模态分解进行了比较。所提出的方法不仅表现出最小的失真，而且还保留了心脏时间序列的逐搏动态。该方法还在两名植入起搏器的受试者的实际MCG测量中得到了验证，突出了所提出方法在一般情况下以及在极嘈杂测量情况下对MCG进行去噪的能力。
