Shulin Zhang

Shulin Zhang


Research Focus & Metrics

Academic Career Length

34Years (1993-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
23 October 2018 in IEEE Transactions on Biomedical Engineering
Verified
Journal Article
Closed access
RTRong Tao
Rong Tao
SZShulin Zhang
Shulin Zhang
XHXiao Huang
Xiao Huang
MTMinfang Tao
Minfang Tao
JMJian Ma
Jian Ma
XXXiaoming Xie
Xiaoming Xie
Objective: This study focused on developing a fast and accurate automatic ischemic heart disease detection/localization methodology. Methods: T wave was segmented from averaged Magnetocardiography (MCG) recordings and 164 features were subsequently extracted. These features were categorized into three groups: time domain features, frequency domain features, and information theory features. Next, we compared different machine learning classifiers including: k-nearest neighbor, decision tree, support vector machine (SVM), and XGBoost. To identify ischemia heart disease (IHD) case, we selected three classifiers with best performance and applied model ensemble to average results. All 164 features were used in this stage. To localize ischemia, we classified IHD group according to stenosis locations, including left anterior descending (LAD), left circumflex artery (LCX), and right coronary artery (RCA). For this task, we used XGBoost classifier and 18 time domain features. Results: For IHD detection, the SVM-XGBoost model achieved best results with accuracy = 94.03%, precision = 86.56%, recall = 97.78%, F-score = 92.79%, AUC = 0.98, and average precision = 0.98. For ischemia localization, XGBoost model achieved accuracy = 0.74, 0.68, and 0.65, for LAD, LCX, and RCA, respectively. Conclusion: we have developed an automatic IHD detection and localization system. We find that 1. T wave repolarization synchronicity is an important factor to distinguish IHD from normal subjects 2. Magnetic field pattern is associated with stenosis location. Significance: The proposed machine learning method provides the clinicians a fast and accurate diagnosis tool to interpret MCG data, boosting its acceptance into clinics. Furthermore, the magnetic pole characteristics revealed by the method shows to be related to ischemia location, presenting the opportunity to noninvasively locate ischemia.

15 November 2021 in IEEE Journal of Biomedical and Health Informatics
Verified
Journal Article
Closed access
RTRong Tao
Rong Tao
SZShulin Zhang
Shulin Zhang
YWYuexia Wang
Yuexia Wang
XMXianqiang Mi
Xianqiang Mi
JMJian Ma
Jian Ma
Guoyan ZhengGuoyan Zheng
Guoyan Zheng
In this paper, we propose an end-to-end deep learning architecture, referred as MCG-Net, integrating convolutional neural network (CNN) with transformer-based global context block for fine-grained delineation and diagnostic classification of four cardiac events from magnetocardiogram (MCG) data, namely Q-, R-, S- and T-waves. MCG-Net takes advantage of a multi-resolution CNN backbone as well as the state-of-the-art (SOTA) transformer encoders that facilitate global temporal feature aggregation. Besides the novel network architecture, we introduce a multi-task learning scheme to achieve simultaneous delineation and classification. Specifically, the problem of MCG delineation is formulated as multi-class heatmap regression. Meanwhile, a binary diagnostic classification label as well as a duration are jointly estimated for each cardiac event using features that are temporally aligned by event heatmaps. The framework is evaluated on a clinical MCG dataset, containing data collected from 270 subjects with cardiac anomalies and 108 control subjects. We designed and conducted a two-fold cross-validation study to validate the proposed method and to compare its performance with the SOTA methods. Experimental results demonstrated that our method outperformed counterparts on both event delineation and diagnostic classification tasks, achieving respectively an average ECG-F1 of 0.987 and an average Event-F1 of 0.975 for MCG delineation, and an average accuracy of 0.870, an average sensitivity of 0.732, an average specificity of 0.914 and an average AUC of 0.903 for diagnostic classification. Comprehensive ablation experiments are additionally performed to investigate effectiveness of different network components.

19 February 2025 in Scientific Reports
Verified
Journal Article
Open access
PDF file
RTRong Tao
Rong Tao
SZShulin Zhang
Shulin Zhang
RZRui Zhang
Rui Zhang
CSChengxing Shen
Chengxing Shen
JMJian Ma
Jian Ma
Guoyan ZhengGuoyan Zheng
Guoyan Zheng
Early diagnosis and localization of myocardial ischemia (MS) and coronary artery stenosis (CAS) play a crucial role in the effective prevention and management of ischemic heart disease (IHD). Magnetocardiography (MCG) has emerged as a promising approach for non-invasive, non-contact, and high-sensitivity assessment of cardiac dysfunction. This study presents a multi-center, AI-enabled diagnosis and localization of myocardial ischemia and coronary artery stenosis from MCG data. To this end, we collected a large-scale dataset consisting of 2,158 MCG recordings from eight clinical centers. We then proposed a multiscale vision transformer-based network for extracting spatio-temporal information from multichannel MCG recordings. Anatomical prior knowledge of the coronary artery and the irrigated left ventricular regions was incorporated by a carefully designed graph convolutional network (GCN)-based feature fusion module. The proposed approach achieved an accuracy of 84.7%, a sensitivity of 83.8%, and a specificity of 85.6% in diagnosing IHD, an average accuracy of 78.4% in localization of five MS regions, and an average accuracy of 65.3% in localization of stenosis in three coronary arteries. Subsequent validation on an independent validation dataset consisting of 268 MCG recordings collected from four clinical centers demonstrated an accuracy of 82.3%, a sensitivity of 83.8%, and a specificity of 81.3% in diagnosing IHD, an average accuracy of 77.3% in localization of five myocardial ischemic regions, and an average accuracy of 65.6% in localization of stenosis in three coronary arteries. The proposed approach can be used as a fast and accurate diagnosis tool, boosting the integration of MCG examination into clinical routine.

8 February 1993 in Applied Physics Letters
Verified
Journal Article
Closed access
PDF file
SZShulin Zhang
Shulin Zhang
KHKuok-San Ho
Kuok-San Ho
YHYongtian Hou
Yongtian Hou
BQBidong Qian
Bidong Qian
PDPeng Diao
Peng Diao
SCShengmin Cai
Shengmin Cai
The photoluminescence spectra from a series of porous silicon samples formed in HF concentrations at constant current density and fixed forming times, are presented. A novel steplike behavior of photoluminescence peak energy was observed. A formation mechanism based on the quantum confinement of porous Si walls has been proposed to explain the experimental results.

20 February 2020 in Chinese Physics B
Verified
Journal Article
Closed access
PDF file
SZShulin Zhang
Shulin Zhang
NCNing Cao
Ning Cao
Magnetocardiography (MCG) measurement is important for investigating the cardiac biological activities. Traditionally, the extremely weak MCG signal was detected by using superconducting quantum interference devices (SQUIDs). As a room-temperature magnetic-field sensor, optically pumped magnetometer (OPM) has shown to have comparable sensitivity to that of SQUIDs, which is very suitable for biomagnetic measurements. In this paper, a synthetic gradiometer was constructed by using two OPMs under spin-exchange relaxation-free (SERF) conditions within a moderate magnetically shielded room (MSR). The magnetic noise of the OPM was measured to less than 70 fT/Hz1/2. Under a baseline of 100 mm, noise cancellation of about 30 dB was achieved. MCG was successfully measured with a signal to noise ratio (SNR) of about 37 dB. The synthetic gradiometer technique was very effective to suppress the residual environmental fields, demonstrating the OPM gradiometer technique for highly cost-effective biomagnetic measurements.

1 April 2024 in Journal of Geriatric Cardiology
Verified
Journal Article
Closed access
PDF file
JCJian-Guo Cui
Jian-Guo Cui
FTFeng Tian
Feng Tian
YMYu-Hao Miao
Yu-Hao Miao
QJQin-Hua Jin
Qin-Hua Jin
YSYa-Jun Shi
Ya-Jun Shi
YCYundai Chen
Yundai Chen
OBJECTIVE To evaluate the role of resting magnetocardiography in identifying severe coronary artery stenosis in patients with suspected coronary artery disease. METHODS A total of 513 patients with angina symptoms were included and divided into two groups based on the extent of coronary artery disease determined by angiography: the non-severe coronary stenosis group (< 70% stenosis) and the severe coronary stenosis group (≥ 70% stenosis). The diagnostic model was constructed using magnetic field map (MFM) parameters, either individually or in combination with clinical indicators. The performance of the models was evaluated using receiver operating characteristic curves, accuracy, sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV). Calibration plots and decision curve analysis were performed to investigate the clinical utility and performance of the models, respectively. RESULTS In the severe coronary stenosis group, QR_MCTDd, S_MDp, and TT_MAC50 were significantly higher than those in the non-severe coronary stenosis group (10.46 ± 10.66 vs. 5.11 ± 6.07, P < 0.001; 7.2 ± 8.64 vs. 4.68 ± 6.95, P = 0.003; 0.32 ± 57.29 vs. 0.26 ± 57.29, P < 0.001). While, QR_MVamp, R_MA, and T_MA in the severe coronary stenosis group were lower (0.23 ± 0.16 vs. 0.28 ± 0.16, P < 0.001; 55.06 ± 48.68 vs. 59.24 ± 53.01, P < 0.001; 51.67 ± 39.32 vs. 60.45 ± 51.33, P < 0.001). Seven MFM parameters were integrated into the model, resulting in an area under the curve of 0.810 (95% CI: 0.765–0.855). The sensitivity, specificity, PPV, NPV, and accuracy were 71.7%, 80.4%, 93.3%, 42.8%, and 73.5%; respectively. The combined model exhibited an area under the curve of 0.845 (95% CI: 0.798–0.892). The sensitivity, specificity, PPV, NPV, and accuracy were 84.3%, 73.8%, 92.6%, 54.6%, and 82.1%; respectively. Calibration curves demonstrated excellent agreement between the nomogram prediction and actual observation. The decision curve analysis showed that the combined model provided greater net benefit compared to the magnetocardiography model. CONCLUSIONS The novel quantitative MFM parameters, whether used individually or in combination with clinical indicators, have been shown to effectively predict the risk of severe coronary stenosis in patients presenting with angina-like symptoms. Magnetocardiography, an emerging non-invasive diagnostic tool, warrants further exploration for its potential in diagnosing coronary heart disease.