Chun-Li Wang

Recognition of Ventricular Extrasystoles Over the Reconstructed Phase Space of Electrocardiogram

Distinguishing ventricular extrasystoles from normal heartbeats is crucial to cardiac arrhythmia analysis. This paper proposes novel morphological descriptors, the major portrait partition area (MPPA) and point distribution percentage (PDP), which are extracted from the reconstructed phase space of the QRS complex. These measures can be linked to QRS width and prolonged ventricular contraction, and offer several advantages over traditional characterization of the QRS structure: it does not require QRS boundary detection, is robust under R-peak misalignment, and including some information from nearby points. The first four principal components of MPPA variables and PDPs in the first and the third quadrants of the phase space diagram were used as inputs of neural networks. The performance of networks in distinguishing premature ventricular contraction events from normal heartbeats were evaluated under a series of 50 cross-validations based on the electrocardiogram data taken from the MIT/BIH arrhythmia database. The sensitivity and specificity obtained using the aforementioned MPPA principal components and PDPs as inputs were similar to those obtained using wavelet features and Hermite coefficients. However, the phase space information performed better in situations of noise contaminations and waveform deformations.

An intelligent classifier for prognosis of cardiac resynchronization therapy based on speckle-tracking echocardiograms

PURPOSE Predicting response after cardiac resynchronization therapy (CRT) has been a challenge of cardiologists. About 30% of selected patients based on the standard selection criteria for CRT do not show response after receiving the treatment. This study is aimed to build an intelligent classifier to assist in identifying potential CRT responders by speckle-tracking radial strain based on echocardiograms. METHODS AND MATERIALS The echocardiograms analyzed were acquired before CRT from 26 patients who have received CRT. Sequential forward selection was performed on the parameters obtained by peak-strain timing and phase space reconstruction on speckle-tracking radial strain to find an optimal set of features for creating intelligent classifiers. Support vector machine (SVM) with a linear, quadratic, and polynominal kernel were tested to build classifiers to identify potential responders and non-responders for CRT by selected features. RESULTS Based on random sub-sampling validation, the best classification performance is correct rate about 95% with 96-97% sensitivity and 93-94% specificity achieved by applying SVM with a quadratic kernel on a set of 3 parameters. The selected 3 parameters contain both indexes extracted by peak-strain timing and phase space reconstruction. CONCLUSIONS An intelligent classifier with an averaged correct rate, sensitivity and specificity above 90% for assisting in identifying CRT responders is built by speckle-tracking radial strain. The classifier can be applied to provide objective suggestion for patient selection of CRT.

Phase-space Reconstruction of Electrocardiogram for Heartbeat Classification

Heartbeat classification is crucial for cardiac arrhythmia analysis. QRS complex presents important characteristics which are beneficial to distinguish abnormal beats from normal beats. In the present study we propose a novel descriptor for QRS complex. The waveform is transformed to a two-dimensional phase space and then mapped to a onedimensional portrait partition area (PPA). The proposed morphological descriptor has advantages of no need to detect Q and S characteristic points, tolerating R-peak misalignment and taking into account temporal relation of data samples. On the basis of 32 records from the MIT/BIH arrhythmia database, normal QRS and premature ventricular contraction (PVC) beats show different phase space portraits and PPA. An artificial neuronal network using PPA as the input feature was built for heartbeat classification. Our results showed that the sensitivity and specificity of distinguishing PVC from normal QRS achieved 0.9699 and 0.9651 in the testing sets, respectively.

Instantaneous Respiratory Estimation from Thoracic Impedance by Empirical Mode Decomposition

Impedance plethysmography provides a way to measure respiratory activity by sensing the change of thoracic impedance caused by inspiration and expiration. This measurement imposes little pressure on the body and uses the human body as the sensor, thereby reducing the need for adjustments as body position changes and making it suitable for long-term or ambulatory monitoring. The empirical mode decomposition (EMD) can decompose a signal into several intrinsic mode functions (IMFs) that disclose nonstationary components as well as stationary components and, similarly, capture respiratory episodes from thoracic impedance. However, upper-body movements usually produce motion artifacts that are not easily removed by digital filtering. Moreover, large motion artifacts disable the EMD to decompose respiratory components. In this paper, motion artifacts are detected and replaced by the data mirrored from the prior and the posterior before EMD processing. A novel intrinsic respiratory reconstruction index that considers both global and local properties of IMFs is proposed to define respiration-related IMFs for respiration reconstruction and instantaneous respiratory estimation. Based on the experiments performing a series of static and dynamic physical activates, our results showed the proposed method had higher cross correlations between respiratory frequencies estimated from thoracic impedance and those from oronasal airflow based on small window size compared to the Fourier transform-based method.

ECG-derived respirations based on phase-space reconstruction of single-lead ECG: validations over various physical activities based on parallel recordings of ECG, respiration, and body accelerations.

ECG derived respiration (EDR) provides a comfortable measurement of respiration and is recently applied to sleep studies. Nevertheless, the mechanic disturbances due to postural changes or other physical activity during long-term recording is less investigated. In the present study, ECG, impedance-based respiration, and body accelerations were parallel recorded during a series of scheduled postures in 8 healthy subjects and during 24 hours in one subject. In addition, a novel EDR method based on phase space reconstruction of single-lead ECG is also proposed. The area under major portrait radius (MPR) curve is employed to quantify the deviation of phase-space loop which is related to respiration. Coherence analysis between the EDR and the impedance-based respiration demonstrated that the MPR-based EDR had better performance than using R peak or QRS area as the EDR feature in the scheduled postures.ECG derived respiration (EDR) provides a comfortable measurement of respiration and is recently applied to sleep studies. Nevertheless, the mechanic disturbances due to postural changes or other physical activity during long-term recording is less investigated. In the present study, ECG, impedance-based respiration, and body accelerations were parallel recorded during a series of scheduled postures in 8 healthy subjects and during 24 hours in one subject. In addition, a novel EDR method based on phase space reconstruction of single-lead ECG is also proposed. The area under major portrait radius (MPR) curve is employed to quantify the deviation of phase-space loop which is related to respiration. Coherence analysis between the EDR and the impedance-based respiration demonstrated that the MPR-based EDR had better performance than using R peak or QRS area as the EDR feature in the scheduled postures.

PATTERNS OF LEFT VENTRICULAR CONTRACTION IN STRAIN VECTOR SPACE RELATED TO BUNDLE BRANCH BLOCK WITH HEART FAILURE BY SPECKLE-TRACKING ECHOCARDIOGRAPHY

The aim of this research is to study bundle branch block (BBB)-related patterns of radial strain in the left ventricles of patients with heart failure by speckle-tracking echocardiography. Twenty-seven left-BBB (LBBB), 10 right-BBB (RBBB), and 11 narrow QRS-complexes (non-BBB) patients and 11 healthy subjects were assessed. Strain fractions used to quantify thickening-during-systole and thinning-during-diastole, and timing parameters defined as time to onset-of-thickening and peak-strain were measured. Principal strain vectors were conducted on the fractions and parameters to analyze mechanical discoordination and dyssynchrony. Heart failure patients show a significantly greater extent of discoordination and dyssynchrony compared with healthy subjects. Significant differences between the LBBB and RBBB groups are demonstrated by deflection, a spatial characteristic of myocardial coordination. New information provided by these findings can provide a better understanding of BBB-related mechanisms of myocardial coordination and may be useful in improving patient selection, electrode placement and subsequent outcomes for cardiac resynchronization therapy.

Phase space analysis of myocardial coordination related to left ventricular ejection fraction by echocardiographic speckle-tracking radial strain.

Phase space reconstruction, which is performed by converting echocardiogram-derived strain data from different ventricular regions into phase space trajectories, is applied in this study to describe nonlinear behaviour of myocardial coordination. A new method was developed to quantify patterns of phase space trajectories. Echocardiograms of 31 healthy individuals and 63 patients with left bundle branch block (LBBB) and different left ventricular ejection fractions (LVEFs) were used to evaluate this method. The LBBB patients were separated into two groups: LBBB with a LVEF≥50% and LBBB with a LVEF<50%. LVEF is used to represent levels of systolic function and disease severity. A classifying map was constructed to separate the reconstructed phase space into three zones and to acquire the parameters Pz1, Pz2 and Pz3 as percentiles of phase points related to the zones. The criteria used to define the zones were cross-tested. Based on these parameters, significant group-related differences in myocardial coordination were observed. Significantly smaller Pz1 and significantly larger Pz2 values were observed in the healthy group, as compared to the patient group, and similar, significant results were obtained for the patients with LVEF≥50%, as compared to the patients with LVEF<50% (p<.05). A significantly larger Pz3 was observed in patients with LVEF<50%, as compared to the other patients (p<.05). In addition, different inter-regional relationships among strain pairs (all, near-side, middle-side and opposite-wall) were examined to create phase space trajectories. Consistent group-related differences were observed when different inter-regional relationships were applied. Expanding the use of the proposed method to assess various pathological factors and therapeutic impacts is promising.

Detecting Endocardial Boundary in Echocardiogram by Anisotropic Filtering and Entropy-Weighted Features

Reading echocardiograms is important to evaluate cardiac function. Due to influence of speckle, shadow and artifacts, analyzing echocardiograms requires more effort and energy than other medical imaging. A computerized method is proposed by this study to automatize the detection of endocardial boundaries based on B-mode echocardiograms in shortaxis view. Local entropy, anisotropic filtering, and cost image technique are used to pre-process the images to enhance the difference of blood region from the segments of myocardium and fix missing edge components. Above 80% of true positive can be achieved by the method when comparing to boundaries identified manually.

An Intelligent Classifier for Heart Health – Based on Radial Myocardial Strain and Electrocardiogram

Heart health is important for anyone everywhere. The electrocardiograms (ECG) and echocardiograms are common-used and critical clinical assessment of the heart. A classifier based on ECG and echocardiograms for identifying heart health and offering objective suggestion was proposed. Clinical data and cardiologists’ diagnosis were used to evaluate the classifier. The accuracy rate to identify heart condition between healthy and ill was about 90% and over. Among the heart disease trials, which of them indicates to cardiac resynchronization therapy (CRT) can be distinguished with an accuracy rate of 75%. This work may facilitate clinical data reading and serve as a heart health consulter when needed. More factors will be investigated in future studies to keep upgrading the classifier.