ECG Signal Research Articles

Design of a low-area hardware architecture to predict early signs of sudden cardiac arrests

Sudden cardiac arrest (SCA) results in an unexpected and untimely death within minutes, and its early prediction can alert cardiac patients to a timely medical diagnosis. To detect early symptoms of an SCA, the detection and classification of ventricular tachycardias (VT) are of utmost importance. In this work, a low-area yet highly accurate hardware architecture for VT classification is proposed based on the detection of premature ventricular contraction (PVC) beats. After pre-processing of the ECG signals using a wavelet-based pre-processing unit, a characteristics-matching algorithm is used to detect the PVC beats, and a low-complexity adaptive decision-based logic classifier is used to classify them into four types of VTs, namely monomorphic, polymorphic, non-sustained VT (NSVT), and sustained VT (SVT). FPGA verification of the hardware architecture for the VT classifier using the Nexys 4 DDR Artix-7 board utilizes 10.4 % of the total available resources and displays the type of VT and the number of PVCs detected to help in determining the severity of SCA and the need for medical attention. The ASIC implementation of the proposed PVC-based VT classification using the SCL 180 nm CMOS technology results in an area overhead of 0.02 mm2 and a power consumption of 3.47 μW for a high accuracy rate of 98.2 %. When compared to the existing CA detection systems for wearable devices, the proposed one consumes the least area while achieving high detection rates.

Visualized Lead Selection for Arrhythmia Classification Based on a Lead Activation Heatmap Using Multi-Lead ECGs.

Visualizing the decision-making process is a key aspect of research regarding explainable arrhythmia recognition. This study proposed a visualized lead selection method to classify arrhythmia for multi-lead ECG signals. The proposed method has several advantages, as it uses a visualized approach to select effective leads, avoiding redundant leads and invalid information. It also captures the temporal dependencies of ECG signals and the complementary information between leads. The method deployed a lead activation heatmap (LA heatmap) based on a lead-wise network to select the proper 5 leads from 12-lead ECG heartbeats extracted from the public 2018 Chinese Physiological Signal Challenge database (CPSC 2018 DB), which were then fed into a ResBiTime network combining bidirectional long short-term memory (Bi-LSTM) networks and residual connections for a classification task of nine heartbeat categories (i.e., N, AF, I-AVB, RBBB, PAC, PVC, STD, LBBB, and STE). The results indicate an average precision of 93.25%, an average recall of 93.03%, an average F1-score of 0.9313, and that the proposed method can effectively extract additional information from ECG heartbeat data.

Improved Bivariate-VAR Model for Extraction of Respiratory Information from Artifact Corrupted ECG and PPG Signals

Improved Bivariate-VAR Model for Extraction of Respiratory Information from Artifact Corrupted ECG and PPG Signals

Priority-based Multi-feature Vector Model Using Convolution Neural Network for Biometric Authentication

Many societal institutions have increased standards for the efficacy and dependability of identification systems in response to the ever-increasing sophistication of computer technology. Biometric identity systems have mostly supplanted the usage of conventional key and smart card systems in highly secure industries. There has been a dramatic increase in the adoption of biometric authentication systems in the last decade. While many biometric systems have relied on exterior physiological traits like fingerprints, iris scans, palm prints, and faces, very few have investigated the possibility of using interior physiological traits as a biometric. It is challenging for the benefits of unimodal biometric technology to be realized in real-world applications because to its inherent constraints, such as the fact that it only proposes a single piece of information and that environmental factors can impact data verification. While passwords have served their purpose of authenticating users and controlling access, they have also revealed their weaknesses. Concerns about forging or impersonating authentication have prompted the proposal of various Artificial Intelligence (AI)-based solutions. A lot of people are interested in user authentication based on ECGs these days. Unlike other biometrics, an electrocardiogram verifies that the person is actual and alive, making it one of the most trustworthy advanced authentication methods. The P, Q, R, S, and T characteristic points of an ECG signal are responsible for its most salient features. Feature selection is a statistical method that can be implemented in a supervised or unsupervised setting via regression or classification. A Convolution Neural Network (CNN) is a type of neural network that extracts and feeds into another neural network, which then classifies those extracted features. The ECG signal serves as input to a feature extraction network. The neural network performs categorization based on the retrieved feature signals. To overcome the problem that the recognition accuracy of traditional ECG identification methods declines with the rise in the number of testing samples at various moments or throughout different heartbeat cycles, this research offers a multi feature reuse model. In this research, an Associated Priority-based Weighted Multi-Feature Vector model using Convolution Neural Network (APbWMFV-CNN) is proposed for ECG signal-based authentication. The proposed model selects the most appropriate features used for accurate biometric model. The comparison results represent that the proposed model performance in user detection is high.

Detection of Sleep Apnea From ECG Signals Using Sliding Singular Spectrum Based Subpattern Principal Component Analysis

Detection of Sleep Apnea From ECG Signals Using Sliding Singular Spectrum Based Subpattern Principal Component Analysis

Detection of Atrial Fibrillation from ECG Signal Using Efficient Feature Selection and Classification

Detection of Atrial Fibrillation from ECG Signal Using Efficient Feature Selection and Classification

1D Convolutional Neural Network Impact on Heart Rate Metrics for ECG and BCG Signals

1D Convolutional Neural Network Impact on Heart Rate Metrics for ECG and BCG Signals

Potato peels-based electrodes for recording ECG and EMG signals

Reliable measurement and acquisition of physiological parameters is among critical tasks in human health monitoring. While the presence of personalized measurement of Electrocardiography (ECG) and Electromyography (EMG) devices enables more frequent data collection, and consequently better health monitoring, it also leads to higher consumption of electrodes. On the other hand, commonly used Ag/AgCl electrodes are not free of charge or biodegradable. Therefore, their extensive use leads to higher costs and higher waste production. In this paper, we present an approach for the development of ECG/EMG electrodes that are based on biodegradable materials (two potato types). We performed successful integration of potato peel-based electrodes with the commercial device as well as with our in-house developed portable system. Measurements on 8 healthy volunteers revealed very small differences when heart rate, inter-beat-interval or normal-to-normal intervals are extracted from the ECG signal with potato peel-based and commercial Ag/AgCl electrodes. Our study demonstrates the feasibility of utilizing biodegradable potato peel-based electrodes for ECG and EMG measurements, showcasing their comparable performance to traditional Ag/AgCl electrodes in capturing physiological parameters. We introduced biodegradable electrodes as a viable and eco-friendly alternative for accurate ECG and EMG measurements, addressing the issues of electrode waste and cost associated with conventional electrodes.

Highly stretchable six-leaf hand structure electrodes prepared by sacrificial layer method, resistant to sweat erosion, for EMG and ECG signal detection

Recently, wearable flexible electronic devices have become more widely used in detecting human physiological signals, so higher demands have been placed on flexible stretchable electrodes directly attached to the human body, which need to have better mechanical tensile properties, anti-fatigue properties, and chemical stability. In this paper, the four-segment hand structure is optimized to obtain the four-leaf hand structure, which reduces the maximum stress value when the electrode is stretched and decreases the probability of electrode fracture. Further, the structure of the electrode is optimized by changing the number of nodes, the six-leaf hand structure electrode is obtained. The six-leaf hand structure electrode has lower tensile stress and stability. The substrate-free six-leaf hand structure electrode prepared by the sacrificial layer method has a much higher stretchability than the electrode with a substrate layer due to the bonding of the substrate-free layer, and there is no significant change in the resistance value during the stretching process. The screen-printed waterproof-coated six-leaf hand structure electrode has good sweat resistance and can be used for ECG and EMG detection.

Integrating Morphological Characteristics with Empirical Mode Decomposition for Robust ECG Signal Classification

Integrating Morphological Characteristics with Empirical Mode Decomposition for Robust ECG Signal Classification

A porcine large animal model of radiofrequency ablation-induced left bundle branch block

Abstract Funding Acknowledgements Type of funding sources: Private company. Main funding source(s): Medtronic Background and Objective ECG features of left bundle branch (LBB) block (LBBB) can be observed in up to 20–30 % of patients that suffer from heart failure with reduced ejection fraction. The resulting mechanical intraventricular asynchrony contributes to the worsening of left ventricular hemodynamics and has a relevant impact on the prognosis of the affected patients. The fact that we still cannot accurately predict which LBBB patients will benefit from cardiac resynchronization therapy (CRT) underscores that our understanding of the pathophysiology of LBBB remains incomplete. Here, a translational animal model of LBBB could be a major contributor to improving our knowledge of the molecular and translational pathomechanisms underlying LBBB. Methods A total of 15 male pigs underwent RF catheter ablation of the LBB, guided by fluoroscopy and intracardiac ECG signals. Before LBBB induction and after a follow-up period of one week or three weeks, a thorough clinical characterization was performed including 12-lead ECG, echocardiography, and electroanatomical mapping. To further test the feasibility of cardiac resynchronization in our model, 3 pigs were implanted with a CRT pacemaker on day seven after LBB ablation. Results Upon LBB ablation, 12-lead ECGs displayed typical characteristics of LBBB like a QRS broadening, a QRS slurring in left lateral leads, and a change of the QRS axis. The QRS width was prolonged from 64.2±4.2 ms to 93,4±7.4 ms and the R wave peak time in V6 was extended from 19.3±3.9 ms to 42.7±4.2 ms. Cardiac electromechanical asynchrony was further confirmed by echocardiography, where the appearance of a septal flash, the prolongation of the septal to posterior wall movement delay and extended left ventricular as well as interventricular electromechanical delays were documented. Electroanatomical mapping demonstrated a shift of the left ventricular breakthrough site. After LBBB-induction, left ventricular activation times, defined as the average of the maximum 10 % of activation times minus the average of the minimum 10 % of activation times were significantly prolonged from 36.3±7.1 ms to 60.8±10.2 ms while right ventricular activation times remained unchanged. Finally, RF-induced LBBB persisted over an observation period of three weeks. The pigs that underwent CRT showed a reduction in QRS duration from 99.1±8.2 ms on day seven after LBBB induction to 75.5±8.2 ms and echocardiography provided additional evidence of successful resynchronization. Conclusions In summary, our porcine large animal model accurately recapitulates the electrical and electromechanical phenotype, observed in LBBB patients. This closed-chest model of LBBB therefore provides an experimental platform to study both, molecular and translational characteristics of cardiac electromechanical asynchrony in a reproducible and clinically relevant setting.

Modified MF-DFA Model Based on LSSVM Fitting

This paper proposes a multifractal least squares support vector machine detrended fluctuation analysis (MF-LSSVM-DFA) model. The system is an extension of the traditional MF-DFA model. To address potential overfitting or underfitting caused by the fixed-order polynomial fitting in MF-DFA, LSSVM is employed as a superior alternative for fitting. This approach enhances model accuracy and adaptability, ensuring more reliable analysis results. We utilize the p model to construct a multiplicative cascade time series to evaluate the performance of MF-LSSVM-DFA, MF-DFA, and two other models that improve upon MF-DFA from recent studies. The results demonstrate that our proposed modified model yields generalized Hurst exponents h(q) and scaling exponents τ(q) that align more closely with the analytical solutions, indicating superior correction effectiveness. In addition, we explore the sensitivity of MF-LSSVM-DFA to the overlapping window size s. We find that the sensitivity of our proposed model is less than that of MF-DFA. We find that when s exceeds the limited range of the traditional MF-DFA, h(q) and τ(q) are closer than those obtained in MF-DFA when s is in a limited range. Meanwhile, we analyze the performances of the fitting of the two models and the results imply that MF-LSSVM-DFA achieves a better outstanding performance. In addition, we put the proposed MF-LSSVM-DFA into practice for applications in the medical field, and we found that MF-LSSVM-DFA improves the accuracy of ECG signal classification and the stability and robustness of the algorithm compared with MF-DFA. Finally, numerous image segmentation experiments are adopted to verify the effectiveness and robustness of our proposed method.

A neurophysiological profiling of the heartbeat-evoked potential in severe acquired brain injuries: A focus on unconsciousness.

Unconsciousness in severe acquired brain injury (sABI) patients occurs with different cognitive and neural profiles. Perturbational approaches, which enable the estimation of proxies for brain reorganization, have added a new avenue for investigating the non-behavioural diagnosis of consciousness. In this prospective observational study, we conducted a comparative analysis of the topological patterns of heartbeat-evoked potentials (HEP) between patients experiencing a prolonged disorder of consciousness (pDoC) and patients emerging from a minimally consciousness state (eMCS). A total of 219 sABI patients were enrolled, each undergoing a synchronous EEG-ECG resting-state recording, together with a standardized consciousness diagnosis. A number of graph metrics were computed before/after the HEP (Before/After) using the R-peak on the ECG signal. The peak value of the global field power of the HEP was found to be significantly higher in eMCS patients with no difference in latency. Power spectrum was not able to discriminate consciousness neither Before nor After. Node assortativity and global efficiency were found to vary with different trends at unconsciousness. Lastly, the Perturbational Complexity Index of the HEP was found to be significantly higher in eMCS patients compared with pDoC. Given that cortical elaboration of peripheral inputs may serve as a non-behavioural determinant of consciousness, we have devised a low-cost and translatable technique capable of estimating causal proxies of brain functionality with an endogenous, non-invasive stimulus. Thus, we present an effective means to enhance consciousness assessment by incorporating the interaction between the autonomic nervous system (ANS) and central nervous system (CNS) into the loop.

From ECG to broadband ECG, focused on the ventricular activation

Abstract Background An ultra-high-frequency ECG (UHF-ECG) technique, recently introduced, provides ventricular activation patterns and identifies electrical dyssynchrony. UHF-ECG interpretation assumes that the ECG amplitude decreases rapidly at higher frequencies with increasing distance from the source. Based on this property, nearby sources can be distinguished. Unfortunately, this assumption has never been experimentally proven. Purpose The study aims to show whether there is a decrease in ECG amplitude with distance from the source and whether this decrease is frequency-dependent. Methods UHF-ECG resting recordings were acquired in 106 subjects. We used a standard electrode setup for a 12-lead ECG. In addition, we recorded UHF-ECG from six extended precordial electrodes. With the help of magnetic resonance images, we measured the distance of all precordial leads to the ventricular geometrical center. For each lead, we determined the amplitude of frequency components by computing the area of the averaged amplitude envelopes in the region of the QRS complex. We analyzed five frequency bands: LF (0.2-20 Hz), MF (20-80 Hz), HF (80-300 Hz), UHF1 (300-500 Hz), and UHF2 (800-1000 Hz). Results The left panel in the figure shows the relative amplitude decrease in five frequency bands with relative distance from the depolarization source. The decay coefficient in the LF band is only 0.56 [0.84, 0.29] median [percentile 25,75] and gradually grows up to 1.23 [1.40, 0.98] in the UHF2 band, p<0.005. It indicates that the amplitude of the ECG signal from distant areas of the ventricles in the LF band drops approximately 50 percent of the original value, and the amplitude in the UHF2 band drops to only 20 percent of the original value. The right panel demonstrates an amplitude decrease with distance and frequency in the background of heart geometry. The circle's area corresponds to the relative signal intensity recorded in a particular body surface electrode V1, V4, and V6 from RV, LV, and septal ventricular segments. The circle marked in red has a size of 1 and is at the smallest distance from the electrode. The right panel includes the LF band (regular ECG) and the UHF2 band of 800-1000 Hz. Conclusion The results show significant differences in ECG signal decay between frequency bands and the ability to localize nearby sources using UHF-ECG. Verifying the principle of frequency dependence of ECG properties opens possibilities for a more accurate description of activation patterns. Combining multiple frequency bands (broadband ECG) has the potential to distinguish the activation of near and distant sources.Different decrease in ECG amplitude

Explainable AI for myocardial infarction using the vectorcardiogram

Abstract Background and purpose Myocardial infarction (MI), a widespread cardiovascular ailment, often lacks immediate patient awareness being early and accurate MI diagnosis crucial for lifesaving interventions. In artificial intelligence studies, ECG signals are extensively explored, while fewer examine the potential of vectorcardiogram (VCG) which offers a three-dimensional cardiac dipole representation. Traditional Machine Learning methods manually extract features; some studies advocate VCG features. Deep Learning, especially Convolutional Neural Networks (CNN), automates feature extraction. Although ECG contains redundant information from the VCG's 12 projections, no studies directly apply CNNs to VCG, nor assess the impact of redundancy on predictive ability. This study aims to bridge this gap by investigating CNNs' diagnostic capabilities with VCG inputs, specifically focusing on improving MI classification. Methods The PTB-XL database was used for this study, containing 10-seconds ECG data from various pathologies, including MI. The database encompasses 9,481 records from healthy patients and 4,129 records from patients with MI. Data was divided into 80% for training and 20% for testing. Consequently, the dataset was expanded to have 52% representing MI and 48% representing healthy cases. We designed a 1D-CNN classifier (Figure 1.A) with three convolutional layers. The network was trained using both ECG and VCG reconstruction through the Inverse Dower Transform. Results The trained model achieved accuracy values of 0.895 and 0.914 for the test set when utilizing ECG and VCG, respectively. These results underscore how the information condensation in the VCG contributes to an enhanced predictive capacity of the model. In a qualitative analysis, activation maps extracted from the CNN illustrate heightened activation in the QRS complex when the disease is present, contrasting with diminished activation in its absence. These maps depict the weight assigned to each signal segment, enabling the classification of signals into healthy and diseased categories. The presented figure exemplifies a notable activation in the QRS loop in the case of myocardial infarction (MI), aligning with the observed alteration in the cardiac signal associated with this pathology (Figure 1.B). Conclusion In this study, an interpretable CNN has been designed for MI detection, with the VCG yielding better results than the ECG for this task due to information compression. This provides a new insight for enhancing classification capability in Deep Learning models, allowing for advancements in the MI detection.

Designing optimal prototype filters for maximally decimated Cosine Modulated filter banks with rapid convergence

An analytic design of a prototype filter for M-channel maximally decimated cosine-modulated Near Perfect Reconstruction (NPR) filter banks is proposed in this work. The prototype filter is created using the least-square (CLS) method with weighted constraints, which is one-dimensional and requires single-parameter optimization. Compared to existing approaches, this suggested method achieves rapid convergence by analytically determining the optimal step size, ensuring the 3 dB cutoff frequency at π/2 M. The simulation results for design examples outperform the techniques in the available literature in terms of amplitude and aliasing distortion, reaching distortion around 2.4489 × 10−4 and 3.4907 × 10−9, respectively. This optimization algorithm's usefulness is further demonstrated with the sub-band coding of ECG signals. Implementing optimal prototype filters has tangible real-world effects, especially in critical sectors like healthcare and communications, improving diagnostics accuracy, data transmission efficiency, and overall performance.

#2570 Non-invasive blood pressure monitoring in dialysis patients using pulse wave velocity

Abstract Background and Aims Dialysis patients are at high risk of circulatory dysregulation and hypotension during treatment. Intradialytic hypotension (IDH) is a potentially life-threatening complication. The main causes are high ultrafiltration volumes and high ultrafiltration rates when patients tend to gain a lot of weight between dialysis treatments. A lack of patient compliance with regard to drinking volumes and a lack of residual diuresis are risk factors for this. Close circulatory monitoring and therapy settings to avoid absolute or relatively high UF rates to prevent IDH are frequently used. These are primarily the frequent non-invasive measurement of blood pressure using cuffs, the measurement of relative blood volume and the adjustment of ultrafiltration rates to these values. Prophylactically, the dialysis time can be extended, ultrafiltration profiles or ultrafiltration breaks can be planned, the dialysate temperature can be lowered and the composition of the dialysate can be adjusted. Despite all attempts, however, the incidence of IDH in everyday clinical practice is too high. The non-invasive measurements with cuffs are not carried out frequently enough, the measurement of the relative blood volume only covers a partial range, sudden changes in vascular tension or cardiac output are not recorded. Method The pulse wave velocity (PWV) is a well-known parameter for describing the propagation of the pressure wave after ejection of the heart beat volume in the arteries. The PWV is dependent on blood pressure, vessel wall tension and cardiac output. Changes in blood pressure can therefore be recognised by measuring the PWV. With the help of a measuring device (PESAS) for the ECG signal and peripheral pulse plethysmography, PWV can be measured non-invasively, continuously, for every heartbeat and, above all, hardly noticed by the patient. The PESAS allows the time-synchronised measurement of ECG and peripheral plethysmography curves and thus the determination of PWV. It is an experimental device. We would like to present a study showing the use of a measuring device (PESAS) in 15 prevalent dialysis patients for the detection of IDH. All study participants were examined during 3 dialysis treatments. Results The measurement of PWV in chronic dialysis patients during treatment is technically feasible. The measurement is tolerated by the patients. Changes in PWV could be detected in the patients during the course of dialysis; these were also dependent on the ultrafiltration rates. Individual patient reactions to volume withdrawal varied, with both increases and decreases in PWV being demonstrated. An acute change in PWV was usually associated with a drop in blood pressure. Continuous measurement allowed IDH events to be detected earlier than intermittent blood pressure measurement with cuffs. Sudden drops in blood pressure could be detected better and earlier with the PESAS system than by measuring the relative blood volume. Conclusion The PESAS system is suitable for monitoring dialysis patients to prevent IDH. Further adjustments and optimisation of the evaluation algorithms are necessary.

Compression of electrocardiogram signals using compressive sensing technique based on curvelet transform toward medical applications

AbstractElectrocardiogram (ECG) signals can be monitored from many patients based on healthcare systems. To enhance these systems, the ECG signals should be collected and then stored in a cloud platform for later analysis. Hence, ECG signals can be utilized to diagnose heart diseases. However, the ECG signals require great internet capacity. So, compression techniques can be implemented to reduce a memory storage capacity for these signals. One of the potential compression techniques is the compressive sensing (CS). This paper proposes a CS technique to compress ECG signals. This technique is used to reduce sampling rates of the ECG signals to be less than the Nyquist rate. Moreover, a framework is suggested for the compression of maternal and fetal ECG signals. The compression of these signals is based on the curvelet transform (CT) to produce sparsity in ECG signals. The MIT-BIH database are utilized for testing the ECG signals. This database includes several ECG signals with various sampling rates, such as aberrant and normal signals. The proposed CS technique achieved a compression ratio (CR) of 15.7 with an accuracy of 98.2%. Also, a percentage root mean difference (PRD) is utilized to calculate the performance of the reconstructed ECG signals. The achieved value of the PRD is 2.0.

Thermally Conductive and UV-EMI Shielding Electronic Textiles for Unrestricted and Multifaceted Health Monitoring

HighlightsAg nanoparticles are evenly plated on BN-embedded fibers as a thermally conductive sheath, bridging the insulating interface between fibers to construct a 3D heat transfer network.The LM-printed e-textile shows outstanding capability for the monitoring of human ECG, sEMG, and EEG signals even under intense EM interference, when commercial electrodes cannot work properly.The e-textile simultaneously manifests excellent EMI shielding (SET > 65, X-band) and UV protection (UPF = 143.1) performance, thus protecting the device and skin from harmful radiation.

Searching for the Best Machine Learning Algorithm for the Detection of Left Ventricular Hypertrophy from the ECG: A Review.

Background: Left ventricular hypertrophy (LVH) is a powerful predictor of future cardiovascular events. Objectives: The objectives of this study were to conduct a systematic review of machine learning (ML) algorithms for the identification of LVH and compare them with respect to the classical features of test sensitivity, specificity, accuracy, ROC and the traditional ECG criteria for LVH. Methods: A search string was constructed with the operators "left ventricular hypertrophy, electrocardiogram" AND machine learning; then, Medline and PubMed were systematically searched. Results: There were 14 studies that examined the detection of LVH utilizing the ECG and utilized at least one ML approach. ML approaches encompassed support vector machines, logistic regression, Random Forest, GLMNet, Gradient Boosting Machine, XGBoost, AdaBoost, ensemble neural networks, convolutional neural networks, deep neural networks and a back-propagation neural network. Sensitivity ranged from 0.29 to 0.966 and specificity ranged from 0.53 to 0.99. A comparison with the classical ECG criteria for LVH was performed in nine studies. ML algorithms were universally more sensitive than the Cornell voltage, Cornell product, Sokolow-Lyons or Romhilt-Estes criteria. However, none of the ML algorithms had meaningfully better specificity, and four were worse. Many of the ML algorithms included a large number of clinical (age, sex, height, weight), laboratory and detailed ECG waveform data (P, QRS and T wave), making them difficult to utilize in a clinical screening situation. Conclusions: There are over a dozen different ML algorithms for the detection of LVH on a 12-lead ECG that use various ECG signal analyses and/or the inclusion of clinical and laboratory variables. Most improved in terms of sensitivity, but most also failed to outperform specificity compared to the classic ECG criteria. ML algorithms should be compared or tested on the same (standard) database.