A robust ECG denoising technique using variable frequency complex demodulation

Adversarial Spatiotemporal Contrastive Learning for Electrocardiogram Signals

Extracting invariant representations in unlabeled electrocardiogram (ECG) signals is a challenge for deep neural networks (DNNs). Contrastive learning is a promising method for unsupervised learning. However, it should improve its robustness to noise and learn the spatiotemporal and semantic representations of categories, just like cardiologists. This article proposes a patient-level adversarial spatiotemporal contrastive learning (ASTCL) framework, which includes ECG augmentations, an adversarial module, and a spatiotemporal contrastive module. Based on the ECG noise attributes, two distinct but effective ECG augmentations, ECG noise enhancement, and ECG noise denoising, are introduced. These methods are beneficial for ASTCL to enhance the robustness of the DNN to noise. This article proposes a self-supervised task to increase the antiperturbation ability. This task is represented as a game between the discriminator and encoder in the adversarial module, which pulls the extracted representations into the shared distribution between the positive pairs to discard the perturbation representations and learn the invariant representations. The spatiotemporal contrastive module combines spatiotemporal prediction and patient discrimination to learn the spatiotemporal and semantic representations of categories. To learn category representations effectively, this article only uses patient-level positive pairs and alternately uses the predictor and the stop-gradient to avoid model collapse. To verify the effectiveness of the proposed method, various groups of experiments are conducted on four ECG benchmark datasets and one clinical dataset compared with the state-of-the-art methods. Experimental results showed that the proposed method outperforms the state-of-the-art methods.

Evaluation of comorbid psychological disorders in functional gastrointestinal disorders patients by vibraimage technology: protocol of a prospective, single-center trial

Introduction

Functional gastrointestinal disorders (FGIDs) affect over 40% of individuals globally, and impact the quality of life. A significant portion of FGIDs patients comorbids with anxiety and depression. Traditional screening tools for psychological disorders may lack comprehensiveness. Vibraimage technology currently enables non-contact, objective analysis of psychological indicators through high-frame-rate cameras and computer analysis of micro-movements. Therefore, this study aims to (1) explore the use of vibraimage technology as a non-contact objective method to assess the psychological status of FGIDs patients, comparing this technology with the Hospital Anxiety and Depression Scale (HADS) to evaluate its screening efficacy, and (2) observe the therapeutic outcomes of FGIDs patients with or without comorbid psychological disorders after the same conventional treatment.

Methods

This is a prospective, single-center observational trial. 276 FGIDs outpatients who visit Peking Union Medical College Hospital will be evaluated simultaneously by HADS and vibraimage technology, then to evaluate the screen efficacy of this technology. The patients will be allocated into two groups (those with or without psychological disorders). The primary endpoint is the overall rate of improvement, specifically referring to the proportion of patients who achieved Likert scores greater than or equal to 4. The secondary endpoints encompass evaluating whether there is a reduction of more than 50% in symptom evaluation scores such as IBS-SSS. Additionally, the study will assess changes in health status and quality of life using SF-36 questionnaires and the patients' satisfaction with treatment. Furthermore, psychological status will be reassessed by vibraimage technology and HADS after treatment to evaluate the effect of combined psychological factors on FGIDs treatment.

Machine learning workflow for edge computed arrhythmia detection in exploration class missions

Deep-space missions require preventative care methods based on predictive models for identifying in-space pathologies. Deploying such models requires flexible edge computing, which Open Neural Network Exchange (ONNX) formats enable by optimizing inference directly on wearable edge devices. This work demonstrates an innovative approach to point-of-care machine learning model pipelines by combining this capacity with an advanced self-optimizing training scheme to classify periods of Normal Sinus Rhythm (NSR), Atrial Fibrillation (AFIB), and Atrial Flutter (AFL). 742 h of electrocardiogram (ECG) recordings were pre-processed into 30-second normalized samples where variable mode decomposition purged muscle artifacts and instrumentation noise. Seventeen heart rate variability and morphological ECG features were extracted by convoluting peak detection with Gaussian distributions and delineating QRS complexes using discrete wavelet transforms. The decision tree classifier's features, parameters, and hyperparameters were self-optimized through stratified triple nested cross-validation ranked on F1-scoring against cardiologist labeling. The selected model achieved a macro F1-score of 0.899 with 0.993 for NSR, 0.938 for AFIB, and 0.767 for AFL. The most important features included median P-wave amplitudes, PRR20, and mean heart rates. The ONNX-translated pipeline took 9.2 s/sample. This combination of our self-optimizing scheme and deployment use case of ONNX demonstrated overall accurate operational tachycardia detection.

A portable household detection system based on the combination of bidirectional LSTM and residual block for automatical arrhythmia detection

OBJECTIVES

Arrhythmia is an important component of cardiovascular disease, and electrocardiogram (ECG) is a method to detect arrhythmia. Arrhythmia detection is often paroxysmal, and ECG signal analysis is time-consuming and expensive. We propose a model and device for convenient monitoring of arrhythmia at any time.

METHODS

This work proposes a model combining residual block and bidirectional long-term short-term memory network (BiLSTM) to detect and classify ECG signals. Residual blocks can extract deep features and avoid performance degradation caused by convolutional networks. Combined with the feature of BiLSTM to strengthen the connection relationship of the local window, it can achieve a better classification and prediction effect.

RESULTS

Model optimization experiments were performed on the MIT-BIH Atrial Fibrillation Database (AFDB) and MIT-BIH Arrhythmia Database (MITDB). The accuracy simulation results on both long and short signal was higher than 99 %. To further demonstrate the applicability of the model, validation experiments were conducted on MIT-BIH Normal Sinus Rhythm Database (NSRDB) and the Long-Term AF Database (LTAFDB) datasets, and the related recognition accuracy were 99.830 and 91.252 %, respectively. Additionally, we proposed a portable household detection system including an ECG and a blood pressure detection module. The detection accuracy was higher than 98 % using the collected data as testing set.

CONCLUSIONS

Hence, we thought our system can be used for practical application.

A novel approach for denoising electrocardiogram signals to detect cardiovascular diseases using an efficient hybrid scheme

Background

Electrocardiogram (ECG) signals are inevitably contaminated with various kinds of noises during acquisition and transmission. The presence of noises may produce the inappropriate information on cardiac health, thereby preventing specialists from making correct analysis.

Methods

In this paper, an efficient strategy is proposed to denoise ECG signals, which employs a time-frequency framework based on S-transform (ST) and combines bi-dimensional empirical mode decomposition (BEMD) and non-local means (NLM). In the method, the ST maps an ECG signal into a subspace in the time frequency domain, then the BEMD decomposes the ST-based time-frequency representation (TFR) into a series of sub-TFRs at different scales, finally the NLM removes noise and restores ECG signal characteristics based on structural self-similarity.

Results

The proposed method is validated using numerous ECG signals from the MIT-BIH arrhythmia database, and several different types of noises with varying signal-to-noise (SNR) are taken into account. The experimental results show that the proposed technique is superior to the existing wavelet based approach and NLM filtering, with the higher SNR and structure similarity index measure (SSIM), the lower root mean squared error (RMSE) and percent root mean square difference (PRD).

Conclusions

The proposed method not only significantly suppresses the noise presented in ECG signals, but also preserves the characteristics of ECG signals better, thus, it is more suitable for ECG signals processing.

Automated cardiac arrhythmia detection techniques: a comprehensive review for prospective approach

Abnormal cardiac functionality produces irregular heart rhythms which are commonly known as arrhythmias. In some conditions, arrhythmias are treated as very dangerous which may lead to sudden cardiac arrest. The incidence and prevalence of cardiac anomalies seeks early detection of arrhythmias using automated classification techniques. In the past, numerous automated arrhythmia detection techniques have been developed that are based on electrocardiogram (ECG) signal analysis. Focusing on the prospective research in this field, this article reports a comprehensive review of existing techniques that are obtained using search engines such as IEEE explore, Google scholar and science direct. Based on the review, the existing techniques are broadly categorized into two types: machine-learning and deep-learning-based techniques. In this study, it is noticed that the performance of the machine-learning-based arrhythmia detection techniques depend on pre-processing of ECG signal, R-peaks detection, features extraction and classification tools while the deep-learning-based techniques do not require the features extraction step. Generally, the existing techniques utilize Massachusetts Institute of Technology-Beth Israel Hospital arrhythmia database to evaluate the classification performance. The classification performance of automated techniques also depends on ECG data used for training and testing of the classifier. It is expected that the performance should be evaluated using a variety of ECG signals including the cases of inter-patient and intra-patient paradigm. The existing techniques also require to deal with the class-imbalance problem. In addition to this, a specific partition-ratio between training and testing datasets should be maintained for fair comparison of performance of different techniques.

Analysis of sympathetic responses to cognitive stress and pain through skin sympathetic nerve activity and electrodermal activity

We explored the non-invasive evaluation of the sympathetic nervous system (SNS) by employing two distinct physiological signals: skin sympathetic nerve activity (SKNA), extracted from electrocardiogram (ECG) signals, and electrodermal activity (EDA), a well-studied marker in the context of the SNS assessment. Our investigation focused on cognitive stress and pain; two conditions closely associated with the SNS. We sought to determine if the information and dynamics of EDA could be derived from the novel SKNA signal. To this end, ECG and EDA signals were recorded simultaneously during three experiments aimed at sympathetic stimulation, Valsalva maneuver (VM), Stroop test, and thermal-grill pain test. We calculated the integral area under the rectified SKNA signal (iSKNA) and decomposed the EDA signal to its phasic component (EDAphasic). An average delay of more than 4.6 s was observed in the onset of EDAphasic bursts compared to their corresponding iSKNA bursts. After shifting the EDAphasic segments by the extent of this delay and smoothing the corresponding iSKNA bursts, our results revealed a strong average correlation coefficient of 0.85±0.14 between the iSKNA and EDAphasic bursts, indicating a noteworthy similarity between the two signals. We also reconstructed the EDA signals with time-varying sympathetic (TVSymp) and modified TVSymp (MTVSymp) methods. Then we extracted the following features from iSKNA, EDAphasic, TVSymp, and MTVSymp signals: peak amplitude, average amplitude (aSKNA), standard deviation (vSKNA), and the cumulative duration during which the signals had higher amplitudes than a specified threshold (HaSKNA). A strong average correlation of 0.89±0.18 was found between vSKNA and subjects' self-rated pain levels during the pain test. Our statistical analysis also included applying Linear Mixed-Effects Models to check if there were significant differences in features across baseline and different levels of SNS stimulation. We then assessed the discriminating power of the features using Area Under the Receiver Operating Characteristic Curve (AUROC) and Fisher's Ratio. Finally, using all the four EDA features, a multi-layer perceptron (MLP) classifier reached the classification accuracies 95.56%, 89.29%, and 67.88% for the VM, Stroop, and thermal-grill pain control and stimulation classes. On the other hand, the highest classification accuracies based on SKNA features were achieved using K-nearest neighbors (KNN) (98.89%), KNN (89.29%), and MLP (95.11%) classifiers for the same experiments. Our comparative analysis showed the feasibility of SKNA as a novel tool for assessing the SNS with accurate classification capability, with a faster onset of amplitude increase in response to SNS activity, compared to EDA.

A Deep Convolutional Autoencoder for Automatic Motion Artifact Removal in Electrodermal Activity Signals: A Preliminary Study

Automatic motion artifact (MA) removal in electrodermal activity (EDA) signals is a major challenge because of the aperiodic and irregular characteristics of EDA. Given the lack of a suitable MA removal algorithm, a substantial amount of EDA data is typically discarded, especially during ambulatory monitoring. Current methods for MA removal in EDA are feasible when data are corrupted with low magnitude artifacts. In this study, we propose a more data-driven deep convolutional autoencoder (DCAE) for automated motion artifact removal in EDA signals. The DCAE was trained using several publicly available datasets. We used both Gaussian white noise (GWN) and real-life induced MA data records collected in a laboratory setting to corrupt the clean EDA signals. We compared the performance of our DCAE network with three state-of-the-art methods using the performance metrics the signal-to-noise ratio (SNR) improvement (SNR_imp), and the mean squared error (MSE). The proposed DCAE provided significantly higher SNR_impand lower MSE compared to three other methods for both synthetically and real-life induced MA. While the work is preliminary, this work illustrates a promising approach which can potentially be used to remove many different types of MA.

Basis pursuit sparse decomposition using tunable-Q wavelet transform (BPSD-TQWT) for denoising of electrocardiograms

The electrocardiogram (ECG) is an essential diagnostic tool to identify cardiac abnormalities. So, the primary issue in an ECG acquisition unit is noise interference. Essentially, the prominent ECG noise sources are power line interference (PLI) and Baseline drift (BD). Therefore, in the study, a new technique called the basis pursuit sparse decomposition (BPSD) using tunable-Q wavelet transform (TQWT) is proposed to remove the PLI and BD present in the ECG recordings. Chiefly, the TQWT method is a wavelet transform with distinct Quality factors (Q) which can adjust the signal to the natural non-stationary behaviour in time and space. Further, the method decomposes the signal into high-Quality factor and low-Quality factor components of wavelet coefficients to eliminate PLI and BD by choosing appropriate redundancy (r) and decomposition levels (J2). The 'r' and 'J' values are chosen based on the trial-and-error method concerning signal-to-noise ratio (SNR). It has been found that the PLI noise has been suppressed significantly with the redundancy of 3 and decomposition levels of 10; more so, the BD has been removed with the redundancy of 4 and decomposition levels of 19. The proposed method BPSD-TQWT was evaluated using the open-source MIT-BIH Arrhythmia database and the real-time ECG recordings collected through a wearable Silver Plated Nylon Woven (Ag-NyW) textile-based ECG monitoring system. The performance was then evaluated using fidelity metrics such as SNR, maximum absolute error (MAX), and normalized cross-correlation coefficient (NCC). The results were compared with IIR filter, stationary wavelet transform (SWT), non-local means (NLM) and local means (LM) methods. Using the proposed method on MIT-BIH Arrhythmia Database, performance evaluation parameters such as SNR, MAX, and NCC were improved by 4.3 dB and 6.8 dB, 0.37 and 0.78, 0.2 and 0.46 compared to IIR and SWT methods respectively. On the other hand, using the proposed method on the real-time datasets, values of SNR, MAX, and NCC were improved by 0.3 dB and 0.6 dB, 0.009 and 0.74 and 0.3 and 0.35 compared to IIR and SWT methods respectively. Finally, it can be concluded that the proposed method shows improved performance over IIR, SWT, NLM and LM methods for PLI and BD removal.

Discrete Wavelet Transform Analysis of the Electroretinogram in Autism Spectrum Disorder and Attention Deficit Hyperactivity Disorder

Background

To evaluate the electroretinogram waveform in autism spectrum disorder (ASD) and attention deficit hyperactivity disorder (ADHD) using a discrete wavelet transform (DWT) approach.

Methods

A total of 55 ASD, 15 ADHD and 156 control individuals took part in this study. Full field light-adapted electroretinograms (ERGs) were recorded using a Troland protocol, accounting for pupil size, with five flash strengths ranging from –0.12 to 1.20 log photopic cd.s.m^–2. A DWT analysis was performed using the Haar wavelet on the waveforms to examine the energy within the time windows of the a- and b-waves and the oscillatory potentials (OPs) which yielded six DWT coefficients related to these parameters. The central frequency bands were from 20–160 Hz relating to the a-wave, b-wave and OPs represented by the coefficients: a20, a40, b20, b40, op80, and op160, respectively. In addition, the b-wave amplitude and percentage energy contribution of the OPs (%OPs) in the total ERG broadband energy was evaluated.

Results

There were significant group differences (p < 0.001) in the coefficients corresponding to energies in the b-wave (b20, b40) and OPs (op80 and op160) as well as the b-wave amplitude. Notable differences between the ADHD and control groups were found in the b20 and b40 coefficients. In contrast, the greatest differences between the ASD and control group were found in the op80 and op160 coefficients. The b-wave amplitude showed both ASD and ADHD significant group differences from the control participants, for flash strengths greater than 0.4 log photopic cd.s.m^–2 (p < 0.001).

Conclusion

This methodological approach may provide insights about neuronal activity in studies investigating group differences where retinal signaling may be altered through neurodevelopment or neurodegenerative conditions. However, further work will be required to determine if retinal signal analysis can offer a classification model for neurodevelopmental conditions in which there is a co-occurrence such as ASD and ADHD.

Atrial Fibrillation Prediction from Critically Ill Sepsis Patients

Sepsis is defined by life-threatening organ dysfunction during infection and is the leading cause of death in hospitals. During sepsis, there is a high risk that new onset of atrial fibrillation (AF) can occur, which is associated with significant morbidity and mortality. Consequently, early prediction of AF during sepsis would allow testing of interventions in the intensive care unit (ICU) to prevent AF and its severe complications. In this paper, we present a novel automated AF prediction algorithm for critically ill sepsis patients using electrocardiogram (ECG) signals. From the heart rate signal collected from 5-min ECG, feature extraction is performed using the traditional time, frequency, and nonlinear domain methods. Moreover, variable frequency complex demodulation and tunable Q-factor wavelet-transform-based time-frequency methods are applied to extract novel features from the heart rate signal. Using a selected feature subset, several machine learning classifiers, including support vector machine (SVM) and random forest (RF), were trained using only the 2001 Computers in Cardiology data set. For testing the proposed method, 50 critically ill ICU subjects from the Medical Information Mart for Intensive Care (MIMIC) III database were used in this study. Using distinct and independent testing data from MIMIC III, the SVM achieved 80% sensitivity, 100% specificity, 90% accuracy, 100% positive predictive value, and 83.33% negative predictive value for predicting AF immediately prior to the onset of AF, while the RF achieved 88% AF prediction accuracy. When we analyzed how much in advance we can predict AF events in critically ill sepsis patients, the algorithm achieved 80% accuracy for predicting AF events 10 min early. Our algorithm outperformed a state-of-the-art method for predicting AF in ICU patients, further demonstrating the efficacy of our proposed method. The annotations of patients' AF transition information will be made publicly available for other investigators. Our algorithm to predict AF onset is applicable for any ECG modality including patch electrodes and wearables, including Holter, loop recorder, and implantable devices.