Deep learning in ECG diagnosis: A review

Artificial Intelligence in Cardiology-A Narrative Review of Current Status

Artificial intelligence (AI) is an integral part of clinical decision support systems (CDSS), offering methods to approximate human reasoning and computationally infer decisions. Such methods are generally based on medical knowledge, either directly encoded with rules or automatically extracted from medical data using machine learning (ML). ML techniques, such as Artificial Neural Networks (ANNs) and support vector machines (SVMs), are based on mathematical models with parameters that can be optimally tuned using appropriate algorithms. The ever-increasing computational capacity of today’s computer systems enables more complex ML systems with millions of parameters, bringing AI closer to human intelligence. With this objective, the term deep learning (DL) has been introduced to characterize ML based on deep ANN (DNN) architectures with multiple layers of artificial neurons. Despite all of these promises, the impact of AI in current clinical practice is still limited. However, this could change shortly, as the significantly increased papers in AI, machine learning and deep learning in cardiology show. We highlight the significant achievements of recent years in nearly all areas of cardiology and underscore the mounting evidence suggesting how AI will take a central stage in the field.

An Effective and Lightweight Deep Electrocardiography Arrhythmia Recognition Model Using Novel Special and Native Structural Regularization Techniques on Cardiac Signal

Recently, cardiac arrhythmia recognition from electrocardiography (ECG) with deep learning approaches is becoming popular in clinical diagnosis systems due to its good prognosis findings, where expert data preprocessing and feature engineering are not usually required. But a lightweight and effective deep model is highly demanded to face the challenges of deploying the model in real-life applications and diagnosis accurately. In this work, two effective and lightweight deep learning models named Deep-SR and Deep-NSR are proposed to recognize ECG beats, which are based on two-dimensional convolution neural networks (2D CNNs) while using different structural regularizations. First, 97720 ECG beats extracted from all records of a benchmark MIT-BIH arrhythmia dataset have been transformed into 2D RGB (red, green, and blue) images that act as the inputs to the proposed 2D CNN models. Then, the optimization of the proposed models is performed through the proper initialization of model layers, on-the-fly augmentation, regularization techniques, Adam optimizer, and weighted random sampler. Finally, the performance of the proposed models is evaluated by a stratified 5-fold cross-validation strategy along with callback features. The obtained overall accuracy of recognizing normal beat and three arrhythmias (V-ventricular ectopic, S-supraventricular ectopic, and F-fusion) based on the Association for the Advancement of Medical Instrumentation (AAMI) is 99.93%, and 99.96% for the proposed Deep-SR model and Deep-NSR model, which demonstrate that the effectiveness of the proposed models has surpassed the state-of-the-art models and also expresses the higher model generalization. The received results with model size suggest that the proposed CNN models especially Deep-NSR could be more useful in wearable devices such as medical vests, bracelets for long-term monitoring of cardiac conditions, and in telemedicine to accurate diagnose the arrhythmia from ECG automatically. As a result, medical costs of patients and work pressure on physicians in medicals and clinics would be reduced effectively.

An Ensemble Deep Learning Model with a Gene Attention Mechanism for Estimating the Prognosis of Low-Grade Glioma

While estimating the prognosis of low-grade glioma (LGG) is a crucial problem, it has not been extensively studied to introduce recent improvements in deep learning to address the problem. The attention mechanism is one of the significant advances; however, it is still unclear how attention mechanisms are used in gene expression data to estimate prognosis because they were designed for convolutional layers and word embeddings. This paper proposes an attention mechanism called gene attention for gene expression data. Additionally, a deep learning model for prognosis estimation of LGG is proposed using gene attention. The proposed Gene Attention Ensemble NETwork (GAENET) outperformed other conventional methods, including survival support vector machine and random survival forest. When evaluated by C-Index, the GAENET exhibited an improvement of 7.2% compared to the second-best model. In addition, taking advantage of the gene attention mechanism, HILS1 was discovered as the most significant prognostic gene in terms of deep learning training. While HILS1 is known as a pseudogene, HILS1 is a biomarker estimating the prognosis of LGG and has demonstrated a possibility of regulating the expression of other prognostic genes.

State-of-the-art Deep Learning Methods on Electrocardiogram Data: A Systematic Review (Preprint)

Background

Electrocardiogram (ECG) is one of the most common noninvasive diagnostic tools that can provide useful information regarding a patient’s health status. Deep learning (DL) is an area of intense exploration that leads the way in most attempts to create powerful diagnostic models based on physiological signals.

Objective

This study aimed to provide a systematic review of DL methods applied to ECG data for various clinical applications.

Methods

The PubMed search engine was systematically searched by combining “deep learning” and keywords such as “ecg,” “ekg,” “electrocardiogram,” “electrocardiography,” and “electrocardiology.” Irrelevant articles were excluded from the study after screening titles and abstracts, and the remaining articles were further reviewed. The reasons for article exclusion were manuscripts written in any language other than English, absence of ECG data or DL methods involved in the study, and absence of a quantitative evaluation of the proposed approaches.

Results

We identified 230 relevant articles published between January 2020 and December 2021 and grouped them into 6 distinct medical applications, namely, blood pressure estimation, cardiovascular disease diagnosis, ECG analysis, biometric recognition, sleep analysis, and other clinical analyses. We provide a complete account of the state-of-the-art DL strategies per the field of application, as well as major ECG data sources. We also present open research problems, such as the lack of attempts to address the issue of blood pressure variability in training data sets, and point out potential gaps in the design and implementation of DL models.

Conclusions

We expect that this review will provide insights into state-of-the-art DL methods applied to ECG data and point to future directions for research on DL to create robust models that can assist medical experts in clinical decision-making.

Deep 3D Neural Network for Brain Structures Segmentation Using Self-Attention Modules in MRI Images

In recent years, the use of deep learning-based models for developing advanced healthcare systems has been growing due to the results they can achieve. However, the majority of the proposed deep learning-models largely use convolutional and pooling operations, causing a loss in valuable data and focusing on local information. In this paper, we propose a deep learning-based approach that uses global and local features which are of importance in the medical image segmentation process. In order to train the architecture, we used extracted three-dimensional (3D) blocks from the full magnetic resonance image resolution, which were sent through a set of successive convolutional neural network (CNN) layers free of pooling operations to extract local information. Later, we sent the resulting feature maps to successive layers of self-attention modules to obtain the global context, whose output was later dispatched to the decoder pipeline composed mostly of upsampling layers. The model was trained using the Mindboggle-101 dataset. The experimental results showed that the self-attention modules allow segmentation with a higher Mean Dice Score of 0.90 ± 0.036 compared with other UNet-based approaches. The average segmentation time was approximately 0.038 s per brain structure. The proposed model allows tackling the brain structure segmentation task properly. Exploiting the global context that the self-attention modules incorporate allows for more precise and faster segmentation. We segmented 37 brain structures and, to the best of our knowledge, it is the largest number of structures under a 3D approach using attention mechanisms.

An Approach for Deep Learning in ECG Classification Tasks in the Presence of Noisy Labels

Cardiovascular disease (CVD) is a serial of diseases with global leading causes of death. Electrocardiogram (ECG) is the most commonly used basis for CVD diagnosis due to its low cost and no injury. Due to the great performance shown in classification tasks with large-scale data sets, deep learning has been widely applied in ECG diagnosis. Manual labeling is a time-consuming and labor-intensive job, which makes it error-prone and easy to labeled wrongly. These noisy labels cause deterioration in performance since deep neural network is easy to over-fitting with noisy labels. However, currently, only limited studies have been concerned with this problem. To alleviate the performance degradation caused by noisy labels, we come up with an optimization method combining data clean and anti-noise loss function. Our method filters the noisy data by data-clean method, followed by training the network with boot-hard loss function. The experiment is carried on MIT-BIH arrhythmia database and we take a 1-D CNN model for test. The result indicates that our optimization method can produce an effective improvement for noisy label problems when the proportion of incorrect labels ranging from 10% to 50%.Clinical Relevance- The proposed algorithm can be potentially applied to deal with the noisy label problem in ECG diagnosis task.

A One-Dimensional Siamese Few-Shot Learning Approach for ECG Classification under Limited Data

Electrocardiogram (ECG) is mainly used by medical domain to diagnose arrhythmia. With the development of deep learning algorithms in the ECG classification field, related algorithms have achieved very high accuracy. However, the training of deep learning algorithms always requires large amounts of samples, while the labeled samples are often lacked in the field of medical signals. Therefore, the performance of deep learning algorithms will be greatly restricted. To overcome the sample scarcity problem, we propose a few-shot ECG classification approach based on the Siamese network. This network architecture first uses two one-dimensional convolutional neural network (CNN) that share weights to extract feature vectors of the paired input signals. Then, L1-distance between the two feature vectors is calculated and inputted into the fully connected layer with an activation function sigmoid to determine whether the input pairs belong to same category. We validated our method on the MIT-BIH arrhythmia database. By experiments, our method performs better than existing networks under the circumstance of extremely few amounts of data.

Automatic 12-Leading Electrocardiogram Classification Network with Deformable Convolution

Electrocardiogram (ECG) is an electrical signal that helps monitor the physiology of the heart. A complete ECG record includes 12 leads, each reflecting features from a different angle of the heart. In recent years, various deep learning algorithms, especially convolutional neural networks (CNN), have been applied to detect ECG features. However, the conventional CNN can only extract the local features and cannot extract the data correlation across the leads of ECG. Based on deformable convolution networks (DCN), this article proposes a new neural network structure (DCNet) to detect ECG features. The network architecture consists of four DCN blocks and a classification layer. For the ECG classification task, in a DCN block, the combination of normal convolution and deformable convolution with better effect was testified by the experiments. Based on the feature learning capability of DCN, the architecture can better extract the characteristics between leads. Using the public 12-leading ECG data in CPSC-2018, the diagnostic accuracy of this architecture is the highest, reaching 86.3%, which is superior to other common network architectures with good results in ECG signal classification.Clinical relevance-In this paper, we proposed an effective automatic ECG classification model that can reduce medical staff workload.

Dual Attention Convolutional Neural Network Based on Adaptive Parametric ReLU for Denoising ECG Signals with Strong Noise

Electrocardiogram (ECG) signal is one of the most important methods for diagnosing cardiovascular diseases but is usually affected by noises. Denoising is therefore necessary before further analysis. Deep learning-related methods have been applied to image processing and other domains with great success but are rarely used for denoising ECG signals. This paper proposes an effective and simple model of encoder-decoder structure for denoising ECG signals (APR-CNN). Specifically, Adaptive Parametric ReLU (APReLU) and Dual Attention Module (DAM) are introduced in the model. Rectified Linear Unit (ReLU) is replaced with the APReLU for better negative information retainment. The DAM is an attention-based module consisting of a channel attention module and spatial attention module, through which the inter-spatial and inter-channel relationship of the input data are exploited. We tested our model on the MIT-BIH dataset, and the results show that the APR-CNN can handle ECG signals with a different signal-to-noise ratio (SNR). The comparative experiment proves our model is better than other deep learning and traditional methods.Clinical Relevance- This paper proposed a method capable of denoising ECG signals with strong noise to alleviate difficulties for further medical analysis.