Abstract:Objective Automated segmentation of intracardiac electrograms and extraction of fundamental cycle length intervals is crucial for reproducible arrhythmia substrate analysis conducted during electrophysiology procedures. The objective of this study is to develop a robust, computationally efficient end-to-end model for the precise electrogram multi-channel delineation using a highly imbalanced dataset. Methods A temporal deep convolutional neural network (CNN) based on the UNet architecture incorporating convolutional layers of varying dilation rates was implicitly regularized through data augmentations (DAs), a domain specific Tversky loss function, and distinct labelling strategies for segments comprising atrial fibrillation (AF). An exploratory study utilizing Bayesian search was conducted to optimize architectural and loss function hyperparameters. The impact of dilated convolutions, data augmentations, and labelling strategies on the performance and generalization capability was assessed through an ablation study. The performance of different models was evaluated using a cross-validation procedure and two independent test datasets derived from two separate patient cohorts containing 326, 84, and 97 electrograms encompassing sinus rhythms, abnormal complexes during ongoing tachycardias, and stimulation protocols. Results A UNet model with optimized loss hyperparameters, a dilated receptive field, and atrial fibrillation (AF) annotated as a positive class (D-UNet-L) achieved an average Sørensen-Dice coefficient (SDC) of 84.9 % on recordings with regular atrial beats across test datasets, surpassing the performance of models without loss optimization (81.5 %), without dilated kernels (81.3 %), and with inversed AF labelling (77.5 %). Notably, the highest average accuracy (Acc) of 95.8 % for AF recordings was obtained by a model trained on negatively assigned AF segments, outperforming D-UNet-L (88.9 %), the model without loss optimization (81.5 %), and the model without dilated kernels (81.3 %). The reference D-UNet-L model exhibited overall root-mean-square errors of 8.3 and 9.0 ms across test datasets. Additionally, 61.5 % and 20.8 % of delineations exhibited absolute errors below 5 ms and 10 ms, respectively. Disabling data augmentation (DAs) resulted in a 2.7 % decrease in validation SDC and a 5.3 % increase in training SDC." Conclusion Generalization capability across independent datasets was improved by employing exponentially weighted Tversky loss. The model's segmentation performance on longer sequences with atrial fibrillation was improved by incorporating dilated convolution kernels. Noise-aware and morphology data augmentations effectively mitigated overfitting potential in a limited training dataset. Label noise introduced by annotating atrial fibrillation sequences into a positive class strengthened regularization of the model, particularly in its ability to identify regular beats. However, it also negatively impacted performance on F-waves. Significance The proposed method diminishes the dependence on manual measurements in complex electrophysiology measurements and enhances utilization of the collected electrophysiology data. The findings underscore the substantial impact of intrinsic regularization techniques and establish guidelines to facilitate methodological decision-making processes.

Unsupervised sequence-to-sequence learning for automatic signal quality assessment in multi-channel electrical impedance-based hemodynamic monitoring

Unsupervised sequence-to-sequence learning for automatic signal quality assessment in multi-channel electrical impedance-based hemodynamic monitoring

Machine Learning-based Signal Quality Assessment for Cardiac Volume Monitoring in Electrical Impedance Tomography

Deep Learning Assessment of Myocardial Infarction From MR Image Sequences.

Multi-task cascaded assessment of signal quality for long-term single-lead ECG monitoring

Joint Spatio-Temporal Features Constrained Self-Supervised Electrocardiogram Representation Learning

Multi-label Classification of Abnormalities in 12-Lead ECG Using Deep Learning.

Variational Autoencoders for Biomedical Signal Morphology Clustering and Noise Detection

Unsupervised Learning-Based Non-Invasive Fetal ECG Muti-Level Signal Quality Assessment

Weak self-supervised learning for seizure forecasting: a feasibility study

A New Automated Signal Quality-Aware ECG Beat Classification Method for Unsupervised ECG Diagnosis Environments

Wearable Electrocardiogram Quality Assessment Using Wavelet Scattering and LSTM

Multimodal Variational Autoencoder for Low-cost Cardiac Hemodynamics Instability Detection

Learning with incomplete labels of multisource datasets for ECG classification

Deep Metric Learning for the Hemodynamics Inference with Electrocardiogram Signals

A Non-Intrusive Neural Quality Assessment Model for Surface Electromyography Signals

CE-SSL: Computation-Efficient Semi-Supervised Learning for ECG-based Cardiovascular Diseases Detection

ECG-SL: Electrocardiogram(ECG) Segment Learning, a deep learning method for ECG signal

Weakly-Supervised Learning via Multi-Lateral Decoder Branching for Guidewire Segmentation in Robot-Assisted Cardiovascular Catheterization

Multi-channel delineation of intracardiac electrograms for arrhythmia substrate analysis using implicitly regularized convolutional neural network with wide receptive field

Spatiotemporal self-supervised representation learning from multi-lead ECG signals