A lightweight deep neural network for personalized detecting ventricular arrhythmias from a single-lead ECG device

Ventricular arrhythmia (VA) is a leading cause of sudden cardiac death. Detecting VA from electrocardiograms (ECGs) using deep learning techniques has potential to improve clinical outcomes. However, developing robust deep learning models for ECG analysis remains challenging due to: (1) inter-subject diversity among different individuals, and (2) intra-subject diversity within the same subject across different physiological state over time. In this study, we address these challenges by introducing enhancements in both the pre-training and fine-tuning stages. In the pre-training stage, we propose a novel approach combining model-agnostic meta-learning (MAML) with curriculum learning (CL) to effectively address inter-subject diversity. MAML efficiently transfer knowledge from large-scale datasets and enables rapid model adaptation to new individuals using limited records. Integrating CL further enhances the effectiveness of MAML by sequentially training models from simpler to more complex tasks. For the fine-tuning stage, we propose an improved pre-fine-tuning strategy specifically designed to manage the intra-subject diversity. We evaluate our methods on three publicly available ECG datasets and one real-world clinical ECG dataset collected using a portable device. Our proposed method achieves ROC-AUC = 0.984 / F1 = 0.940 with only 10 beats per class per subject on the test set and also achieves ROC-AUC = 0.965 / F1 = 0.937 on a real-world clinical collected data. Experimental results demonstrate that our proposed approach outperforms existing comparative methods across all evaluation metrics, and have a tendency to address intra-subject diversity. Ablation studies confirm that the combination of MAML and CL leads to more uniform performance across individuals, and our enhanced pre-fine-tuning technique substantially improves model adaptation to individual-specific data.Author summary: Ventricular arrhythmias are dangerous irregular heart rhythms that can lead to sudden cardiac death if they are not detected in time. Today, many people carry small, single-lead electrocardiogram (ECG) devices that record their heart signals at home or during daily activities. These devices create new opportunities for early warning, but existing computer algorithms often struggle to recognize abnormal rhythms in different people, or even in the same person as their condition changes.We developed a new lightweight deep learning system called MetaVA that learns to “adapt” to each individual quickly. Instead of needing thousands of training samples, the system can adjust itself with only a few labeled heartbeats from the user. We also designed a simple preparation step that helps the model stay accurate even as a person’s heart signals vary over time.In tests on large public databases and a clinical study with a handheld ECG device, our method reached high accuracy while requiring very little computing power. This means MetaVA could be embedded in wearable or portable devices, offering timely, personalized detection of life-threatening arrhythmias in everyday life.