Electrocardiogram (ECG) signals are indispensable for diagnosing heart conditions and capturing the subtle rhythms of cardiac function. As wearable single-lead ECG devices gain wider adoption, there is an urgent need for robust and efficient algorithms capable of handling large-scale, unannotated data. We address this challenge with NERULA (Non-contrastive ECG and Reconstruction Unsupervised Learning Algorithm), a self-supervised framework specifically tailored for single-lead ECG signals. NERULA’s dual-pathway architecture combines ECG reconstruction and non-contrastive learning to extract detailed cardiac features. Our patch-masking strategy, using both masked and inverse-masked signals, enhances model robustness against real-world incomplete or corrupted data. The non-contrastive pathway aligns representations of masked and inverse-masked signals, while the reconstruction pathway comprehends and reconstructs missing features. We show that combining generative and discriminative paths into the training spectrum leads to better results by outperforming state-of-the-art self-supervised learning benchmarks in various tasks, demonstrating superior performance in ECG analysis, including arrhythmia classification, gender classification, and age regression. NERULA’s dual-pathway design offers a robust, efficient solution for comprehensive ECG signal interpretation.