A Mamba-augmented model is proposed that inserts selective state-space blocks at the convolutional bottleneck, combining local feature extraction with long-range temporal modeling at linear complexity, and substantially improves ST/T-change diagnoses, which depend on broad, context-sensitive waveforms.
Abstract
Electrocardiogram (ECG) recordings are corrupted by non-stationary noise sources that degrade diagnostic reliability, particularly in ambulatory and long-duration recordings. Deep learning denoisers exist, but convolutional architectures are limited by their receptive field, transformer-based models scale quadratically with sequence length, and diffusion-based approaches incur prohibitive inference cost. We propose a Mamba-augmented model that inserts selective state-space blocks at the convolutional bottleneck, combining local feature extraction with long-range temporal modeling at linear complexity. We comprehensively evaluate the proposed model with respect to reconstruction fidelity, noise robustness, recording-length scaling, and downstream diagnostic classification across over 40 pathology classes. On synthetic and real datasets, our model achieves the highest SNR and lowest RMSE, with the Mamba advantage increasing with sequence length and in low-SNR regimes. On classification with two independent classifiers, the proposed Mamba-based models achieve the best macro AUROC among all denoisers and improve over their convolutional base models. Calibration is more nuanced and classifier-dependent: denoising improves Binary Cross-Entropy and Brier score on Inception1D but often fails to beat the noisy input on ResNet1D-Wang, and the lead-specific Mamba variant is the only denoiser to improve both calibration metrics over the noisy baseline on both classifiers. Per-class analysis reveals a morphology-dependent benefit: Mamba substantially improves ST/T-change diagnoses, which depend on broad, context-sensitive waveforms.
Electrocardiogram (ECG) recordings are frequently corrupted by nonstationary artifacts such as baseline wander, muscle activity, and electrode motion, which can distort waveform morphology, reduce signal quality, and compromise downstream analysis. Although recent deep learning approaches have improved signal-level fid...
Thanaporn Viriyasaranon, J. Choi· IEEE journal of biomedical a...· 0 citations
It is demonstrated that hybrid spatio-temporal architectures can achieve diagnostic performance comparable to strong convo-lutional baselines while offering significantly improved trans-parency through a quantitative comparison of representative models and visual analysis of performance—interpretability trade-offs.
Rhivu Dutta, N. Suma· JIMS8I - International Journ...· 0 citations
A hybrid CNN-SSM-Attention backbone for 12-lead ECG classification is introduced, which provides strong supervised baselines under a compact parameter budget and improves transfer, particularly in reduced-label settings and under both full fine-tuning and LoRA-based adaptation.
Y. Bazi, Sarah Aljuhani, M. M. Al Rahhal et al.· arXiv.org· 0 citations
Automatic recognition of abnormal heartbeats from a single electrocardiogram lead is limited by two coupled difficulties: the clinically important beats are the rarest ones, and the classifier must be small enough to run on the ambulatory hardware that records them. This study presents a compact fully convolutional mod...
Pradeep Kumar· Natural Resources for Human...· 0 citations
An orthogonal-polynomial imaging framework is introduced that encodes 12-lead ECGs into a single two-dimensional image through an invertible transformation prior to coarse-graining, producing compact image representations that retain sufficient information for accurate classification.
Karan Singh, Pingal Pratyush Nath, U. Sinha et al.· Frontiers in Artificial Inte...· 0 citations
To facilitate the analysis of electrocardiogram (ECG) signals collected in challenging environments, denoising algorithms must be adaptable to diverse noise sources and varying intensities. This study aims to improve the generalizability of denoising models trained on specific datasets, allowing effective performance i...
Jia-Feng Qiu, Shi-Sheng Chen, Fang-Zhuo Jin et al.· IEEE journal of biomedical a...· 0 citations
Related blog posts
MIT News · Artificial Intelligence· news.mit.eduOct 13, 2026
Machine-learning networks can learn without detailed instructions from a designer. Those that do may offer insight into how the brain reorganizes itself during learning.
MIT News · Artificial Intelligence· news.mit.eduSep 29, 2026
Professor Sherry Turkle’s new book, “Artificial Intimacy,” offers a withering critique of chatbots and the antisocial dynamics she believes they encourage.