What question did this study set out to answer?

This research aims to improve the interpretability of ECG arrhythmia classification through a new deep learning architecture.

January 10, 2026Discover Applied Sciences0 citationsOpen Access

Ots-net: unlocking mechanistic interpretability in ECG arrhythmia classification

TDThanh Dinh DaoVietnamese-German University DTDang Trieu TonVietnam National University Ho Chi Minh City KTKien TrangMonash University Malaysia

Key Result

OTS-Net achieved 99.09% accuracy in ECG arrhythmia classification with only 103 K parameters, providing unprecedented interpretability of model insights.

Key Points

This research aims to improve the interpretability of ECG arrhythmia classification through a new deep learning architecture.
Introduced OTS-Net, integrating a spatio-temporal convolutional backbone with artificial Kuramoto oscillatory neurons.
Employed dynamic synchronization modeling to reflect cardiac pacemaker cell behavior.
Utilized the MIT-BIH Arrhythmia Database for performance evaluation.
Developed an analytical suite for visualizing the model's internal states and synchronizations.
Achieved 99.09% accuracy with only 103 K parameters.
Provided unprecedented insights into the model's internal anatomy and physiology.
Enabled diagnostic explanations based on measurable phase-locking and network coherence.

Structured PICO

Population

99,952 annotated heartbeats from 48 half-hour ambulatory ECG recordings of 47 subjects from the MIT-BIH Arrhythmia Database, categorized into Normal (N), Supraventricular Ectopic (S), Ventricular Ectopic (V), and Fusion (F) classes.

Intervention

Oscillatory Temporal Synchronization Net (OTS-Net), a deep learning architecture integrating a spatio-temporal convolutional backbone with Artificial Kuramoto Oscillatory Neurons (AKOrN).

Outcome

ECG arrhythmia classification accuracy

OTS-Net provides a highly accurate, parameter-efficient, and mechanistically interpretable deep learning architecture for ECG arrhythmia classification by modeling biological synchronization dynamics.

Main Result

Absolute Event Rate: 0% vs 0%

Abstract

Deep learning models have achieved remarkable accuracy in electrocardiogram (ECG) arrhythmia classification, yet their “black-box” nature remains a major barrier to clinical trust. This paper introduces the Oscillatory Temporal Synchronization Net (OTS-Net), a novel, biophysically-plausible architecture designed to address this performance-interpretability deadlock. OTS-Net integrates a spatio-temporal convolutional backbone with Artificial Kuramoto Oscillatory Neurons (AKOrN), shifting the paradigm from static pattern recognition to dynamic synchronization modeling. By emulating the collective behavior of cardiac pacemaker cells, our approach achieves competitive performance −99.09% accuracy on the MIT-BIH Arrhythmia Database with an ultra-efficient 103 K parameters-while offering a framework for unprecedented mechanistic interpretability. We present an analytical suite that decodes the model’s internal state, visualizing its learned “anatomy” (coupling matrix) and emergent “physiology” (class-specific synchronization patterns). This enables diagnostic explanations based on quantifiable disruptions in phase-locking and network coherence, rather than opaque feature correlations. This work demonstrates that by grounding AI in the fundamental principles of biological synchronization, it is possible to create systems that are not only accurate and efficient, but are also inherently transparent and trustworthy.

Ask AI

Helpful

Bookmark

View Full Paper