PulseExploreJournal ClubDebatesTrendingResearchersJournals
Instagram
HomeExploreJournal ClubTrending
Synapse
⌘+K
Synapse
August 22, 2025Medical Engineering & Physics5 citations

Graph-enhanced deep learning for ECG arrhythmia detection: An integration of CNN-GNN-BiLSTM approach

View Full Paper
Ask AI
Bookmark
Share

Why the study?

Early and accurate detection of cardiac arrhythmias is crucial for preventing severe cardiovascular events.

Does a CNN-GNN-BiLSTM integrated framework improve ECG arrhythmia detection accuracy compared to conventional deep learning approaches in benchmark ECG datasets?

Population

ECG recordings from MIT-BIH, PTB, Chapman-Shaoxing, and a combined 11-class dataset

Comparison

CNN-GNN-BiLSTM integrated framework vs conventional deep learning approaches

Design

Model development and validation study

Key result

The proposed CNN-GNN-BiLSTM framework achieved 96.0% overall accuracy and up to 99.89% accuracy on the MIT-BIH dataset for automated ECG arrhythmia classification.

Authors

PMPiyush MahajanAKAmit Kaul

Discussion

Loading...

Member takes

Overview

AI models may enhance ECG arrhythmia detection; leaves open prospective clinical validation before practice change.

Structured PICO

Does a CNN-GNN-BiLSTM integrated framework improve ECG arrhythmia detection accuracy compared to conventional deep learning approaches in benchmark ECG datasets?

P
Population
Three benchmark ECG datasets (MIT-BIH, PTB, and Chapman-Shaoxing) and a combined 11-class dataset
I
Intervention
CNN-GNN-BiLSTM integrated framework for automated ECG arrhythmia classification
C
Comparator
Conventional deep learning approaches
O
Outcome
Overall accuracy and dataset-specific accuracy for arrhythmia classification

A novel CNN-GNN-BiLSTM deep learning framework achieves high accuracy for automated ECG arrhythmia detection, offering a scalable solution for AI-driven cardiac monitoring.

Cite This Study

Mahajan et al. (2025) studied Cardiac arrhythmias. CNN-GNN-BiLSTM integrated framework vs. Conventional deep learning approaches was evaluated on Overall accuracy. The proposed CNN-GNN-BiLSTM framework achieved 96.0% overall accuracy and up to 99.89% accuracy on the MIT-BIH dataset for automated ECG arrhythmia classification.

synapsesocial.com/papers/6a6282b6044166ff3f7d8645https://doi.org/10.1016/j.medengphy.2025.104418
View Full Paper
Ask AI
Bookmark
Share