What question did this study set out to answer?

The aim is to enhance phonocardiogram signal classification for early detection of cardiovascular diseases using a deep learning framework.

April 21, 2026Open Access

MGO-CRNN: a bio-inspired deep learning framework for accurate phonocardiogram classification

Q: What is the clinical evidence from this study?

Study design: Other. Population: Cardiovascular diseases (abnormal heart sounds) (n=1568). Intervention: MGO-CRNN (Mountain Gazelle-Optimized Convolutional Recurrent Neural Network) vs. Unoptimized models and traditional tuning methods. Primary outcome: Classification accuracy (PhysioNet 2022 dataset).

Key Result

The MGO-CRNN model improved phonocardiogram classification accuracy to 99.67% on the PASCAL dataset and 99.99% on the PhysioNet 2022 dataset compared to unoptimized models.

Key Points

The aim is to enhance phonocardiogram signal classification for early detection of cardiovascular diseases using a deep learning framework.
Developed an optimized hybrid CNN-RNN model using the Mountain Gazelle Optimization algorithm for hyperparameter tuning.
Implemented onset peak detection and segmentation methods for precise heartbeat localization and analysis.
Utilized SMOTE to address class imbalance in the datasets.
MGO-CRNN model accuracy improved from 54.23% to 99.67% on the PASCAL dataset and from 99.50% to 99.99% on PhysioNet2022.
Significantly reduced classification errors compared to traditional tuning methods.
Showed effectiveness in enhancing model performance and generalization.

Structured PICO

Does the MGO-CRNN model improve the classification accuracy of phonocardiogram signals compared to traditional tuning methods?

Population

Phonocardiogram (PCG) recordings from two benchmark datasets: PASCAL and PhysioNet 2022 (5,272 PCG recordings from 1,568 subjects collected during mass screening campaigns in Brazil).

Intervention

Mountain Gazelle-Optimized Convolutional Recurrent Neural Network (MGO-CRNN) with onset peak detection and Synthetic Minority Over-sampling Technique (SMOTE)

Comparator

Traditional hyperparameter tuning methods (grid search, Bayesian optimization) and prior CNN-RNN hybrids

Outcome

Classification accuracy of phonocardiogram signals

The MGO-CRNN model demonstrates exceptionally high accuracy in classifying phonocardiogram signals, offering a promising tool for automated cardiovascular disease screening.

Main Result

Absolute Event Rate: 99.99% vs 99.5%

Abstract

Abstract Accurate classification of phonocardiogram (PCG) signals is vital for the early detection of cardiovascular diseases (CVDs), a leading cause of global mortality. However, conventional machine learning and deep learning approaches often suffer from suboptimal feature extraction, class imbalance, and inefficient hyperparameter tuning. This study proposes an optimized hybrid CNN-RNN (CRNN) model, enhanced through the Mountain Gazelle Optimization (MGO) algorithm for automated hyperparameter tuning. The model incorporates onset peak detection for precise heartbeat localization, segmentation for structured analysis, and Synthetic Minority Over-sampling Technique (SMOTE) to address class imbalance. Evaluated on two benchmark datasets—PASCAL and PhysioNet2022—the MGO-CRNN model demonstrated substantial improvements in classification accuracy. Specifically, accuracy on the PASCAL dataset improved from 54.23% to 99.67%, and on PhysioNet 2022 from 99.50% to 99.99% after optimization. The optimized model significantly reduced classification errors and outperformed traditional tuning methods such as grid search and Bayesian optimization. These results underscore the effectiveness of MGO in enhancing model performance and generalization. The integration of advanced signal processing techniques and intelligent optimization provides a robust, real-time solution for AI-assisted cardiac screening and clinical decision support.

Mark Helpful

Bookmark

Relay

View Full Paper