Key result
Attention-CNN outperforms baseline CNN and AST for classification with an AUC of ~0.97.
Why the study?
Cardiovascular diseases are a primary cause of death worldwide, driving demand for reliable, non-invasive diagnostic tools such as deep learning models for automatic heart sound classification.
Does an Attention-CNN model improve the classification of normal versus abnormal heart sounds compared to baseline CNN and AST models in phonocardiogram recordings?
Does an Attention-CNN model improve the classification of normal versus abnormal heart sounds compared to baseline CNN and AST models in phonocardiogram recordings?
Absolute Event Rate: 0.9711% vs 0.9636%
An Attention-CNN model provides a highly accurate and computationally balanced approach for automated heart sound classification, outperforming standard CNN and transformer architectures.
Background: Cardiovascular diseases remain one of the primary causes of death worldwide, thus generating an ever-continuing market demand for authentic and non-invasive diagnostic tools. The current research focuses on the automatic classification of heart sound signal recordings (phonocardiograms, PCGs) with deep learning models to facilitate detection of cardiac abnormalities. Methods: Three deep learning approaches were compared in a common experimental framework: a baseline Convolutional Neural Network (CNN), an Attention Augmented CNN (Attention-CNN), and Audio Spectrogram Transformer (AST). Experiments were performed on the PhysioNet/CinC Challenge 2016 data set using Mel-spectrogram representations. The performance of all the architectures was evaluated using a 5-fold cross-validation protocol with Area Under the Receiver Operating Characteristic Curve (AUC), Accuracy, and F1 metrics. Training used the Adam optimizer with binary cross-entropy loss, and record-level predictions were obtained by aggregating segment-level probabilities for each recording. Results: Attention-CNN got the best results with an AUC of 0.9711 and an F1-score of 0.7679, beating the baseline CNN (AUC = 0.9636) and AST (AUC = 0.9243). The use of channel attention resulted in better feature discrimination and fewer false positive predictions. While AST demonstrated the ability to capture global contextual features through self-attention, its performance was constrained by computational resources and limited training period. Conclusion: The findings suggest that Attention-CNN is a preferable balanced relationship between performance and computational requirements for PCG-based heart sound classification. However, the study has limitations in terms of data size and availability of computational resources, especially for transformer models. Future work will continue with larger scale training and validation to further test model generalizability in clinical and telemedicine practice.
No takes yet. Share an insight, caveat, or question.
Althaph et al. (2026) studied Heart sound abnormalities (n=3,126). Attention Augmented CNN (Attention-CNN) vs. Baseline CNN and Audio Spectrogram Transformer (AST) was evaluated on Area Under the Receiver Operating Characteristic Curve (AUC). Attention-CNN achieved the highest classification performance with an AUC of 0.9711, outperforming the baseline CNN (AUC 0.9636) and Audio Spectrogram Transformer (AUC 0.9243).
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: