H-LDM: Hierarchical Latent Diffusion Models for Controllable and Interpretable PCG Synthesis from Clinical Metadata

2025-11-18

AI Summary

To address the scarcity of high-quality, expert-annotated phonocardiogram (PCG) data for cardiovascular disease diagnosis, this paper proposes a controllable and interpretable PCG generation framework grounded in structured clinical metadata. Methodologically, we design a hierarchical latent diffusion model integrated with a multi-scale variational autoencoder to disentangle representations of cardiac rhythm, heart sounds, and murmurs; additionally, a medical attention mechanism is introduced to enhance clinical interpretability in text-to-signal generation. Evaluated on the PhysioNet/CirCor dataset, our model achieves a Fréchet Audio Distance of 9.7, a 92% attribute disentanglement rate, and 87.1% clinical validity—yielding a +11.3% improvement in rare-disease classification accuracy. To the best of our knowledge, this is the first work to enable high-fidelity, disentangled PCG synthesis conditioned on clinical metadata, establishing a novel paradigm for few-shot cardiovascular AI diagnosis.

Technology Category

Natural Language Processing: Code Generation / Program Synthesis from Natural LanguageMachine Learning: Deep Generative Models & AutoencodersComputer Vision: Diffusion Models for Vision

Application Category

Semantics and Knowledge: Methods, algorithms and applications for the development of semantic models, knowledge graphs and other forms of structured data models with machine-interpretable semanticsUser Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalizationEconomics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

Phonocardiogram (PCG) analysis is vital for cardiovascular disease diagnosis, yet the scarcity of labeled pathological data hinders the capability of AI systems. To bridge this, we introduce H-LDM, a Hierarchical Latent Diffusion Model for generating clinically accurate and controllable PCG signals from structured metadata. Our approach features: (1) a multi-scale VAE that learns a physiologically-disentangled latent space, separating rhythm, heart sounds, and murmurs; (2) a hierarchical text-to-biosignal pipeline that leverages rich clinical metadata for fine-grained control over 17 distinct conditions; and (3) an interpretable diffusion process guided by a novel Medical Attention module. Experiments on the PhysioNet CirCor dataset demonstrate state-of-the-art performance, achieving a Fréchet Audio Distance of 9.7, a 92% attribute disentanglement score, and 87.1% clinical validity confirmed by cardiologists. Augmenting diagnostic models with our synthetic data improves the accuracy of rare disease classification by 11.3%. H-LDM establishes a new direction for data augmentation in cardiac diagnostics, bridging data scarcity with interpretable clinical insights.

Problem

• Generates clinically accurate PCG signals from structured metadata
• Enables fine-grained control over 17 distinct cardiac conditions
• Addresses data scarcity in cardiovascular disease diagnosis

Innovation

• Hierarchical Latent Diffusion Model for PCG synthesis
• Multi-scale VAE with physiologically-disentangled latent space
• Medical Attention module guiding interpretable diffusion process