QualityFM: a Multimodal Physiological Signal Foundation Model with Self-Distillation for Signal Quality Challenges in Critically Ill Patients

2025-09-08

AI Summary

Poor photoplethysmography (PPG) and electrocardiography (ECG) signal quality in intensive care units (ICUs) and operating rooms leads to false alarms and diagnostic inaccuracies. Existing methods suffer from weak generalizability, heavy reliance on large-scale labeled data, and limited cross-task transferability. This paper introduces QualityFM, a multimodal foundation model for signal quality assessment. QualityFM features a novel dual-path architecture integrating self-distillation with windowed sparse attention, coupled with a joint power-spectrum and phase-spectrum reconstruction loss—enabling unsupervised signal quality modeling without requiring extensive annotations. It enhances frequency-domain fidelity and long-sequence modeling capability. Pretrained on 21 million waveform segments across three model scales, QualityFM achieves state-of-the-art performance in ventricular tachycardia false-alarm suppression, atrial fibrillation detection, and cuffless blood pressure estimation—demonstrating superior generalizability and cross-task adaptability.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Computer Vision: Multi-modal Vision
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• Security and Privacy: Large-scale security measurements
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Photoplethysmogram (PPG) and electrocardiogram (ECG) are commonly recorded in intesive care unit (ICU) and operating room (OR). However, the high incidence of poor, incomplete, and inconsistent signal quality, can lead to false alarms or diagnostic inaccuracies. The methods explored so far suffer from limited generalizability, reliance on extensive labeled data, and poor cross-task transferability. To overcome these challenges, we introduce QualityFM, a novel multimodal foundation model for these physiological signals, designed to acquire a general-purpose understanding of signal quality. Our model is pre-trained on an large-scale dataset comprising over 21 million 30-second waveforms and 179,757 hours of data. Our approach involves a dual-track architecture that processes paired physiological signals of differing quality, leveraging a self-distillation strategy where an encoder for high-quality signals is used to guide the training of an encoder for low-quality signals. To efficiently handle long sequential signals and capture essential local quasi-periodic patterns, we integrate a windowed sparse attention mechanism within our Transformer-based model. Furthermore, a composite loss function, which combines direct distillation loss on encoder outputs with indirect reconstruction loss based on power and phase spectra, ensures the preservation of frequency-domain characteristics of the signals. We pre-train three models with varying parameter counts (9.6 M to 319 M) and demonstrate their efficacy and practical value through transfer learning on three distinct clinical tasks: false alarm of ventricular tachycardia detection, the identification of atrial fibrillation and the estimation of arterial blood pressure (ABP) from PPG and ECG signals.

Problem

• Addressing poor signal quality in ICU and OR physiological monitoring
• Overcoming limited generalizability and cross-task transferability in signal analysis
• Reducing false alarms and diagnostic inaccuracies in critical care settings

Innovation

• Multimodal foundation model for physiological signals
• Self-distillation strategy between quality-differentiated encoders
• Windowed sparse attention in Transformer architecture