AI Insight
This study developed a comprehensive inference-time augmentation (ITA) framework for improving atrial fibrillation detection from photoplethysmography (PPG) signals without requiring model retraining. The framework applies 13 different signal augmentation methods with optimized parameters during the prediction phase, tested on over 400 patients and approximately 9,800 hours of recordings. Results showed improvements in detection accuracy of up to 8.5% AUROC and 10.6% AUPRC for GPT-PPG models, with reduced false positive rates of up to 4.4%.
Why it matters
This approach provides a practical solution for improving physiological signal classification in deployed medical devices where retraining models is impractical or impossible. The model-agnostic nature means it can be applied to existing systems to enhance reliability of atrial fibrillation detection and potentially other physiological signal analysis applications.
Understand the Science
Abstract: Objective: Accurate classification of physiological signals in real-world deployments is challenged by sensor noise, motion artifacts, and distribution shifts between training and deployment data. Inference-time augmentation (ITA), which applies augmentations during inference rather than retraining, offers a simple, model-agnostic mechanism to improve robustness. However, ITA application to physiological signals has remained narrow in scope, relying on limited augmentation methods with fixed, unoptimized parameters. This work proposes a unified ITA framework to address that gap.
Approach: The framework incorporates 13 augmentation methods spanning time-domain, amplitude-domain, frequency-domain, and artifact-injection transformations, with hyperparameters optimized via Bayesian optimization. We evaluate on atrial fibrillation (AF) detection from 30-second PPG signals using GPT-PPG and ResNet across five datasets comprising more than 400 patients and ${sim}$9,800 hours of recording.
Main results: Standard ITA consistently improved AUROC (up to 8.5% for GPT-PPG and 0.7% for ResNet) and AUPRC (up to 10.6% for GPT-PPG and 0.8% for ResNet). Selective ITA further reduced average FPR by up to 4.4% (GPT-PPG) and 1.3% (ResNet) on non-AF datasets.
Significance: These findings establish ITA as a practical, model-agnostic approach for improving PPG-based AF classification reliability in deployment settings where retraining is not feasible, with broader applicability to physiological signal analysis.