An Explainable Fusion of ECG and SpO₂-Based Models for Real-Time Sleep Apnea Detection

Area of Science:

• Biomedical Engineering
• Artificial Intelligence in Healthcare
• Sleep Medicine

Background:

• Obstructive sleep apnea (OSA) is a prevalent condition with significant health risks.
• Current diagnosis via polysomnography (PSG) is resource-intensive and patient-unfriendly.
• There is a growing need for accessible and automated OSA detection methods.

Purpose of the Study:

• To develop an explainable AI model for real-time apnea detection.
• To utilize electrocardiogram (ECG) and blood oxygen saturation (SpO2) data for enhanced accuracy.
• To provide visual explanations for AI-driven diagnostic decisions, improving clinical interpretability.

Main Methods:

• Development of an AI model integrating ECG and SpO2 signals.
• Implementation of visual explanation techniques for model transparency.
• Exploration of model fusion strategies to boost detection performance.

Main Results:

• The fused AI models demonstrated improved accuracy in detecting OSA.
• Visual explanations highlighted key signal features influencing diagnostic conclusions.
• The model offers transparent and interpretable insights into OSA detection.

Conclusions:

• The proposed explainable AI model offers a transparent and accurate approach to real-time OSA detection.
• Visual explanations enhance clinical trust and decision-making in OSA diagnosis.
• This advancement holds potential for improved patient care and early OSA management.