An ECG oximetry system for identifying obstructive and central apnoea events

Area of Science:

• Biomedical Engineering
• Medical Signal Processing
• Sleep Medicine

Background:

• Sleep apnea is a common disorder characterized by repeated breathing cessation during sleep.
• Accurate detection of different apnea types (central, obstructive) is crucial for diagnosis and treatment.
• Current methods often require manual analysis or specialized equipment, limiting accessibility.

Purpose of the Study:

• To develop and evaluate an automatic algorithm for processing simultaneous electrocardiogram (ECG) and oximetry signals.
• To accurately identify epochs of central apnea, obstructive apnea, and normal breathing.
• To assess the algorithm's performance using a validated sleep apnea database.

Main Methods:

• An algorithm was developed to process ECG and oximetry signals concurrently.
• The algorithm utilizes time and spectral features from ECG-derived heart rate and respiration.
• It also incorporates features capturing desaturations from the oximeter sensor.
• Performance was evaluated using leave-one-record-out cross-validation on the Physionet Sleep Apnea Database.

Main Results:

• The algorithm achieved 80% specificity in classifying three epoch types (central apnea, obstructive apnea, normal breathing).
• Sensitivity for pure central apnea was 44%, and for obstructive apnea was 35%.
• When central and obstructive apnea epochs were combined, the system achieved 81% sensitivity.

Conclusions:

• The developed automatic algorithm demonstrates potential for processing cardiorespiratory signals to detect sleep apnea.
• While sensitivities for individual apnea types require improvement, combining them shows higher detection rates.
• Further refinement could enhance the algorithm's clinical utility for sleep apnea screening and monitoring.