New tracheal sound feature for apnoea analysis.

Area of Science:

• Biomedical Engineering
• Sleep Medicine
• Signal Processing

Background:

• Sleep apnoea syndrome is prevalent and frequently underdiagnosed in the general population.
• Accurate detection of apnoea events is crucial for timely diagnosis and management.
• Existing methods may require patient-specific adjustments due to variations in tracheal sound amplitudes.

Purpose of the Study:

• To develop and validate a novel tracheal sound feature for differentiating apnoea events from non-apnoea periods.
• To create an amplitude-independent feature for simplified application in sleep apnoea screening.
• To assess the efficacy of the new feature in separating apnoea events across multiple patients.

Main Methods:

• Development of a new feature analyzing the local spectral content of tracheal sound signals.
• Comparison of median, maximum, and mean smoothing techniques for feature extraction.
• Testing the feature on overnight recordings from 10 sleep apnoea patients (1,107 apnoea events, 72 hours of non-apnoea time).
• Ensuring the feature has a constant range (0 to 1) irrespective of signal amplitude.

Main Results:

• The developed tracheal sound feature demonstrated good separation between apnoea and non-apnoea events.
• An overall separation accuracy of 80.8% was achieved across all patients.
• The amplitude-independent nature of the feature simplifies its application without patient-specific calibration.

Conclusions:

• The novel tracheal sound feature shows significant promise for objective sleep apnoea event detection.
• Amplitude independence is a key advantage, facilitating broader clinical applicability.
• This feature offers a beneficial tool for screening and diagnosing sleep apnoea syndrome.