Tidal Volume Level Estimation Using Respiratory Sounds.

Area of Science:

• Biomedical Engineering
• Respiratory Physiology
• Signal Processing

Background:

• Respiratory sounds offer a noninvasive method for estimating respiratory flow and tidal volume.
• Current methods require calibration, limiting their utility in home environments.
• Accurate tidal volume estimation during sleep is crucial for diagnosing respiratory disorders.

Purpose of the Study:

• To develop and validate a novel, calibration-free respiratory sound analysis method for qualitative tidal volume level estimation during sleep.
• To differentiate between normal breathing, simple snoring, and obstructive snoring using acoustic parameters.
• To assess the correlation between estimated tidal volume levels and nocturnal oxygen saturation.

Main Methods:

• Respiratory sounds were filtered and segmented into one-minute clips.
• Agglomerative hierarchical clustering (AHC) categorized clips into normal breathing, snoring, or uncertain.
• K-means algorithm classified snoring into simple or obstructive, with tidal volume estimated based on snoring duration or breathing pause intervals.

Main Results:

• The proposed method successfully categorized respiratory sounds and estimated tidal volume levels.
• Tidal volume estimations demonstrated high accuracy and robustness when compared with lowest nocturnal oxygen saturation (LoO2) data.
• The method was validated on the open-source PSG-Audio dataset, recording simultaneous polysomnography (PSG) and tracheal sounds.

Conclusions:

• The developed respiratory sound analysis method provides a robust and accurate approach for qualitative tidal volume estimation during sleep.
• This calibration-free technique holds significant potential for convenient home-based respiratory monitoring.
• Further research can explore its application in diagnosing and managing sleep-disordered breathing.