Acoustic breath-phase detection using tracheal breath sounds

Area of Science:

• Biomedical Engineering
• Respiratory Physiology
• Acoustic Signal Processing

Background:

• Current breathing flow estimation relies on tracheal sounds, but breath phase detection is often inaccurate.
• Existing methods assume alternating breath phases or use a second lung sound channel, which is unreliable or labor-intensive.

Purpose of the Study:

• To develop an automatic and accurate breath-phase detection method using only tracheal sounds.
• To overcome limitations of current methods, such as unreliability in long recordings and the need for additional equipment.

Main Methods:

• A novel method analyzing breath sound parameters from tracheal sounds was developed.
• The method utilizes sound envelope features (duration, volume, shape) for phase identification.
• It requires only pre- and post-breath sound segments and is independent of flow level.

Main Results:

• The proposed method achieved 95.6% accuracy, 95.5% sensitivity, and 95.6% specificity.
• Tested on 93 healthy individuals across 4 flow levels.
• Successfully identified breath phases without assuming alternation or using extra channels.

Conclusions:

• The developed method provides an accurate and automatic solution for breath-phase detection using tracheal sounds.
• This technique enhances respiratory monitoring by eliminating the need for secondary channels or phase alternation assumptions.
• Offers a more reliable and less labor-intensive approach for breathing analysis.