Distinguishing normal and abnormal tracheal breathing sounds by principal component analysis

Area of Science:

• Pulmonary Medicine
• Biomedical Engineering
• Signal Processing

Background:

• Respiratory diseases present diagnostic challenges.
• Traditional methods for analyzing breathing sounds (BS) can be subjective.
• Objective, quantitative methods are needed for improved respiratory disease diagnosis.

Purpose of the Study:

• To investigate the potential of analyzing tracheal breathing sounds (BS) using spectral analysis for differentiating between normal subjects and patients with respiratory diseases.
• To explore the application of Principal Component Analysis (PCA) for automated classification of respiratory conditions based on BS spectral features.

Main Methods:

• Tracheal breathing sounds (BS) were recorded from 10 normal subjects and 8 patients with various respiratory diseases (bronchial asthma, sarcoidosis, fibrosing lung disease, chronic bronchitis, radiation pneumonitis).
• Frequency spectra of BS were generated using Fast Fourier Transform (FFT).
• Spectral features were extracted by dividing spectra into frequency bands, and Principal Component Analysis (PCA) was employed for classification.

Main Results:

• Significant differences were observed in the frequency spectra of BS between normal subjects and patients.
• Patients exhibited significantly higher peak amplitude frequency and mean frequency compared to normal subjects.
• Principal Component Analysis (PCA) demonstrated a clear separation between normal and abnormal tracheal BS using 10, 20, and 40 spectral features.

Conclusions:

• Principal Component Analysis (PCA) of tracheal breathing sounds (BS) shows promise as a novel, automated method for diagnosing respiratory diseases.
• Spectral analysis of BS provides objective, quantifiable data that can distinguish between healthy and diseased lung states.
• This approach could lead to more efficient and accessible respiratory disease screening and diagnosis.