Optimal time-window duration for computing time/frequency representations of normal phonocardiograms in dogs.

Area of Science:

• Biomedical Engineering
• Cardiovascular Physiology
• Signal Processing

Background:

• Phonocardiogram (PCG) analysis is crucial for diagnosing heart conditions.
• Accurate time-frequency representation (spectrogram) is essential for PCG interpretation.
• The choice of time-window duration significantly impacts spectrogram quality.

Purpose of the Study:

• To determine the optimal time-window duration for computing phonocardiogram spectrograms.
• To evaluate the impact of different time-window durations on signal analysis.
• To optimize the linear relationship assessment between intracardiac and thoracic PCGs.

Main Methods:

• Studied four dogs using intracardiac and thoracic phonocardiogram (PCG) measurements.
• Computed power and cross-spectrograms using Fast Fourier Transform (FFT) and a decaying sine-cosine window.
• Calculated coherence spectrograms to assess linear relationships and determine optimal time-window duration.

Main Results:

• Optimal time-window duration for PCG spectrograms identified as 16-32 ms.
• Windows <16 ms resulted in poor frequency resolution (>62.5 Hz) and spread low frequencies.
• Windows >32 ms improved frequency resolution but distorted time-domain signal properties.

Conclusions:

• A time-window duration between 16 and 32 ms is optimal for phonocardiogram spectrogram analysis.
• Deviations from this range lead to inaccurate frequency resolution and time-varying property representation.
• Optimal windowing ensures accurate coherence function amplitude between intracardiac and thoracic PCGs.