Detection of Pathologic Heart Murmurs Using a Piezoelectric Sensor

Figures

• 

  Figure 1. The appearance of the piezoelectric sensor (PS). The upper panel shows the bottom surface of the PS. The lower panel shows its top surface, which faces the patients. The diameter of the sensor is 100 mm. The sensor is covered with an insulation material.

• 

  Figure 2. Comparison of heart sound signals and murmurs between the previous and modified PS. Simultaneous recording of ECG and the PS waveforms recorded with the previous PS in the supine (a) and prone (b) positions and the modified PS in the prone position (c). The PS waveforms were high-pass-filtered with a cut-off frequency of 70 Hz.

• 

  Figure 3. Piezoelectric sensor (PS waveforms and heart sound signals with various cut-off frequencies. (a) PS outputs filtered by high-pass filtering with various cut-off frequencies indicated by numerals. The first (S1) and second (S2) heart sounds were distinguished by ECG signals. The red waveform represents the envelope of the filtered signals. Electrocardiogram (ECG) signal was shown on the top and the blue wave form represents the original output. (b) Output of electronic stethoscope. There is no simultaneous recording ECG signal. (c)–(e) The relationship between SNRs1 (c), SNRs2 (d), and SNRsys (e) are presented. Filled and open symbols indicate the data for PS and electronic stethoscope, respectively. Each point indicates the mean and standard error of 10 consecutive heart beats. SNRsys, signal-to-noise ratio during systole; SNRs1, signal-to-noise ratio during S1; SNRs2, signal-to-noise ratio during S2

• 

  Figure 4. A representative recording using the piezoelectric sensor (PS) in a patient with congenital heart defects with systolic murmur (left) and that of a control (right). Electrocardiogram (ECG) signal (upper), PS output (middle), and PS output filtered by high-pass filtering (bottom) are presented.

• 

  Figure 5. The signal-to-noise ratio (SNRsys) of a systolic murmur in nine patients with congenital heart defects (CHDs) and nine healthy neonates. The SNRsys of systolic murmur was significantly higher in patients with CHDs than in healthy neonates (p < 0.01).

• 

  Figure 6. Scatter plot, linear regression line, and Spearman’s correlation coefficient between the signal-to-noise ratio (SNRsys) of a systolic murmur recorded by the piezoelectric sensor and that recorded by an electronic stethoscope (n = 18). Black circle indicates the SNRsys of the congenital heart defects (CHDs) group, while the white circle shows the SNRsys of the control group. The red line indicates the regression line.