Noise in Blood Pressure Signal Acquisition

Pages 7-12 (read for background and context, we will be covering the frequency domain aspects later on)

Read and analyze Examples 2 and 3: pages 12-15 in DSP Applications in Medicine.pdf

Example 2 (page 12)

1. Draw a diagram showing the input signal(s), processing elements, outputs.

2. Indicate which of the figures you are looking at are in the time domain and which in the frequency domain

3. Explain why aliasing is occurring and how is it manifested. How was the exact aliased frequency computed?

4. The case study is set in the UK. If you were to repeat the same analysis in the US, where the power (mains) frequency is 60Hz, how would the values of sampling frequency required to prevent aliasing would need to change. If you were to maintain the sampling frequency at 67Hz, at what frequency would the aliased signal appear?

Example 3 (page 14,15)

5. How is the mean value calculated? How is the peak-to-peak amplitude calculated for the specific signal given in Figure 11?

6. Why is the DC component important to consider when measuring BP?

Example 2: The acquisition of a blood pressure signal

This example illustrates a blood pressure (BP) signal which contains clinically relevant components up to about 20 Hz. The example includes three figures: figure 8 is in the time domain showing the original signal while figures 9 is in the frequency domain showing the power spectrum of the signal where the power (mains) frequency is 20 Hz and sharp peak is at 50 Hz, and figure 10 is in frequency domain showing the clinically relevant frequency band of the BP signal when sampled at 67 Hz. The original BP signal is contaminated by mains noise & other noise at and below 50Hz. To remove this noise, a Lowpass (anti-alias) filter can be utilized; however, the correct sampling rate must be used. The correct sampling rate would need to be chosen above or at 100Hz because the highest frequency present in the signal should be at 50 Hz even though this is not the highest frequency present, it is the highest desired frequency which determines the minimum sampling rate required. Aliasing is occurring in this example because the sampling frequency chosen was 67 Hz. Therefore, a large aliasing peak can be seen at 17 Hz in figure 10 which is a visual of the clinically relevant frequency band of BP signal since 67 Hz (sampling frequency) minus 50 Hz (maximum frequency desired) is 17 Hz. To prevent this peak from occurring, a sampling frequency at or above 100 Hz must be chosen as this is the Nyquist rate of the signal or the low-pass (anti-alias) filter with a cut-off frequency at 20 Hz (power (mains) frequency)) could be used if the sampling rate could be reduced to a value above 40 Hz. The sampling frequency should always be greater than or equal to 2 times the maximum frequency desired/obtained from the signal as that is the definition of the Nyquist rate. The inputs, processing elements, and outputs of this example can be seen in the image below.

If the case study was set in the US as oppose the UK where it were originally set, and the same analysis was repeated where the power (mains) frequency is 60Hz the values of the sampling frequency would be required to change to prevent aliasing. If the sampling frequency remained at 67 Hz then, aliasing would occur at 7 Hz because 67 Hz (sampling frequency) minus the 60 Hz (mains frequency) is 7 Hz. The sampling frequency would need to be 2 times the largest frequency desired which is 60 Hz in the case of the US so, the Nyquist Rate would need to be 120 Hz as opposed to 100 Hz in the UK. Moreover, even if the noise is removed via an anti-alias filter at 50 Hz, the highest frequency is 60 Hz which means the sampling frequency would need to increase to capture all the clinically relevant BP signal without aliasing.

Example 3: Arterial blood pressure

This example illustrates an arterial blood pressure signal which shows the fluctuations of the pressure in an artery during the cardiac cycle. This increases when the heart contracts, and decreases as the blood drains away while the heart relaxes. There is also a 'notch' (the 'dichrotic notch'), associated with the closure of cardiac valves, as the pressure decreases. In figure 11, the mean value is calculated to be 124 mmHg using the formula see in the image below. The mean blood pressure gives the average value during the recorded period and it is the DC part of the signal; moreover, it is important to consider because it constitutes the fluctuations about the AC component of the signal (the residual when the mean is subtracted). The peak-to-peak amplitude is 74 mmHg this is calculated via subtracting the maximum value of the signal which is 174 mmHg from the minimum value of the signal which is 100 mmHg.