Raw impedance signals contain motion artifacts and noise. The system applies several signal processing steps:
The raw impedance signal is first processed sequentially through a 50 Hz notch filter, a median filter, and a low-pass filter to obtain a relatively clean impedance signal.
The signal is then differentiated once to obtain the ICG signal. Peak points of the ICG signal, namely the C points, are detected using the differential threshold method. Based on the C points in the ICG signal, the time windows for detecting the B and X points are determined. Within each respective time window, the detection is performed as follows.
B Point Detection Algorithm: Within the B point time window, the maximum value of the second derivative of the ICG corresponds to the B point. If the maximum value is less than 15% of the Cpeak, the zero-crossing point of the ICG signal is taken as the B point.
If no zero-crossing point is found, the point closest to 0.15 × Cpeak is used as the B point.
X Point Detection Algorithm: Within the X point time window, a local maximum of the ICG signal is first identified, which represents the O point. Then, the position with the maximum slope relative to the O point within this time window is selected as the X point. If no such point is found, the minimum value within the time window is taken as the X point.
The Cpeak is defined as (dZ/dt)max, and the time difference between the B and X points represents the LVET. Figure 6 shows the flowchart of the ICG feature point detection algorithm.
Figure 6. Flowchart of ICG feature point detection.
- LVET = X − B
- (dZ/dt)max = Cpeak
- These feed into the SV formula
This structured detection approach:
- Improves repeatability
- Handles variable heart rates
- Maintains robustness under physiological variability
Result:
Figure 7 shows the raw human impedance signals collected using the setup in Figure 5.
Figure 8 shows the ICG signals that were created after processing these impedance signals with the ICG feature‑point detection algorithm.
The important points C, B, and X—used to calculate the stroke volume (SV)—are clearly marked on the signals.
Applicable Part Numbers
| DigiKey Part Number | Manufacturer Part Number |
|---|---|
| 175-MAX30009ENA+TTR-ND,175-MAX30009ENA+TCT-ND,175-MAX30009ENA+TDKR-ND | MAX30009ENA+T |
| 505-MAX30009ENA±ND | MAX30009ENA+ |
| 175-MAX20356AEWY+TR-ND | MAX20356AEWY+ |
| 505-MAX20356AEWY+TTR-ND | MAX20356AEWY+T |
Related Article:
Why is the MAX30009 AFE ideal for bioimpedance and ICG applications?
How does the four-electrode bioimpedance measurement method work?
What is Impedance Cardiography (ICG) and how does it measure stroke volume?
How does impedance cardiography estimate stroke volume and LVET?