Robust Arm Impedocardiography Signal Quality Enhancement Using Recursive Signal Averaging and Multi-Stage Wavelet...

Area of Science:

• Biomedical Engineering
• Physiological Measurement

Background:

• Impedance cardiography (ICG) is a non-invasive method for assessing hemodynamic parameters like cardiac output and stroke volume (SV).
• Conventional ICG uses thorax-based recordings, limiting its use in ambulatory settings.
• Upper-arm brachial artery ICG offers potential for convenient, wearable sensor devices for long-term health monitoring.

Purpose of the Study:

• To investigate the efficacy of wavelet-based denoising and recursive signal averaging for enhancing upper-arm ICG signals.
• To evaluate the accuracy and precision of beat-by-beat (BbyB) ICG waveform feature metrics derived from arm recordings compared to thorax recordings.
• To characterize the functional relationship and analyze linear regression trends between arm and thorax ICG metrics.

Main Methods:

• Collected upper arm and thorax ICG data from 15 healthy subjects.
• Applied a third-order Savitzky-Golay FIR filter for prefiltering raw ICG signals.
• Utilized a multi-stage wavelet-based denoising strategy with recursive signal-averaging for Arm-ICG signal enhancement.
• Evaluated denoising performance using a 700 ms frame from a 600-beat ensemble-averaged signal as the gold standard.

Main Results:

• The denoising strategy achieved a beat inclusion rate (BIR%) of 80.9% for Arm-ICG and 100% for Thorax-ICG (correlation > 0.95).
• Error rates (ER%) for Arm-ICG waveform features (A, B, C, VET) were 0.83%, 11.1%, 3.99%, and 5.2%, respectively.
• Recursive averaging with a 36-beat buffer yielded the best Arm-ICG BbyB denoising, with < 3.3% error in time metrics.
• Linear regression between arm and thorax stroke volume (SV) showed a coefficient of determination (R² ) of 0.84.

Conclusions:

• Wavelet-based denoising and recursive averaging effectively enhance upper-arm ICG signal quality for hemodynamic monitoring.
• Upper-arm ICG provides a viable, albeit less precise, alternative to thorax-based ICG for certain ambulatory applications.
• The study establishes a characterized functional relationship between arm and thorax ICG metrics, supporting wearable sensor development.