[Processing of impedance cardiogram differential for non-invasive cardiac function detection].

Area of Science:

• Cardiovascular Physiology
• Biomedical Signal Processing
• Medical Instrumentation

Context:

• Accurate detection of feature points in impedance cardiogram (ICG) signals is crucial for calculating hemodynamic parameters.
• Existing methods for ICG signal analysis may suffer from noise and inaccuracies.
• Thoracic bioimpedance measurements rely on precise ICG waveform interpretation.

Purpose:

• To develop and validate a novel, highly accurate method for detecting key feature points (A, B, C, X) in impedance cardiogram (ICG) signals.
• To improve the precision of hemodynamic parameter calculation derived from thoracic bioimpedance.
• To enhance the reliability of ICG signal analysis in diverse patient populations.

Summary:

• A new method combining adaptive ensemble empirical mode decomposition (EEMD) with wavelet thresholding for ICG signal denoising is proposed.
• Following denoising, a combination of difference analysis and adaptive segmentation, based on EEMD, is employed for feature point detection (A, B, C, X).
• The algorithm achieved a 99.72% accuracy rate on 30 diverse ICG signals from cardiac patients.

Impact:

• The improved accuracy in feature point detection directly enhances the precision of cardiac hemodynamic parameters calculated via thoracic bioimpedance.
• This advancement offers a more reliable tool for non-invasive cardiovascular assessment.
• Facilitates more accurate clinical decision-making based on bioimpedance data.