Physiowise: A Physics-aware Approach to Dicrotic Notch Identification.

Area of Science:

• Cardiovascular Physiology
• Biomedical Engineering
• Signal Processing

Background:

• The dicrotic notch (DN) in arterial blood pressure (ABP) waveforms is crucial for assessing cardiovascular function.
• Aging and pathological vascular stiffness reduce DN prominence, complicating automated identification.
• ABP waveform deformations from noise or hemodynamic instability further challenge DN detection.

Purpose of the Study:

• To develop a robust and generalizable method for automatic dicrotic notch identification.
• To improve the accuracy of DN detection in challenging ABP waveforms, including those with deformations.
• To address the limitations of current methods in aging and pathological conditions.

Main Methods:

• A physics-aware approach, Physiowise (PW), was developed.
• The method uses a cardiovascular model to augment ABP waveforms, reducing deformations.
• Predefined rules are applied to the augmented signal for DN localization.
• A hybrid methodology allowing user-defined rules was also proposed.

Main Results:

• Physiowise (PW) demonstrated a 52% overall mean error improvement.
• Detection accuracy increased by 16% within the critical 30ms error range.
• The method was validated on in-vivo data from pigs under hemorrhage and sepsis.

Conclusions:

• The proposed Physiowise (PW) method effectively enhances dicrotic notch identification in ABP waveforms.
• This approach offers improved accuracy and robustness, particularly in cases of vascular stiffness and waveform deformation.
• The hybrid option provides flexibility for integrating with existing or custom detection algorithms.