Cuffless hemodynamic monitoring with physics-informed machine learning models.

Area of Science:

• Biomedical Engineering
• Cardiovascular Physiology
• Wearable Technology

Background:

• Cuffless wearable devices for blood pressure (BP) monitoring are limited by methods lacking theoretical foundations, such as pulse wave analysis or pulse arrival time.
• These limitations lead to inaccuracies due to physiological and experimental confounders, reducing clinical utility.
• Continuous hemodynamic monitoring is crucial for managing cardiovascular health.

Purpose of the Study:

• To develop a smartwatch device for cuffless hemodynamic monitoring using real-time electrical bioimpedance (BioZ) sensing.
• To establish the biophysical relationship between BioZ signals and BP through advanced modeling.
• To enable accurate estimation of BP and blood velocity using a physics-informed neural network.

Main Methods:

• Development of a smartwatch with integrated electrical bioimpedance (BioZ) sensing.
• Utilizing a multiscale analytical and computational modeling framework to link BioZ and BP.
• Implementation of a signal-tagged physics-informed neural network incorporating fluid dynamics for BP and blood velocity estimation.

Main Results:

• Successfully elucidated the biophysical relationship between BioZ and BP, identifying key influencing parameters.
• Demonstrated accurate cuffless BP and blood velocity monitoring in healthy individuals and patients with cardiovascular conditions.
• Validated the approach across various conditions, including rest, physical activity, and autonomic challenges.

Conclusions:

• Electrical bioimpedance (BioZ) technology is feasible for cuffless BP and blood velocity monitoring.
• The developed smartwatch addresses critical limitations of existing cuffless wearable BP devices.
• This technology holds potential for transforming ambulatory and at-home hemodynamic assessments.