Soft, Multi-Wavelength Photoplethysmography Enables Reliable Neonatal Blood Pressure Monitoring Via Error...

Area of Science:

• Biomedical Engineering
• Neonatal Medicine
• Physiological Monitoring

Background:

• Continuous noninvasive blood pressure monitoring is crucial for neonatal care.
• Existing cuffless photoplethysmography (PPG) methods face challenges due to poorly understood acquisition errors.
• Overall error metrics can obscure performance variations under different signal quality and motion conditions.

Purpose of the Study:

• To develop a soft, multi-wavelength wearable device for complementary PPG signal acquisition.
• To create the NEO-BP clinical dataset, synchronizing PPG with invasive blood pressure in neonates.
• To analyze PPG-based blood pressure estimation errors across various acquisition and physiological conditions.

Main Methods:

• A novel soft multi-wavelength wearable with reflective and transmissive optical paths was developed.
• The NEO-BP dataset was prospectively collected from 42 neonates, including synchronized PPG and invasive arterial blood pressure.
• A green-red-infrared (G+R+IR) model was evaluated on a segment-level test set (n=9378) without quality exclusion.

Main Results:

• The G+R+IR model achieved mean absolute errors of 9.67 mmHg (systolic) and 6.17 mmHg (diastolic) blood pressure.
• Error analysis revealed increased deviations under conditions of lower signal quality and increased motion.
• A subgroup error profile was established, identifying specific error-prone acquisition states.

Conclusions:

• The developed wearable and dataset enable a deeper understanding of cuffless PPG blood pressure estimation errors.
• Quality-aware interpretation of segment-level estimates is supported by identifying error patterns.
• These findings provide a foundation for future models to predict the error risk of individual blood pressure readings.