Classification of vascular function in upper limb using bilateral photoplethysmographic signals

Area of Science:

• Biomedical Engineering
• Cardiovascular Physiology
• Medical Diagnostics

Background:

• Diabetes mellitus is a significant risk factor for cardiovascular disease.
• Early detection of vascular dysfunction is crucial for managing diabetic complications.
• Photoplethysmography (PPG) offers a non-invasive method for assessing peripheral vascular function.

Purpose of the Study:

• To investigate the utility of bilateral PPG signals in differentiating between healthy individuals and those with diabetes.
• To explore the application of advanced signal processing techniques for vascular function assessment.

Main Methods:

• Simultaneous recording of PPG signals from index fingers of 40-50 year old healthy and diabetic subjects.
• Modeling of PPG pulses using the ARX440 model and retention of high-fitness segments.
• Application of Principal Component Analysis (PCA) to reduce dimensionality and identify key differentiating features.
• Classification of subjects using Linear Discriminant Analysis (LDA) with a leave-one-subject-out cross-validation method.

Main Results:

• The first principal component, explaining 78.2% of the variance, was significantly higher in diabetic subjects (P < 0.0001).
• The seventh principal component, explaining 0.02% of the variance, was significantly lower in diabetic subjects (P < 0.05).
• LDA achieved 71.7% sensitivity and 70.2% specificity in classifying subjects, with 81% sensitivity for males.

Conclusions:

• Bilateral PPG signal analysis, combined with PCA and LDA, can effectively distinguish between healthy and diabetic individuals.
• PPG-based vascular function assessment shows promise for early detection of vascular changes associated with diabetes.
• This non-invasive technique may aid in the early identification of individuals at risk for cardiovascular complications.