Applying neural network techniques to plethysmographic pulse shape analysis

Assessment of bilateral photoplethysmography for lower limb peripheral vascular occlusive disease using color relation analysis classifier

This paper proposes the assessment of bilateral photoplethysmography (PPG) for lower limb peripheral vascular occlusive disease (PVOD) using a color relation analysis (CRA) classifier. PPG signals are non-invasively recorded from the right and left sides at the big toe sites. With the time-domain technique, the right-to-left side difference is studied by comparing the subject's PPG data. The absolute bilateral differences construct various diminishing and damping patterns. These difference patterns in amplitude and shape distortion relate to the grades of PVOD, including the normal condition, lower-grade disease, and higher-grade disease. A CRA classifier is used to recognize the various patterns for PVOD assessment. Its concept is derived from the HSV color model and uses the hue, saturation, and value to depict the disease grades using the natural primary colors of red, green, and blue. PPG signals are obtained from 21 subjects aged 24-65 years using an optical measurement technique. The proposed CRA classifier is tested using the physiological measurements, and the tests reveal its practicality for monitoring PPG signals.

Relationship between blood pressure and finger photoplethysmographic waveform during oxygen desaturation test: a model fitting approach

A modified 3-element windkessel model was applied to study the relationship between brachial arterial blood pressure and the photoplethysmographic waveform from pulse oximeters. Data were recorded from 12 healthy volunteers who underwent the oxygen desaturation study. During about 30 minutes recording period, the SpO2 value was regulated down till about 70%. After preprocessing, singular value decomposition (SVD) algorithm was then used to get the best fit of the model parameters.

RESULT

the fitting error (RMSE) was 1.07 +/- 0.48 mmHg. The time constant of the model shown significant difference between the highest and the lowest saturation group.