Estimating Correlation between Arterial Blood Pressure and Photoplethysmograph

Photoplethysmograph (PPG) and Arterial Blood Pressure (ABP) are good indicators of cardiovascular performance. Although ABP is more widely employed, the invasive procedure of signal acquisition may cause skin rashes and inconvenience to the patients. Also, it does not allow continuous monitoring of cardiac activity. PPG, on the other hand, uses infrared light to measure the blood volume changes, which is a simple, non-invasive and can be used for continuous measurement. This paper focuses on analyzing the similarity between ABP and PPG using various features like average slope, peak position, time period, elasticity, amplitude of the signal. A segmentation algorithm was used to segment out cycles of ABP and PPG from physionet database taken from 19 patients with respiratory failure and the values of each feature were extracted for each person. Considering the population, using Pearson's correlation coefficient, the coefficient for the average slope of the PPG and peak to peak amplitude of ABP was found to be 0.55 indicating that other factors such as vessel diameter, thickness must be considered. The upstroke time period of both ABP and PPG was found to have a small difference in the range of 0.02s to 0.1s, whereas the time period of the heart cycles remained the same irrespective of the disease or healthy condition. The peak value of both ABP and PPG was found to occur with constant time difference. The elasticity with peak to peak amplitude of ABP was found to have a correlation of 0.822, and with systolic blood pressure, a correlation of 0.7622. When considered for individuals, parameters like the diastolic average slope of PPG and systolic blood pressure were found to have a good correlation coefficient ranging from 0.6 to 0.96 among other parameters which include systolic average slope, maximum and minimum slope of PPG, and the diastolic blood pressure.