Analysis of the photoplethysmographic signal by means of the decomposition in principal components

Abstract

We study the plethysmographic signal using principal component analysis (PCA). By decomposing the signal using this method, we are able to regenerate it again, preserving in the process the functional relationships between the components. We have also found the relative contributions of each specific component to the signal. First return maps have been made for the series of residues of the decomposition. Further analysis using spectral methods has shown that the residues have a 1/f -like structure, which confirms the presence and conservation of this component in the signal and its relative independence with respect to the oscillating component (Hernández et al 2000 Rev. Cubana Inform. Medica 1 5). Our conclusions are that: (i) PCA is a good method to decompose the plethysmographic signal since it preserves the functional relationships in the variables, and this could be potentially useful in finding new clinically relevant indices; (ii) the 1/f process of the plethysmographic signal is preserved in the residues of the decomposed signal when PCA is used; (iii) clinically relevant parameters can potentially be obtained from photoplethysmographic signals when PCA is used.

Improving Pulse Rate Measurements during Random Motion Using a Wearable Multichannel Reflectance Photoplethysmograph

“…The primary approach to reduce the effects of motion artifact is the implementation of software-based algorithms that attempt to extract a clean PPG waveform from the motion-corrupted PPG signal [ 3 , 4 , 5 , 6 , 7 , 8 , 9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 , 18 ]. Conventional signal filtering is helpful, but not very effective when MA has no predetermined frequency band.…”

“…Alternate algorithms have been developed to extract the clean PPG waveform from the motion-corrupted signal based on fundamental components of the PPG signal. These methods include: principle component analysis (PCA) [ 13 ], independent component analysis (ICA) [ 14 , 15 , 16 ], and singular spectral analysis (SSA) [ 17 ]. Most recently, algorithms based on filtering out the motion frequency as taken from the accelerometer spectra have been useful in separating motion signal from PPG signal [ 23 , 24 ].…”

Motion robust remote photoplethysmography in CIELab color space

“…10,11 To overcome this difficulty, various creative methods have been developed, including independent component analysis 11,12 and principle component analysis. 13,14 These blind source separation algorithms help to improve the quality of rPPG, but often fail to recover rPPG signal from serious movement artifacts. Another effort is to divide the face area into many small regions, and then reduce the body movement-induced artifacts on the rPPG from the entire face area by adjusting the weighting factors for different regions.…”

A Wearable System for Real-Time Continuous Monitoring of Physical Activity

“…The automatic detection of MAs and its separation from pulse recordings is a nontrivial exercise in computer signal processing mainly due to their significant band overlapping. Numbers of solutions have been proposed ranging from moving average filters [ 38 – 40 ] to adaptive algorithms (least mean squares adaptive algorithm [ 41 – 45 ], Kalman filters [ 46 ], time-frequency methods and wavelet transform [ 47 , 48 ], principal component analysis [ 49 ]); the main pros and cons are detailed in [ 50 ]. Within others, the software used to analyze the HR from the proposed device uses the normalized least mean squares (NLMS) adaptive algorithm presented in [ 41 ].…”