Micromotion Artefact Reduction of a Wrist Worn PPG Sensor Using Green Light PPG and Surface EMG

Photoplethysmography (PPG) is a noninvasive technology that uses a light source and a photodetector at the surface of the skin to monitor the volumetric variations of blood circulation for measurements of heart rate and more recently, mean arterial pressure. However, PPG signals are highly susceptible to motion artefacts. In wrist worn PPG, motion artefact reduction often uses an accelerometer as motion reference but micromotions such as finger movements are not accurately captured in an accelerometer and tend to mask with the PPG spectrum making it complicated to remove. This paper presents an approach to reduce the micromotion artefacts in a PPG signal using green light optics and surface electromyography (sEMG). The corrupted PPG signal was initially filtered using 2
^nd
order IIR Chebyshev Type I Bandpass Filter. A combination of Continuous Wavelet Transform spectral subtraction and Filtered-x Least Mean Square algorithm were then used to further correct the PPG signal. Combination thereof showed the highest correlation and degree of agreement between the recovered and resting PPG as compared to using these independently. The recovered PPG signal was evaluated by comparing it to the resting PPG using Pearson linear regression, Passing-bablok regression and Bland-Altman plots.