Motion significantly affects arterial flow, violating the assumption of mutual independence and indicating that care must be taken when applying independent component analysis to wearable sensor data.
Motion artifact reduction and separation become critical when medical sensors are used in wearable monitoring scenarios. Previous research has demonstrated that independent component analysis (ICA) can be applied to pulse oximeter signals to separate photoplethysmographic (PPG) data from motion artifacts, ambient light, and other interference in low-motion environments. However, ICA assumes that all source signal component pairs are mutually independent. It is important to assess the statistical independence of the source components in PPG data, especially if ICA is to be applied in ambulatory monitoring environments, where motion artifacts can have a substantial effect on the quality of data received from light-based sensors. This paper addresses the statistical relationship between motion artifacts and PPG data by calculating the correlation coefficients between arterial volume variations and motion over a range of stationary to high-motion conditions. Analyses indicate that motion significantly affects arterial flow, so care must be taken when applying ICA to light-based sensor data acquired from wearable platforms.
Yao et al. (Sat,) conducted a other in Motion artifacts in wearable pulse oximeter signals. Independent Component Analysis (ICA) was evaluated on Correlation coefficients between arterial volume variations and motion. Motion significantly affects arterial flow, violating the assumption of mutual independence and indicating that care must be taken when applying independent component analysis to wearable sensor data.