Comparison of methods for adaptive removal of motion artifact

Motion artifact is a significant source of noise in an ambulatory ECG monitor and and can occur frequently during Holter recording and stress testing. We have investigated methods for removing motion artifact from ECG signals using adaptive noise removal. We have investigated both the use of an electrode/skin impedance signal and a signal from a physical sensor (Measurand Shape Sensor (TM)) mounted on the electrode for adaptively modeling and removing motion artifact. The skin/electrode signal and physical sensor signals can both be used to produce equivalent noise reduction, but, in addition to requiring the sensors, higher order adaptive filters (5/sup th/ order vs. 3/sup rd/ order) were required when the physical sensor signal was used. Guided by the results of these tests, we have designed and implemented a portable ECG recorder using 120 Hz impedance based motion artifact removal to investigate the utility of motion artifact removal for Holter and stress recording.