• J Y A Foo, S J Wilson, G R Williams, M Harris, and D M Cooper.
• School of Information Technology and Electrical Engineering, University of Queensland, St Lucia, Australia. jong@itee.uq.edu.au
• Australas Phys Eng S. 2004 Dec 1; 27 (4): 165-73.

AbstractMotion artefact is a common occurrence that contaminates photoplethysmographic (PPG) measurements. To extract timing information from signals during artefact is challenging. PPG signal is very sensitive to artefacts and can be used in applications like, pulse transit time (PTT) as part of the polysomnographic studies. A correlation cancellation or signal processing approach is implemented with the adaptive cancelling filter concept and a triaxial accelerometry. PPG signals obtained from a Masimo (Reference) pulse oximeter is used as reference to compare with the reconstructed PPG signals. Different hands are used for each PPG source, one stationary while the other involves typical movements during sleep. A second Masimo pulse oximeter is used to register intensity of timing errors on commercial PPG signals. 108 PTT measurements are recorded in three different movements with PTT estimates from unprocessed PPG signals showing 35.51+/-27.42%, Masimo 50.02+/-29.40% and reconstructed 4.32+/-3.59% difference against those from the Reference PPG. The triaxial accelerometry can be used to detect the presence of artefact on PPG signals. This is useful in PTT measurements when signal contaminated with artefacts are required for further analysis, especially after and during arousals in sleep. The suggested filtering model can then reconstruct these corrupted PPG signals.

17 keywords...

hide…