Examining the ability of accelerometers to differentiate low amplitude movements and the influence of signal processing

This thesis investigated the effects of different signal processing methodologies on the amplitude of preschool children’s free-living raw accelerometer data and further clarified the ability of Actigraph accelerometers to consistently measure low amplitude movements. The first set of studies compared the effects of different signal processing techniques, including filtering and composite measure approaches, on the composition of raw accelerometer data over a measurement period, and during specific activities. Through the application of different processing techniques, the signal content is significantly impacted, thereby altering the interpretation of acquired data. The second set of studies were performed in a controlled laboratory setting to determine the cross-generational comparability of different Actigraph models in their ability to differentiate low amplitude movements and how different setting parameters affect these results. Results confirmed that the monitors are able to consistently differentiate low amplitude oscillations and that the effect of chosen setting parameters significantly impacts these measures.