Accelerometer-derived physical activity is often broadly represented across a spectrum of time spent in different intensities (sedentary (SED), light physical activity (LPA), moderate physical activity (MPA), vigorous physical activity (VPA) and/or moderate-to-vigorous physical activity (MVPA). However, most studies investigating associations between physical activity and health and developmental outcomes have targeted only selected parts of this spectrum. This approach leads to a loss of information from accelerometry data and it creates at least two problems for interpretation of study results: 1) It ignores the possible influence of other intensities and 2) it increases susceptibility of residual confounding. Accordingly, associations across the whole physical activity intensity spectrum should be examined to obtain a complete picture and to facilitate improved interpretations of how physical activity relates to various outcomes.
Strong multicollinearity between intensity variables across the physical activity spectrum, represents a major limitation for common statistical methods such as ordinary least squares multiple linear regression. Thus, statistical approaches that can overcome this challenge are needed. A number of different analytic approaches are now being incorporated in the field of physical activity epidemiology, including isotemporal substitution models, compositional data analysis, and multivariate pattern analysis. These have different features and limitations.
In this project we aim to explore and develop multivariate pattern analysis as applied to accelerometry physical activity data. Multivariate pattern analysis can provide a more detailed analysis of how the entire intensity spectrum associates with health and developmental outcomes. Thus, we aim to determine the association patterns – the signature – of physical associated with diverse outcomes across the life span.
