Developing Objective Assessment of Physical Activity and Sedentary Behavior for Adults with Down Syndrome

Mississippi State University, Mississippi State MS

Investigators

Linked publications & trials

Abstract

There is an essential gap in understanding the levels and patterns of physical activity and sedentary behavior of adults with Down syndrome (DS): objective assessment of these behaviors with accelerometers has not been developed for this population. If this gap remains unfilled, it may impede efforts to objectively quantify the problem of inactivity in adults with DS and to develop effective interventions for improving physical activity and reducing health disparities in this population. Objective assessment of physical activity and sedentary behavior for adults with DS will advance by considering their pathologically low aerobic fitness. It may also improve by applying to accelerometer output innovative machine-learning modes. The long-term goals of the present applicants are to understand the determinants of physical activity and sedentariness in adults with DS and develop effective interventions for improving their physical activity and health. The objective of this application, in pursuit of the long-term goals, is to calibrate accelerometer output for measuring physical activity and sedentary behavior in adults with DS. The central hypothesis is that machine learning applied to accelerometer output will permit accurate prediction of physical activity type and intensity in adults with DS. The hypothesis has been formulated on the basis of preliminary data and previous work that has successfully developed machine learning models for individuals without disabilities. The rationale for the proposed research is that accurate measurement of physical activity and sedentariness will lead to better monitoring of these behaviors in adults with DS and to more precise quantification of the effects of interventions for improving their health. Guided by the applicants? preliminary data and previous work, and a critical review of the empirical knowledge-base, the hypothesis will be tested by pursuing the following two specific aims: (1) compare the relationship between accelerometer output and energy expenditure across various physical activities and sedentary behaviors between adults with and without DS; and (2) develop machine learning models from accelerometer output and evaluate their accuracy for predicting physical activity type and energy expenditure in adults with and without DS. To achieve Aim 1, output from hip- and wrist-worn triaxial accelerometers and energy expenditure expressed relative to aerobic fitness will be measured during physical activities and sedentary behaviors in adults with and without DS. To achieve Aim 2, machine learning models for predicting physical activity type and intensity will be developed in samples of adults with and without DS, and their prediction accuracy will be tested in two additional samples. The approach is innovative, because: (a) it will calibrate accelerometer output while accounting for aerobic fitness; (b) it will refine calibration by utilizing machine learning models. The proposed research is significant because it will allow researchers and public health agencies to more precisely monitor the levels and patterns of physical activity and sedentariness in adults with DS. Such knowledge is critical for the development of interventions for improving physical activity, health, and quality of life in adults with DS.