Fig. 1

Objective measures of physical activity. Objective measures can use closed-source devices (Apple Watch, Fitbit, etc.) or open-source accelerometers (Actigraph, Axivity, etc.). Raw data is combined into a vector sum and then activity is categorized by intensity, using cut points for activity counts (a clinical measure of the sum of activity per period of time) or by machine learning to categorize activity type or intensity using select features of the accelerometer data