Date of Award

May 2024

Degree Type


Degree Name

Doctor of Philosophy



First Advisor

Scott J Strath

Committee Members

Ann M Swarts, Chris C Cho, Rohit J Kate


direct observation, indirect calorimetry, physical activity, sedentary behavior, validation, wearable camera


Physical activity (PA) is a complex behavior to measure given its sporadic nature and its all-encompassing definition as bodily movement produced by skeletal muscles that result in increased energy expenditure. Sedentary behavior (SB) is another complex health behavior to measure as it is more exactly defined as sitting, reclining, or lying while having an energy expenditure ≤ 1.5 metabolic equivalents (METs). For both PA and SB, it is important to measure the dimensions of frequency, intensity, duration and type of activity being performed when associating the volume of PA and SB with different health outcomes. However, commonly used measurement tools within the field do not fully capture all dimensions of PA and SB, nor are they able to do so while providing information about behavioral intent or the domain of PA and SB being performed (i.e. Household chores, Leisure activities, Occupational activities, or Transportation). The wearable camera (WC) is a tool seeing increased use for health behavior research that has the potential to fully capture all dimensions and domains of PA and SB.The purpose of this dissertation is to establish the validity of WC still-images (IMGs) to assess all dimensions of PA and SB, as well as different domains of PA and SB. Specifically, the aims of this dissertation are to: 1) establish the convergent validity of WC IMGs to assess posture against a validated thigh-worn accelerometer; 2) establish the criterion validity of WC IMGs to assess the frequency and duration of PA and SB domains, PA and SB behaviors, and postures against video-recorded direct observation (DO); and 3) establish the criterion validity of WC IMGs to assess the frequency and duration of PA intensity compared with indirect calorimetry (IC). Results showed that WC IMG annotations were accurate and precise when compared with a thigh-worn monitor for time spent in sitting postures, but overestimated movement time and underestimated standing time. In comparison to criterion DO video, WC IMGs were accurate and equivalent on all estimates of PA and SB domains, across 10 different PA and SB activity types, for sitting posture, for standing posture within the Household domain and movement within the Leisure and Transportation domains. Finally, in comparison to IC, WC IMGs were accurate and equivalent for sedentary, moderate, low and moderate-to-vigorous PA intensity across steady-state behaviors and postures. In conclusion, WC IMGs can be used as a criterion measure for assessing the frequency and duration for different types of PA and SB. Future work is warranted to estimate point estimates of PA intensity using WC annotations as time series data within advanced statistical methods, as well as automate annotations of WC IMG data.