CRIB: A Novel Method for Device-Based Physical Behavior Analysis

Physical behaviors (e.g., sleep, sedentary behavior, and physical activity) often occur in sustained bouts that are punctuated with brief interruptions. To detect classify these interrupted bouts, researchers commonly use wearable devices specialized algorithms. Most algorithms examine the data chronological order, initiating terminating whenever specific criteria met. Consequently, may encapsulate or overlap later periods also meet activation termination (i.e., alternative bout solutions). In some cases, it is desirable to compare solutions before making a final classification. Thus, comparison-focused needed, which can be used isolation concert their chronology-focused counterparts. this technical note, we present algorithm called CRIB (Clustered Recognition of Interrupted Bouts). It uses agglomerative hierarchical clustering facilitate comparison different solutions, classification being made favor smallest number comply user-specified limits on number, individual duration, cumulative duration interruptions). For demonstration, assess moderate vigorous activity accelerometer from National Health Nutrition Examination Survey, include against results two established Our discussion explores strengths limitations CRIB, as well potential considerations applications for using future studies. An online vignette ( https://github.com/paulhibbing/PBpatterns/blob/main/vignettes/CRIB.pdf ) available assist users implementing R.