Visualizing dynamic patterns of binge and purge episodes using passive sensor data

Abstract Bulimia nervosa (BN) and binge-eating disorder (BED) are similar in that both are characterized by recurrent episodes of binge-eating, but diverge in that BN but not BED includes recurrent inappropriate compensatory behaviors to limit weight gain or counteract the effects of a binge. Binge (B) and purge (P) “events” are often conceptualized and studied as a series of brief snap shots where changing circumstances increase or decrease the likelihood of these disordered behaviors. We propose that B and P episodes can be better conceptualized as changes in patterns that unfold over time and that shifts between normative and dysfunctional patterns are observable using passive sensing methodologies and may be precluded by physiological or behavioral warning signs. In this paper, we introduce conceptual and theoretical approaches to consider B and P events from this perspective. We then use two illustrative case examples from a larger parent study to demonstrate how recent technologies can be used to collect intensive longitudinal data and how such data can be used to visualize normative and dysfunctional patterns. Finally, we close with discussion of the implications of this conceptualization and the corresponding data collection and visualization methods for basic science and intervention in eating disorders.