Enhancing Nutritional Ingestive Behavior Microstructure Detection: Video Annotation and Passive Sensing Approaches

Understanding the microstructure of ingestive behavior (IB) is critically important to the development and success of interventions to change eating rates and produce more optimal food energy intake outcomes. Detailed measurement of IB microstructure is needed to guide development of real-time sensing approaches that can support such interventions. This article summarizes novel measurement and inference strategies around both digital video and inertial motion sensors in a structured laboratory protocol.

Digital video footage was annotated for chews and bites and analyzed with generalized additive models to assess differences in IB for each of four meal courses varying by food texture.

Significant differences were revealed in IB microstructure in the form of nonlinear patterns of annotated video footage and initial sensing tests, indicating an optimal sensor location over the jaw's condyle bone.

Findings of an intensive longitudinal multicourse full meal protocol reflect important differences in nonlinear trends of eating behavior for diverse texture foods. These differences inform further development of technology-aided measurement strategies, provide an experimental protocol for fieldwide IB inquiry, and reveal expected fundamental differences in ingestion rates. Further inquiry into the underlying causes of nonlinearities for high UPF foods, along with sensor measurements, is warranted.