Long-term weight control is difficult to achieve and requires permanent changes in eating behavior. Emerging wearable sensor technology enables accurate and objective measurement of ingestive behavior, and real-time analysis of the sensor data paves the way for development of individually tailored and immediately delivered intervention (just-in-time adaptive Intervention; JITAI) to change eating behavior. Grounded in empirically and theoretically supported behavior change strategies for weight control, the proposed project relies on the synergy of wearable sensor technology, machine learning, behavioral science, personalized medicine, and nutrition to deliver and test such JITAIs. We previously developed a wearable sensor, the Automatic Ingestion Monitor (AIM), that automatically and accurately detects eating and characterizes meal microstructure (e.g., eating duration, rate of ingestion). These data can also be used to accurately estimate energy intake. The goals of this project are to: 1) use the AIM to study two common behavioral patterns observed among individuals with overweight/obesity, namely, excessive total daily energy intake (EI) and fast eating rate; 2) define the optimal personalized triggering metrics for two JITAIs targeting these behaviors; and 3) evaluate JITAIs? effects on daily energy intake and targeted behaviors. In fulfillment of these goals, we will first conduct a study to characterize the target eating behaviors, then simulate and define triggering metrics for personalized JITAIs to change targeted eating behaviors and decrease EI. The JITAIs are rooted in self-regulation theory (SRT): setting a behavioral goal and monitoring progress toward that goal, with feedback to reinforce success. To enable the SRT-informed JITAIs, we will first use the AIM to collect data about ingestive behaviors quantified by objective, sensor-measured metrics from 90 adults with overweight/obesity who will wear the device for one week in free living conditions. Second, using the collected dataset, we will: a) analyze individual curves of cumulative daily EI and rate of eating within eating episodes to define triggering parameters for personalized JITAI delivery, and b) numerically simulate JITAI delivery and effects. We will then conduct a second study to evaluate the immediate effect of JITAIs on EI and ingestive behavior in free living participants. We will conduct a within-subjects trial with 128 adults wearing the AIM for 7 weeks. To personalize JITAIs, the AIM will learn individual eating patterns over a 1-week run-in period. Each JITAI will be delivered for two weeks (weeks 2-3 and 5-6) in a randomized crossover design with the resulting daily EI and ingestive behavior compared to baseline and the acceptability of the JITAIs assessed via questionnaire. On washout weeks 4 and 7, participants will continue to wear the AIM (no JITAIs) to assess persistence of intervention effects. The proposed project is the first step in demonstrating that AIM-based JITAIs can alter a variety of eating behaviors associated with excess EI.
Achievement of changes in eating behaviors that facilitate long-term weight loss and maintenance is elusive. Emerging wearable sensor technology allows for accurate and objective measurement of ingestive behavior. Real-time analysis of the sensor data paves the way for individually tailored just-in-time adaptive interventions (JITAIs) based on empirically and theoretically supported behavior change strategies for healthy eating and weight control. The proposed project relies on synergy of wearable sensor technology, machine learning, behavioral science, personalized medicine, and nutrition to test two such JITAIs driven by the Automatic Ingestion Monitor (AIM), a device that automatically detects and characterize eating behavior in real-time. The information provided by the AIM will be used to implement and test personalized, adaptable behavioral interventions aimed at the reduction of energy intake.
