Pedometers measure steps, but what if there were similar automatic devices that measured eating behavior? Evidence from nutrition research shows that the speed, timing, and duration of an individual's eating behavior may be linked to obesity and other health problems. has long been shown to be strongly associated with Although feeding behavior can be accurately measured in a controlled laboratory setting, a blind spot exists when researchers attempt to study how participants actually eat “in nature.”
A new National Institutes of Health-funded project by three scientists from the University of Rhode Island and the University of Texas at Austin aims to use AI-enabled wearable technology to uncover real-world eating behavior. is. The four-year, $2.4 million grant from the National Institute of Diabetes and Digestive and Kidney Diseases was awarded to URI nutrition professor Kathleen Melanson, psychology professor Theodore Walls, and University of Texas at Austin electrical and computer engineering professor Edison.・Supporting Thomas' research.
They plan to develop a system that detects detailed information about meal movements and, in some cases, every bite or chew. Researchers plan to combine more than 60 years of expertise in nutrition, behavioral statistics, and engineering to develop a new interdisciplinary project that will provide researchers with more complete data on the nutritional habits and behaviors of study participants. .
Eating behavior data collected in the lab is most accurate, but because people don't live in the lab, we don't know what they do in their daily lives in the real world. You want to compare the results of your new system to the system you already have in your lab to ensure that your data is being collected properly. The goal is to use this system in research to be able to test interventions for modifying intake behaviors. ”
Kathleen Melancon, URI Professor of Nutrition
The study uses two devices: a typical smartwatch and an unobtrusive, custom-made sensor that is placed on the participant's jawline. A smartwatch captures participants' arm and wrist movements as they perform typical eating gestures, measuring speed and frequency. This is combined with data captured by a small, button-sized sensor that records the movements of the jaw as it chews, recording the speed and intensity of the movements.
“This study, through several successive experiments from a laboratory environment to a 'natural environment,' moved from the internal validity of inferences in a laboratory-controlled environment to an inference based on external validity in the real world. It's a gradual transition,” Walls said. Its research produces statistical tools for understanding real-time, intensive, longitudinal health behavior data.
Researchers will study participants across four progressive stages. After being fitted with the discreet system, they will consume a prescribed diet measured according to standardized laboratory procedures and close supervision of researchers. The next stage will involve cafeteria-style meals, still with close assistance from researchers. The test then moves to a restaurant setting, giving participants more control over their meals. These stages more closely reflect real-world eating situations.
“That's what we're trying to answer: When can you tell when someone is eating something? It may sound like a very simple question, but in a very controlled environment. We found that it's very difficult to do that if you're not in,” Thomas said. “When we speak, our jawbone does move, but it doesn't move as rhythmically as it does when we chew food. We're trying to harness these cues with sensors and AI algorithms. Let's connect the two devices and see if we can come up with a more powerful system for detecting this kind of eating behavior.”
Finally, study participants return to their normal lives, but wear sensors to monitor their eating habits. Researchers will measure data such as eating speed, chewing speed, oral processing speed, how long food stays in the mouth before swallowing, and how quickly food disappears from the plate.
“These chewing and oral processing behaviors, such as eating too quickly, eating in large quantities, not pausing between bites, and not chewing thoroughly, can lead to people consuming too many calories before satiety signals occur. Yes,” Melancon said. “Assessing these behaviors can therefore help develop systems that can be used for interventions that help people adapt their intake behaviors to maximize satiety and aid energy intake.”
Walls added that this study will allow researchers to add other sensors, perhaps ones that monitor facial skin stretch. “We want to ensure that we can deliver that progress from the lab in a way that actually works in our overall behavioral monitoring approach. Our 'customers' will be able to use this system These are people who want to start clinical trials. This stage is exactly where it's at. We've started measuring, but after that we plan to test an actual program to help people manage their eating behavior. ”
The participants in this study will be those who are most likely to benefit from this research: those at risk of obesity-related harm. The researchers plan to recruit participants from Latino communities in both Rhode Island and Texas, allowing them to explore their unique cultural eating habits. Members of these communities, on average, have higher rates of obesity and related health problems.
sauce:
University of Rhode Island
