Identify 5 Overeating Types for Targeted Interventions

What if your smart watch could sense when you're about to raid the fridge, and gently steer you toward a healthier choice instead?

Northwestern University scientists are bringing that vision closer to reality with a lifestyle medicine program that uses three wearable sensors—a necklace, a wristband and a body camera—to capture real-world eating behavior in unprecedented detail and with respect for privacy.

The study, "Unveiling overeating patterns within digital longitudinal data on eating behaviors and contexts," was published in the journal npj Digital Medicine.

"Overeating is a major contributor to obesity, yet most treatments overlook the unconscious habits that drive it," said corresponding author Nabil Alshurafa, associate professor of behavioral medicine at Northwestern University Feinberg School of Medicine and of computer engineering at Northwestern's McCormick School of Engineering.

In a new study, 60 adults with obesity wore the three sensors and used a smartphone app to track meal-related mood and context snapshots (i.e. who they're with, what they're doing) for two weeks. The study yielded thousands of hours of video and sensor data and revealed that overeating is far from one-size-fits-all. Instead, it falls into five distinct patterns:

Take-out feasting: Gorging on delivery and take-out meals

Evening restaurant reveling: Social dinners leading to excess food intake

Evening craving: Late-night snack compulsion

Uncontrolled pleasure eating: Spontaneous, joyful binges

Stress-driven evening nibbling: Anxiety-fueled grazing

"These patterns reflect the complex dance between environment, emotion and habit," Alshurafa said. "What's amazing is now we have a roadmap for personalized interventions."

The findings lay the groundwork for a new diagnostic era in which scientists profile individuals into one of the five patterns and deploy tailored interventions. Alshurafa's team is already working with clinicians to pilot trials of personalized behavior-change programs based on these findings, he said.

"What struck me most was how overeating isn't just about willpower," said lead author Farzad Shahabi, a Ph.D. student in Alshurafa's lab.

"Using passive sensing, we were able to uncover hidden consumption patterns in people's real-world behavior that are emotional, behavioral and contextual. Seeing the patterns emerge from the data felt like turning on a light in a room we've all been stumbling through for decades.

"Our long-term vision is to move beyond one-size-fits-all solutions and toward a world in which health technology feels less like a prescription and more like a partnership."

A bodycam with a respect for privacy

During the early days of this research, Alshurafa asked Northwestern's police department to loan him a police bodycam to see how he might design a camera that captures eating behavior in the real world. He programmed the camera to only record the wearer's food-related actions to preserve bystander privacy.

Called HabitSense, the bodycam is the first patented Activity-Oriented Camera (AOC) that uses thermal sensing to trigger recording only when food enters the camera's field of view. Unlike egocentric cameras, which capture a scene from the perspective of the wearer, or broad surveillance, AOCs record activity, not the scene, which reduces privacy concerns while capturing critical data.

A necklace that records eating behaviors

In addition to HabitSense and a wrist-worn activity tracker similar to a FitBit or Apple Watch, study participants wore a necklace designed by Alshurafa and his team called NeckSense. It is the first technology to precisely and passively record multiple eating behaviors, detecting in the real world when people are eating, including how fast they chew, how many bites they take and how many times their hands move to their mouths.