For decades, weight-loss advice has tended to treat overeating as a fairly straightforward problem: people consume more food than they need, so the solution is to help them consume less. Portion control, calorie counting and mindful eating all follow, to some extent, from that premise.
But overeating at a restaurant with friends may have very little in common with standing in the kitchen late at night, repeatedly reaching for something to eat after a stressful day. The calories still count, but the behavior that produced them — and therefore the strategy most likely to change it — may be entirely different.
That is the idea behind a study from Northwestern University published in npj Digital Medicine. Instead of relying exclusively on food diaries or bringing participants into a laboratory, researchers followed 65 adults with obesity through their everyday lives, using wearable sensors, a smartphone app and a small camera to capture not only when they ate but some of what was happening around them at the time.
The scale of the observation was unusual. Researchers collected more than 6,300 hours of footage over 657 days, producing 2,302 meal observations. Participants also reported their mood, cravings and sense of control around food, giving investigators a way to connect the mechanics of eating with the circumstances in which it occurred.
When all of that information was analyzed together, overeating did not emerge as a single, uniform behavior. Instead, researchers identified five recurring patterns.
One involved large takeout or delivery meals. Another appeared primarily during evening restaurant meals, often in social settings. A third was characterized by evening cravings, while another involved spontaneous eating driven largely by pleasure and accompanied by a diminished sense of control. The final pattern involved evening grazing associated with stress and anxiety.
The researchers gave the patterns memorable names — “Take-out Feasting,” “Evening Restaurant Reveling,” “Evening Craving,” “Uncontrolled Pleasure Eating” and “Stress-driven Evening Nibbling” — although they are not clinical diagnoses and should not be treated as five fixed types of overeaters. The study was relatively small and included only people with obesity, so the patterns will need to be replicated in broader populations.
What makes the findings interesting is not so much the five categories themselves as what separates them. Someone who overeats primarily when ordering takeout may be responding to large portions, convenience or the food environment. Someone who eats excessively at night while anxious may be dealing with an entirely different set of triggers. And a person who routinely eats more in restaurants may have few difficulties regulating intake when eating alone at home.
This distinction has practical implications because conventional dietary advice tends to concentrate on the food rather than the circumstances surrounding it. A nutrition plan can specify what someone should eat for dinner, but it may do little to explain why that person follows the plan without difficulty on most evenings and abandons it after a difficult day at work.
The technology provided an unusually clear illustration of this point. Sensors could detect physical characteristics of eating, including bites and chewing, and those signals helped researchers identify episodes of overeating. But the predictions improved substantially when information about mood, cravings and context was added.
In other words, the researchers learned more when they knew not only what the mouth and hands were doing, but what was happening to the person attached to them.
That observation fits with a broader change in obesity research. Excess energy intake remains fundamental to weight gain, but explaining why excess intake occurs requires looking beyond calories. People eat while celebrating and socializing, but also while bored, tired or anxious. They may order more food because they are hungry, or because they have reached the end of a long day and cooking has become one decision too many.
Much of that complexity has traditionally been described as a problem of self-control. The Northwestern researchers are exploring whether it might instead be possible to identify an individual's recurring pattern and intervene at the moment when that pattern typically develops.
A person who repeatedly overeats takeout, for example, might benefit most from an intervention while choosing what to order, before the food arrives. Someone whose excess eating consistently follows evening stress would probably need a different strategy, delivered at a different time. Social overeating could require yet another approach.
The researchers hope that wearable technology may eventually recognize some of these situations automatically, although that possibility remains far ahead of what the current study demonstrates. The intensive monitoring used in the research — including a specialized necklace and food-activated camera — is hardly something most people would want to wear indefinitely.
The more immediate value of the research may therefore be much simpler. It shifts attention away from overeating as a character flaw and toward overeating as a behavior with a context.
The study does not tell us that calories no longer matter, nor does it offer a technological solution to obesity. What it does suggest is that two people consuming the same excess calories may have arrived there by very different routes.
That makes the usual question — How much are you eating? — only part of the investigation.
For someone repeatedly struggling with overeating, a more revealing question may be: When does it happen, and what is usually happening around you when it does?
Understanding that pattern may prove more useful than simply being told, once again, to eat less.
Study: Shahabi F, et al. Unveiling overeating patterns within digital longitudinal data on eating behaviors and contexts. npj Digital Medicine. 2025. DOI: 10.1038/s41746-025-01698-9.
Add comment
Comments