Food Noise: The First U.S. Prevalence Data, How Researchers Measure It, and the Trial Evidence on GLP-1 Drugs

"Food noise" is a popular name for thoughts about food that keep coming back and feel unwanted, whether or not you are hungry. The term spread on TikTok and in patient forums, mostly from people who said a GLP-1 drug such as semaglutide (Ozempic, Wegovy) or tirzepatide (Mounjaro, Zepbound) made those thoughts go quiet.

Research has started to catch up. The first nationally representative U.S. survey, published by Penn State researchers in 2026, found that 17.98% of adults said they experience food noise often or always. Controlled trials did not measure "food noise" by that name, but they did measure food cravings and control of eating, and semaglutide and tirzepatide improved both compared with placebo.

Three cautions matter. Scientists have not agreed on a definition, so prevalence estimates range from about 4% to 18% depending on how the question is asked. In a 60-week semaglutide trial, the drop in food preoccupation seen at week 20 was gone by weeks 40 and 60, even though people kept eating less. And no published study has tracked food noise after people stop a GLP-1 drug.

Penn State nutrition researchers Daisuke Hayashi and Travis Masterson have written several of the early papers on the topic. Masterson said the term came from conversations with people who used GLP-1 drugs and noticed a sudden drop in intrusive thoughts about food. In a 2023 review, Hayashi, Masterson, and colleagues noted anecdotal reports that patients on semaglutide experienced "less rumination and obsessive preoccupation about food."

The same team later studied how the idea spreads online. They analyzed 99 of the top TikTok videos under the hashtag #FoodNoise, collected in June 2024, when the hashtag had more than 3,600 videos. The videos averaged about 1.17 million views. About 71% were patient testimonies, 49% mentioned medications (mostly GLP-1 drugs), and 86% described food noise as negative. Only 22% of the creators were health professionals.

So the term started as a patient description of a drug effect. That history matters, because it means most early reports come from people who were already pleased with their medication.

There is no single scientific definition yet. Two research groups have proposed one each:

• A group led by Emily Dhurandhar and David Allison at Indiana University defined food noise in 2025 as "persistent thoughts about food that are perceived by the individual as being unwanted and/or dysphoric and may cause harm." In their view, intensity and intrusiveness set it apart from routine food thoughts. They compared it to rumination.
• The Penn State group uses a definition built on food cue reactivity: "heightened and/or persistent manifestations of food cue reactivity, often leading to food-related intrusive thoughts and maladaptive eating behaviors." Food cue reactivity is how the body and brain respond to the sight, smell, or thought of food.

Food noise overlaps with ideas that eating researchers have studied for decades. These include food cravings, preoccupation with food, sensitivity to food rewards, and loss of control over eating. A review in the journal Appetite, published online in September 2026 by researchers at Macquarie University in Sydney and American University in Washington laid out five possible explanations: hunger signals from the body, food cue reactivity, food addiction, the backfire effect of trying not to think about food, and a weaker ability to block out food memories. The authors said little is known about what food noise is or what causes it.

A second 2026 commentary in Appetite, by an international group that included Hayashi and Masterson, raised a different concern. It asked whether food noise is truly a separate condition or a new label for familiar experiences. It also warned that the term could turn ordinary thoughts about food into something treated as a disease.

Two questionnaires now exist.

The Food Noise Questionnaire (FNQ) was developed at Pennington Biomedical Research Center with researchers from WW International (formerly Weight Watchers) and published in the journal Obesity in 2025. It has five items. In 396 adults, it showed high internal consistency and gave similar scores when people took it again about a week later. Scores lined up closely with two existing measures of food preoccupation, which supports its validity but also shows how much it overlaps with older tools. Several authors were WW employees or shareholders, and the developers said its clinical usefulness still needs testing. An Italian version tested in 1,087 adults in 2026 showed similar measurement properties.

The Indiana University group built a separate tool, the Ro Allison Indiana Dhurandhar Food Noise Inventory (RAID-FN Inventory), and published validation work in 2025 and 2026.

The Penn State prevalence survey used neither tool. It asked a single question: "How often do you experience 'Food Noise'?" Answers ranged from 1 (never) to 5 (always). A single question is quick, but it relies on each person's own idea of what food noise means.

The Penn State survey, published in Obesity Science and Practice, added the food noise question to an online survey of 1,000 U.S. adults weighted to match the national population. It was run in May 2026.

• 17.98% said they experience food noise often or always.
• 50.89% said they experience it at least occasionally.
• Frequent food noise was more common in women (22.70%) than men (12.68%), in younger adults (26.50%), and in Asian respondents (32.22%). The Asian group had only about 50 people, and Hayashi said "there is no existing data or theory to support this result."
• It was reported by 56.77% of people who described losing control over their eating.

In a model that adjusted for other factors, the strongest link was loss of control over eating (odds ratio 5.95). Current GLP-1 use (odds ratio 1.60) and female sex (odds ratio 1.47) were also linked. An odds ratio compares the odds of an outcome between groups; it is not a percentage.

Penn State's news release said "over 80%" of respondents experienced food noise at least some of the time. That figure does not match the paper's abstract, which reports 50.89% at least occasionally. This article uses the published numbers.

A second national estimate, posted as a preprint in October 2026 and not yet peer reviewed, used the FNQ in 3,646 U.S. adults surveyed in fall 2025. Depending on the cutoff used, about 4% to 9.5% of adults met criteria for food noise, rising to about 6% to 14% among adults with obesity. The gap between that estimate and Penn State's 18% shows how much the answer depends on the definition and the question.

In the Penn State survey, frequent food noise was reported by 25.76% of current GLP-1 users and 26.78% of former users, compared with 16.30% of people who had never used one. That looks backward if the drugs quiet food noise.

A one-time survey cannot show cause and effect. Two explanations fit the data without contradicting the trials. First, people with heavy food noise may be the ones most likely to seek a GLP-1 drug, so users start with more of it. Second, users may notice it more. Masterson said the result "could arise from an awareness issue," adding: "If GLP-1s reduce food noise, those who use GLP-1s might be hyperaware of any food noise they do experience." Many users also learned the term because of the drug.

The survey did not record how long people had used the drugs, which drug they took, or their dose, so it cannot answer whether food noise falls during treatment. Trials are better suited to that question.

No large randomized trial has used a food noise questionnaire as its main outcome. Several placebo-controlled studies measured closely related things, such as cravings and control of eating, using the Control of Eating Questionnaire.

• Semaglutide 2.4 mg, 20 weeks. In a Novo Nordisk trial of 72 adults with obesity, people on semaglutide ate 35% less at an all-you-can-eat lunch than people on placebo. They also reported better control of eating and fewer and weaker food cravings. Weight fell 9.9% on semaglutide and 0.4% on placebo.
• Semaglutide 2.4 mg, 2 years (STEP 5). In 174 participants who completed the questionnaire, semaglutide improved craving control and cravings for savory foods at weeks 20, 52, and 104 compared with placebo. Improvements in cravings for sweet foods were reported at weeks 20 and 52.
• Oral semaglutide 50 mg, 20 weeks. In 61 adults with obesity, the daily pill reduced energy intake at lunch and was linked to fewer cravings and better control of eating than placebo. The 50 mg dose was a study dose; it is not the dose of the approved Wegovy pill.
• Tirzepatide, 6 weeks. In an Eli Lilly phase 1 trial of 114 adults without diabetes, tirzepatide cut lunch intake by about 525 calories compared with placebo at week 3. It also reduced food cravings, overall appetite, the tendency to overeat, and reactions to food in the environment. Brain scans showed less activity in some reward-related areas when people viewed high-fat, high-sugar food photos.

One survey used the FNQ directly. In Novo Nordisk's INFORM survey of 550 U.S. adults taking injectable semaglutide, the median FNQ score fell from 13 to 6 out of 20. Agreement with statements describing food noise dropped from 47% to 63% before treatment to 15% to 20% after. But people rated their food noise before treatment from memory, there was no comparison group, and the survey was run by Novo Nordisk with a paid market research firm.

In September 2026, Novo Nordisk said a 52-week brain imaging study of its investigational drug CagriSema showed reduced "food noise," cravings, and food-related thoughts at 22 and 52 weeks. The company's release gave no food noise scores or sample size.

The best long-term evidence suggests that some of the effect on thoughts and feelings about food wears off, even though people keep eating less.

In a 60-week trial at the University of Pennsylvania, 120 adults were randomly assigned to semaglutide 2.4 mg or placebo, published in the American Journal of Clinical Nutrition in 2026. At week 20, the semaglutide group had larger drops in past-week hunger and food preoccupation. By weeks 40 and 60, the two groups no longer differed on appetite measures. Yet the semaglutide group still ate about 240 to 290 fewer calories at a test lunch at weeks 20, 40, and 60. The authors concluded that eating less remains a key way the drug works, "even though some subjective benefits decline."

A case report in Nature Medicine in 2025 gave an unusual look inside the brain. University of Pennsylvania researchers recorded electrical activity in the nucleus accumbens, a reward center, of a 60-year-old woman with an implanted brain device for treatment-resistant loss-of-control eating. She had regained weight after gastric bypass surgery and was taking tirzepatide for type 2 diabetes. During months 2 to 4 after the implant, which followed a tirzepatide dose increase, she had few episodes of severe food preoccupation. During months 5 to 7, a specific low-frequency brain signal returned, followed by "breakthroughs" of severe food preoccupation despite being on the maximum dose. It is one person, so it cannot show how common this pattern is.

Other evidence is mixed. In a small 24-week study of 28 adults starting semaglutide or tirzepatide without nutrition support, overall food cravings did not change, though people with the highest cravings at the start had the largest drops. In interviews with 31 patients taking obesity medications, University College Dublin researchers found that people described reduced food noise during active weight loss; some in the later, stable phase felt the drug was working less well and considered switching.

No published study has measured food noise before and after people stop a GLP-1 drug. Claims that it "comes roaring back" rest on personal accounts, not data.

What is known is that weight tends to return. In the STEP 1 extension, people who had lost 17.3% of their body weight on semaglutide regained 11.6 percentage points within a year of stopping, about two-thirds of what they had lost. Appetite returning is a likely part of that, but the extension did not measure cravings or food noise.

For more on staying on treatment or stopping safely, see How to Stay Off GLP-1s: The 2026 Maintenance Playbook and Why So Many People Quit GLP-1 Drugs.

Food noise is not only a drug question. A 2026 study in Obesity surveyed 60 adults whose weight had stabilized about 13 months after bariatric surgery. They had lost an average of 24.8% of their body weight at their lowest point. People who scored higher on weight self-stigma, meaning they judged themselves negatively because of their weight, reported more food noise on the FNQ. The study cannot show which came first, and the authors called for research that follows patients over time.

Constant thoughts about food can be a sign of an eating disorder, especially when they come with episodes of eating that feel out of control. The National Institute of Mental Health describes binge-eating disorder this way: "People with binge-eating disorder regularly lose control of their eating and eat unusually large amounts of food." Signs include eating very large amounts in a short time, such as within two hours, and feeling distressed, ashamed, or guilty about eating. In the Penn State survey, loss of control over eating was the factor most strongly linked to frequent food noise.

NIMH lists psychotherapy and medication among the treatments for binge-eating disorder. If food thoughts are taking over your day, or you often eat past the point of comfort and feel unable to stop, talk with a doctor or mental health professional before deciding that a weight-loss drug is the answer. GLP-1 drugs are not FDA-approved to treat binge-eating disorder.

If food noise has changed sharply after starting or stopping a GLP-1 drug, that is worth raising with your prescriber, along with any changes in mood. See GLP-1 Drugs and Mood for that evidence.