Food Noise Is Real — And Science Just Measured It

For years, people on GLP-1 medications have described the same relief in almost the same words.

The noise stopped.

"Food noise" was a patient word first — something people felt, named for each other in online groups, and couldn't point to on any test. Clinicians heard it so often it became shorthand. But it lived in self-report, which is a polite way of saying some people doubted it was real.

This autumn, a brain scanner went looking for it. It found something.

The short version

At EASD 2026 — the European Association for the Study of Diabetes annual meeting, held in Milan at the end of September — Novo Nordisk presented a study that put people in an fMRI scanner and showed them pictures of high-calorie food. After about a year on CagriSema, an investigational drug that pairs semaglutide with a second appetite hormone, the brain's response to those food cues looked different — in regions tied to reward, sensory processing, and self-control. The same people reported less hunger, less craving, and less of the mental chatter about food.

The thing patients had been describing for years, in other words, showed up on a scan.

First, what "food noise" actually is

Food noise isn't hunger. That's the part that's easy to miss.

Hunger is a physical signal with an off switch — you eat, it quiets. Food noise is the commentary: the intrusive, repetitive, unwanted thoughts about food that keep running whether or not your stomach is empty. Researchers who have tried to pin the term down describe exactly that — persistent, dysphoric food-related thoughts that feel more like rumination than appetite, and that are distinct from ordinary hunger.

If you've never had a loud version of it, it's hard to explain. If you have, you don't need it explained. That gap is a big part of why the experience went unmeasured for so long.

What the study actually measured

fMRI — functional MRI — doesn't read thoughts. It tracks blood flow in the brain as a stand-in for which regions are working harder. Show someone a photo of a doughnut and watch which circuits light up, and you get an objective-ish window onto a subjective pull.

That's the design here: a randomized, placebo-controlled study that scanned adults with overweight or obesity while they viewed high-calorie food images, over roughly 52 weeks. The researchers weren't asking people how they felt and taking their word for it. They were watching the brain's food-cue reactivity directly, and comparing drug to placebo.

Why "changed," not "switched off," matters

It would be neater to say the drug turns food noise down like a dial. The honest version is more interesting: food-cue reactivity was altered across those regions, not uniformly suppressed everywhere. Reward, sensory, and behavioural-control circuits don't all move in one direction, and reading an fMRI as "less activity = less wanting" is the kind of shortcut that gets science journalism in trouble.

What's robust is the pairing. People felt the noise quiet, and their measured brain response to food looked different from placebo. When the story you tell yourself and the signal in the scanner agree, that's a stronger result than either one alone.

(For how a gut hormone reaches the brain's appetite and reward circuitry in the first place, our piece on the gut-brain conversation walks through the anatomy.)

What this does — and doesn't — mean for you

The validation is the real takeaway. People who were told, implicitly, that food noise was a matter of willpower now have a brain-imaging study describing the thing they've been reporting. That matters for how seriously the experience gets taken, and for the people who felt a little bit crazy for noticing it.

A few honest limits, though:

• CagriSema is investigational. As of this writing it isn't an approved medication, and nothing here is a suggestion to seek it, switch to it, or ask for it. What to take, and whether to take anything, is a conversation for you and your prescriber — full stop.
• This is one study, early, and presented at a conference rather than the final word in a peer-reviewed journal. Food-cue fMRI is a research tool, not a test you'll get at a check-up.
• A changed scan isn't a promise about any one person. Group averages in a trial don't tell you how your own head will feel.

The bottom line

For years, the most reliable evidence that food noise existed was that so many people described it the same way, unprompted. That was always good evidence. It's just that "trust me, the noise is gone" doesn't persuade a skeptic.

Now there's a scanner picture to go with the testimony. It doesn't change anything for the people who had already lived it — but it moves food noise from a thing patients say to a thing science can point at. That's a quietly big deal, and it's worth saying plainly: the people describing this were right.

If you want the human side of what happens after the noise stops, read what else changes when the food noise stops. For the biology underneath it, see how GLP-1s join the gut-brain conversation.