# Why the Appetite Comes Back

By Nora Bissett, Pricing Editor (https://glpwatchdog.org/authors/nora-bissett) | Published: 2026-08-20 | Updated: 2026-10-30

Canonical page: https://glpwatchdog.org/research/why-the-appetite-comes-back

> Weight loss slows and appetite for particular foods creeps back. A Science study found a specific brain circuit doing exactly that, in mice, whose neurons can be switched off with light.

## Key points

- IN MICE. Optogenetics and photometry cannot be done in people.
- Dopamine neurons encoding palatability were suppressed by semaglutide, then RECOVERED during repeated treatment.
- Inhibiting those neurons at the moment of eating reversed the recovery. That is the causal step.
- A mechanism for the plateau that is not 'the drug stopped working'.
- Nothing actionable follows, and inventing a takeaway would be the failure.

Weight loss on these drugs slows, and many people describe their appetite for particular foods creeping back while still taking them. A study in *Science* found a specific brain circuit doing exactly that. Mice on repeated semaglutide recovered both their taste for palatable food and the dopamine activity underlying it.[1] **Everything here is in mice.** It is a mechanism worth understanding, not a finding about you. In people, [a small trial found semaglutide sharpened taste recognition while dimming the brain’s response to pictures of food](https://glpwatchdog.org/research/taste-got-better-not-worse).

> **This work uses optogenetics and fiber photometry, which switch individual neurons on and off with light and record their activity.** Neither is possible in humans. What follows is rodent neuroscience, and hedonic eating in a mouse is related to, but not the same as, what people mean by food noise.

## What they found

Hedonic eating is eating for pleasure rather than need, like the second helping of something delicious when hunger is already gone. The researchers traced a pathway running from a region near the locus coeruleus to the ventral tegmental area, and found that dopamine neurons there **encode palatability**: their activity tracks how good the food is, and driving them up or down raises or lowers how much of it an animal eats.

Semaglutide suppressed the responsiveness of those neurons during eating. That is a plausible mechanism for the drug reducing pleasure-driven consumption specifically, rather than only signaling fullness.

> Then, with repeated treatment, the appetite came back, and so did the neurons.

That recovery is the finding. Over continued semaglutide, mice regained both their consumption of palatable food and the dopamine activity that accompanies it. The suppression was not permanent.

## The step that makes it a causal claim

Two things recovering together is a correlation. The researchers then inhibited those dopamine neurons at the moment of eating, and the recovery of appetite was reversed.

That is what separates this from an observation. If switching the circuit off undoes the effect, the circuit is doing the work rather than merely accompanying it. Their conclusion is that hedonic eating activates these neurons, which then sustain further eating. The result is a self-reinforcing loop that *opposes* the appetite reduction the drug produces.

> **That is a mechanism for the plateau, and it is a more precise idea than “the drug stops working.”** The drug may keep doing exactly what it did on day one, while a separate circuit progressively reasserts itself against it. Those are different situations with different implications, and only one of them sounds like failure.

## What it does and does not explain

Weight loss curves in every trial of this class flatten over time, and the flattening is usually attributed to metabolic adaptation, in which the body defends its weight by lowering energy expenditure. That is real and well documented. This work proposes something running alongside it: a reward pathway recovering its influence over food choice.

It would fit what people report. The return described is often not general hunger but a specific pull toward particular foods, which is what a palatability-encoding circuit would produce rather than what a broad hunger signal would.

It does not explain everything, and it does not establish anything in humans. The equivalent experiment cannot be run in people, so testing the idea here means indirect evidence: imaging, food-choice measures, and drugs that act on the same pathway. Very little of that has been done.

## Why there is no advice at the end of this

Some articles end with something to do. This one does not, and inventing a takeaway would be a mistake. Nobody can inhibit their own ventral tegmental dopamine neurons, and no drug currently prescribed does it deliberately.

What it offers is a frame. If appetite for particular foods returns during treatment, that is a documented phenomenon with a candidate biological mechanism, not a personal failure of resolve and not proof the treatment has quit. The evidence on what happens to weight over the long run is in [the semaglutide timeline](https://glpwatchdog.org/research/semaglutide-results-timeline), and what patients themselves call the effect is in [food noise, what it is](https://glpwatchdog.org/research/food-noise-what-it-is).

## Frequently asked questions

### Why does appetite come back on a GLP-1?

In mice, dopamine neurons in the ventral tegmental area that encode how good food tastes were suppressed by semaglutide, then recovered their activity during continued treatment, and appetite for palatable food recovered with them. Whether the same happens in people is untested.

### Does this mean the drug stops working?

That is not what the study describes. It suggests a separate circuit progressively reasserting itself against a drug that may still be doing what it always did. That is a different situation from the drug failing.

### Was this in humans?

No. It used optogenetics and fiber photometry in mice, techniques that cannot be applied to people. Hedonic eating in a mouse is related to but not the same as what humans describe as food noise.

### How do they know the neurons caused it?

Because inhibiting those neurons at the moment of eating reversed the recovery of appetite. Switching a circuit off and seeing the effect disappear is what distinguishes a causal role from an accompanying one.

### Is there anything to do about it?

Nothing actionable follows from this study, and it would be dishonest to invent something. Its value is a frame: appetite returning during treatment is a documented phenomenon with a candidate mechanism, not a failure of willpower.

## References

1. Zhu Z, Gong R, Rodriguez V, et al. Hedonic eating is controlled by dopamine neurons that oppose GLP-1R satiety. Science. 2025. PMID: 40146831. https://pubmed.ncbi.nlm.nih.gov/40146831/

---

Cite as: GLP Watchdog, Nora Bissett, "Why the Appetite Comes Back", published 2026-08-20, updated 2026-10-30. https://glpwatchdog.org/research/why-the-appetite-comes-back

Canonical HTML page: https://glpwatchdog.org/research/why-the-appetite-comes-back

Medical disclaimer: Nothing here is medical advice. We report on this subject; we do not treat anyone. Whether to begin a medication, supplement, or treatment, whether to adjust it, and whether to stop are questions for a clinician licensed to answer them, who knows your history.
