Researchers report that semaglutide — the compound sold as Ozempic and Wegovy — may work in the brain in almost the opposite way to what was expected, activating hunger-linked neurons that appear to be necessary for sustaining fat loss. The finding, in mice, is genuinely counterintuitive.

What was expected

The brain contains populations of neurons that drive feeding behaviour. The standard account of how these drugs work is that they suppress those hunger circuits — the drug quietens the signal, appetite falls, people eat less.

That story matches what patients report and has been the working explanation across the field.

What the researchers report finding is closer to the reverse: the hunger-linked neurons are activated rather than silenced, and that activation appears to be required for fat loss to be sustained. If it holds, the mechanism is not simple suppression but something more like a regulatory loop in which those circuits play an active part.

Why a surprising mechanism matters

A reasonable question is why this matters if the drugs already work. Two reasons.

The first is that drug development targets mechanisms. Efforts to build better versions — fewer side effects, oral rather than injected, cheaper — are guided by a model of what the existing compound does. If the model is wrong in an important respect, some of that work is aimed at the wrong thing.

The second concerns what happens when people stop. Weight regain after discontinuation is one of the most consistent findings about this drug class, and it is the central practical problem. A mechanism involving circuits that remain active during treatment is a more plausible starting point for explaining rebound than one in which those circuits were simply switched off.

It was mice

This belongs near the top rather than buried, because it governs how much weight the finding can carry.

Mouse work is how neural circuits get mapped, because it permits interventions impossible in people — switching specific neuron populations on and off and observing what follows. That precision is exactly why the finding is interesting and exactly why it may not transfer. Mouse and human feeding regulation differ, and the history of obesity research contains a long list of mechanisms that were clear in rodents and did not hold up.

Nothing here changes what any patient should do, and nothing here is a reason to alter a prescribed treatment.

The wider point about mechanism

Medicine is full of treatments that work while the explanation of why remains contested or wrong. Aspirin was in use for decades before its mechanism was understood; the account of how several classes of psychiatric drug work has been substantially revised.

This is not a failure of the field. It is what happens when clinical evidence — does it work, is it safe — is easier to establish than mechanistic evidence. It is also a reason to hold confident explanations of how a drug works more loosely than the evidence that it does.

This describes laboratory research and is not medical advice. Questions about weight or medication belong with a doctor.

Sources

• ScienceDaily — semaglutide and hunger-linked neurons, 20 August 2026