People taking semaglutide and similar drugs kept reporting something their prescriptions did not promise. The nightly drink stopped feeling automatic. Cigarettes lost urgency. The mental image of a cold beer stopped pulling.
Neuroscientists trying to explain that ran into an inconvenient problem. The brain regions everyone associates with reward turned out to be poorly equipped to do the job.
Attention has shifted to a structure that most textbooks file under emotional regulation, and whose most famous experimental result involved aggressive animals.
The Obvious Suspects Lack the Receptors
The ventral tegmental area and the nucleus accumbens are the dopamine-associated regions that have anchored reward research for decades. They are the first place anyone would look for a mechanism.
The difficulty is receptor density. Candidate regions for reward-modulating activity have a relative paucity of GLP-1 receptors, making them unlikely sites of direct drug action. A drug cannot exert a strong direct effect on a region that lacks the receptor it binds.
That forces the search one step upstream to the lateral septum.
Its research history is unusual. In 1953, behavioral researchers Joseph Brady and Walle Nauta coined the term septal rage after animals with damage to the region showed increased aggression, while direct stimulation of the region reduced it. For most of the following decades, the lateral septum was studied as a structure involved in emotion and aggression.
More recent connectivity work has reframed it as a hub at the center of a network linking many regions with varied functions, and the aggression finding now looks like one output among many.
Memory, Place and What Is Good Here
The lateral septum draws much of its primary input from the hippocampus, which is the reason this account is interesting rather than merely anatomical.
The hippocampus is where long-term episodic memories are formed. It also contains place cells, neurons that fire according to where a person or animal is in space.
That signal about where and when is forwarded onward. Research has shown that the lateral septum contains place cells that respond strongly to rewards. Functionally, they append information about what is good in a particular place to the hippocampal signal about where the place is, then share that with dopamine-producing machinery.
This is why the region is a candidate for craving specifically rather than consumption generally. A craving is rarely abstract. It is tied to a bar, a time of day, a routine, a mental image. The lateral septum is positioned where those elements would be bound together, and it is densely loaded with GLP-1 receptors.
The framework was laid out by Robert Munn, a senior lecturer at the University of Otago, in an explainer published by The Conversation.
The Animal Evidence Is Specific, and It Is Animal Evidence
The strongest direct evidence to date came from rodent studies published in eBioMedicine by researchers at the University of Gothenburg's Sahlgrenska Academy.
Infusing the GLP-1 receptor agonist exendin-4 directly into the lateral septum reduced alcohol intake in rodents in a dose-dependent way, without affecting food or water consumption. Blocking the receptor in that region increased alcohol intake. Activation also blunted alcohol's rewarding effects, reducing locomotor stimulation, place preference, and accumbal dopamine release, apparently through a GABA-A receptor-dependent mechanism.
The selectivity is the notable part. A general sedative effect would have suppressed food and water intake as well, but it did not.
Additional work has followed. A 2026 study in Neuron described a septal inhibitory circuit that constrains alcohol reward and mediates liraglutide's suppressive effect on alcohol intake in mice. Munn's own laboratory has reported in a preprint that liraglutide reduced the power of high-frequency oscillations and theta rhythm in the lateral septum of rats, which the authors interpret as a change in how the region communicates with others. That result has not completed peer review, which is worth holding in mind before treating it as settled.
All of this is rodent neuroscience. It establishes that a circuit exists and can be manipulated. It does not establish that the same circuit produces the craving reduction people report at clinical doses, nor does it distinguish between wanting something less and enjoying it less, a difference that matters greatly to patients.
What Human Evidence Exists, and What Doctors Can Actually Prescribe
The human data is encouraging and still limited.
An earlier phase 2 trial published in JAMA Psychiatry in 2025 randomized 48 non-treatment-seeking adults with alcohol use disorder to nine weeks of low-dose semaglutide or placebo. It found reduced alcohol consumed during a laboratory self-administration task, with medium to large effect sizes for grams consumed, alongside reductions in craving. It was small, short, and not designed to measure real-world drinking outcomes.
A larger trial followed. Published in The Lancet in May 2026, it randomized 108 treatment-seeking adults with moderate to severe alcohol use disorder and comorbid obesity to 26 weeks of once-weekly semaglutide at 2.4 mg or placebo, with standard cognitive behavioral therapy offered to everyone. Heavy drinking days fell substantially more in the semaglutide group, and 88 of the 108 participants completed the trial. According to an NIH summary of the trial, the frequency of heavy drinking days decreased in both groups, but significantly more with semaglutide.
The trial's own limits matter. It was conducted at a single center; every participant also had obesity; all were offered therapy, and there was no follow-up after the 26 weeks ended. Whether the effect holds in people without obesity or persists after treatment stops is unresolved.
No GLP-1 drug is approved by the FDA for alcohol use disorder or any other addiction. Semaglutide is approved for type 2 diabetes and for weight management, and any use for cravings is off-label. These drugs carry real side effects, including gastrointestinal effects and, in a minority of patients, more serious complications requiring monitoring.
Anyone struggling with alcohol should not wait for this research to mature. Effective treatments already exist, including approved medications and behavioral therapies, and a physician or addiction specialist can discuss which fit a person's situation.
Key Questions Answered
Why are researchers looking at the lateral septum instead of the usual reward regions?
Because the classic dopamine-associated regions carry relatively few GLP-1 receptors, they are unlikely direct targets. The lateral septum sits one step upstream and is densely populated with those receptors.
What does the lateral septum actually do?
It receives heavy input from the hippocampus, contains its own place cells that respond strongly to rewards, and projects to dopamine-producing structures. That positions it to bind together information about place, context, and value.
What did the animal studies show?
Infusing a GLP-1 receptor agonist directly into the lateral septum reduced alcohol intake in rodents in a dose-dependent way without affecting food or water intake, and blocking the receptor there increased alcohol intake.
Does that mean GLP-1 drugs treat addiction in people?
Not yet established. The circuit work is in rodents. Human trials are encouraging but limited, and no GLP-1 drug is approved for alcohol use disorder or any other addiction.
What do the human trials show?
A 48-person phase 2 trial found reduced laboratory drinking and craving over nine weeks. A 108-person, 26-week randomized trial in adults with alcohol use disorder and obesity found significantly larger reductions in heavy drinking days with semaglutide added to cognitive behavioral therapy.
What are the limits of the larger trial?
It was single-center; everyone had comorbid obesity; all participants were offered therapy; and there was no follow-up after treatment ended.
What should someone struggling with alcohol do now?
Speak with a physician or addiction specialist. Approved medications and behavioral therapies already exist and do not require waiting for this research to mature.
