Scientists Find Ozempic May Alter Brain Structure, Prompting Questions About GLP-1 Drug Safety

Allison Shapiro was not looking for a controversy. The assistant professor at the University of Colorado Anschutz Medical Campus had been scanning the brains of adolescents and young women taking GLP-1 receptor agonists for a hormonal condition when she noticed something unexpected: within months of treatment, the participants' brain connections in the salience network — a set of structures that helps direct attention — had multiplied extensively [1][2].

"We didn't expect to see this effect, and we really don't know what it means," Shapiro said [1].

Her findings, drawn from brain scans of 13 teens and young women with polyendocrine metabolic ovarian syndrome, add to a growing body of research suggesting that drugs like semaglutide (sold as Ozempic and Wegovy) and tirzepatide (Mounjaro and Zepbound) are doing more than suppressing appetite and lowering blood sugar. They appear to be reshaping the brain itself — and the scientific community is divided on whether that should be a cause for concern or celebration.

The Scale of Exposure

The question of brain effects would be academic if only a handful of patients were involved. They are not.

Source: MedRxiv / GoodRx Analysis

Data as of Mar 1, 2026CSV

GLP-1 receptor agonist prescriptions in the United States have increased by more than 500% between 2019 and 2024 [3]. Between January 2019 and March 2026, nearly 2.9 million patients received a GLP-1 prescription across a large US health system network, accounting for over 14.7 million total prescriptions [4]. First-time prescribing of anti-obesity semaglutide increased by over 50% in the first quarter of 2026 alone [4]. Novo Nordisk has reported more than 50 million patient-years of real-world semaglutide safety data globally [5].

Source: HealthVerity Real-World Data

Data as of Dec 1, 2025CSV

Between 2020 and 2023, roughly one in four patients prescribed semaglutide or tirzepatide did not carry a type 2 diabetes diagnosis, indicating that a substantial and growing share of GLP-1 use is driven by weight loss and other emerging indications [6]. In December 2025, the World Health Organization issued its first global guideline endorsing GLP-1 medicines for treating obesity [7].

With this many patients on these drugs — many for extended durations — the neurological findings carry population-level significance.

What the Brain Scans Show

The brain changes documented so far span several regions and types of neural activity.

The salience network. Shapiro's team at Colorado Anschutz found that GLP-1 treatment produced extensive changes in brain connectivity within the salience network within just a few months [1][2]. This network, which helps the brain decide what deserves attention, is also implicated in addiction and compulsive behavior.

The hypothalamus. The findings pointed researchers toward the hypothalamus, a small but powerful brain region that regulates hunger, stress, sleep, and hormones — and which contains high concentrations of GLP-1 receptors [1]. Preclinical studies have confirmed that semaglutide reaches hypothalamic tissue despite limited blood-brain barrier penetration [8].

Reward circuitry. A 2026 study published in Nature by Ali D. Güler's team at the University of Virginia identified a specific neural pathway connecting the hindbrain, central amygdala, and dopamine neurons in the nucleus accumbens [9]. GLP-1 drugs dampen dopamine release through this circuit, making high-calorie foods less appealing at a neurological level. The pathway runs from the nucleus tractus solitarius, which projects GLP-1-secreting neurons to GABA neurons in the ventral tegmental area, reducing dopamine activity [10].

Frontal and temporal lobes. Earlier imaging studies have suggested a potential slowdown in volume loss in frontal, temporal, and parietal lobes — areas involved in planning, memory, emotion, and sensory integration [1]. However, these findings remain preliminary, and researchers have not published precise measurements of volume change in absolute terms that would allow comparison with normal age-related brain atrophy (typically 0.2–0.5% per year after age 60).

The absence of standardized dose-response data is a significant limitation. Most published studies have used standard therapeutic doses of semaglutide (0.5–2.4 mg weekly) or tirzepatide, but no study has yet mapped brain changes across a range of doses and treatment durations in a large cohort with sufficient statistical power to establish clear thresholds.

Do Structural Changes Mean Functional Harm?

This is the central question, and the evidence points in multiple directions.

Signals of benefit. A 12-month propensity-score matched cohort study of semaglutide use for type 2 diabetes found that the drug was not associated with worsened neurological or psychiatric outcomes [11]. A 2026 randomized clinical trial examined semaglutide for cognitive dysfunction in major depressive disorder, reflecting growing interest in the drug's potential cognitive benefits [8]. Preclinical evidence consistently shows that GLP-1 receptor agonists reduce substance intake, relapse-like behaviors, and cue-induced drug seeking across multiple drug classes [10][12]. Lorenzo Leggio, clinical director at the National Institute on Drug Abuse, is actively investigating GLP-1 drugs for addiction treatment applications [1].

Signals of concern. Patient reports describe emotional flattening, reduced motivation, diminished interest in hobbies, lower sexual desire, and brain fog [1][2]. Some patients have reported reduced pleasure capacity — a state clinicians call anhedonia. These reports remain anecdotal and have not been systematically quantified in large-scale trials, but they are consistent enough to have drawn the attention of multiple research groups.

The critical distinction. Structural brain change does not automatically equal functional harm. The brain remodels itself throughout life — learning a new language, exercising regularly, or recovering from a stroke all produce measurable structural changes. The question is whether GLP-1-driven changes impair function, improve it, or simply reflect a brain adapting to a new metabolic state.

The Case That Brain Changes Are a Feature, Not a Bug

Several researchers argue that GLP-1-driven neuroplasticity in reward and appetite circuits is precisely what makes these drugs effective — and that the brain changes should be understood as therapeutic.

Source: OpenAlex

Data as of Jan 1, 2026CSV

Academic interest in this question has surged. Publications on GLP-1 neurological effects grew from 122 papers in 2011 to 2,328 in 2025 — a nearly 20-fold increase [13].

Güler's team at the University of Virginia found that GLP-1 drugs "reduce not just hunger, but the desire to pursue rewarding food" by modifying how the brain assigns value to high-calorie foods [9]. In an era when obesity affects 41% of American adults [14], a drug that recalibrates a dysfunctional reward system — rather than simply making patients feel full or nauseous — could represent a meaningful advance.

Source: WHO Global Health Observatory

Data as of Dec 31, 2024CSV

Daniel Drucker at the University of Toronto, a pioneer in GLP-1 drug development, has emphasized that GLP-1 receptors are naturally present throughout the brain and that the drugs interact with existing biological systems rather than creating artificial ones [1]. Preclinical research also shows GLP-1 agonists may reduce neuroinflammation, support cerebral blood flow, and protect against amyloid plaque buildup associated with Alzheimer's disease [15].

The counterargument from skeptics: just because a system is natural doesn't mean pharmacological manipulation of it at supraphysiological doses is safe. And reducing the desire for pleasurable food is pharmacologically adjacent to reducing the capacity for pleasure altogether.

Funding, Conflicts, and the Credibility Question

The funding landscape for GLP-1 brain research is mixed, and both sides of the debate carry potential conflicts.

On the industry side, Novo Nordisk has directly funded some brain-related studies and sponsored the large Evoke and Evoke+ Alzheimer's trials involving more than 3,800 participants [16]. Those trials found semaglutide was not significantly better than placebo at preserving brain function after two years, though it improved some biological markers tied to Alzheimer's [16]. Some investigators who have published on GLP-1 brain effects serve as consultants for Novo Nordisk or Takeda, though at least one has stated he does not accept personal payment for such services [17].

On the independent side, the NIH's National Institute of Diabetes and Digestive and Kidney Diseases has funded GLP-1 research in adolescents [17], and Güler's University of Virginia study was NIH-funded [9]. The National Institute on Drug Abuse's investigation into GLP-1 drugs for addiction is also federally funded [1].

The competing drug angle is less prominent in this case. Unlike the statin or SSRI debates, where rival drug classes stood to benefit from safety scares, GLP-1 drugs currently have no direct pharmacological competitors for obesity treatment at comparable efficacy. The primary commercial tension is between Novo Nordisk (semaglutide) and Eli Lilly (tirzepatide), both of which have financial interests in the GLP-1 class succeeding.

What the FDA Has — and Hasn't — Required

On January 13, 2026, the FDA took a notable step: it requested that manufacturers remove information about the risk of suicidal ideation and behavior from GLP-1 drug labels [18]. This followed a comprehensive review, including a meta-analysis of 91 placebo-controlled trials involving 107,910 patients, which found no increased risk of psychiatric adverse events [18][19]. The European Medicines Agency had launched its own review of the suicidal ideation signal in 2023, and the FDA's 2026 decision effectively closed that chapter [18].

However, the FDA has not required specific post-market studies of long-term CNS structural effects. The agency has added postmarketing language regarding gastrointestinal complications (intestinal obstruction, severe constipation, fecal impaction), indicating that class-wide surveillance remains active for some organ systems [18]. But brain imaging studies, long-term cognitive testing, and neurological monitoring are not currently mandated.

The FDA has not concluded that reported mental effects — emotional flattening, reduced motivation, anhedonia — represent widespread problems [1]. The agency's benefit-risk calculus currently favors the drugs given the substantial cardiovascular and metabolic benefits: Wegovy is FDA-approved to reduce the risk of major adverse cardiovascular events including death, heart attack, and stroke in adults with heart disease and obesity [5].

The Historical Parallel: SSRIs and Delayed Discovery

The SSRI precedent is instructive. Selective serotonin reuptake inhibitors were approved in the late 1980s and promoted as safer alternatives to older antidepressants. FDA approval did not require long-term safety trials at the time [20]. Over three decades, post-market surveillance and independent research revealed a series of effects that were not characterized at approval: akathisia (severe restlessness), post-SSRI sexual dysfunction that can persist after discontinuation, metabolic disruption, and possible cognitive effects [20][21].

The 2007 FDA Amendments Act significantly strengthened the agency's authority to require post-market safety studies, representing a regulatory shift that did not exist when SSRIs were approved [20]. Whether this enhanced authority will be applied to GLP-1 drugs' neurological profile remains to be seen.

The statin parallel is somewhat different. Long-term statin use was associated with reports of cognitive complaints, prompting the FDA to add a warning to statin labels in 2012 — but subsequent large-scale studies generally failed to confirm a causal link, and the label language was later softened.

The Populations at Greatest Unknown Risk

The gap between who is taking GLP-1 drugs and who has been studied for neurological effects is most pronounced in three groups.

Adolescents. Shapiro's Colorado Anschutz study specifically involved adolescents and young women, a population whose brains are still undergoing major developmental remodeling through the mid-20s [1][2]. Pediatric endocrinologist Melanie Cree at the same institution had been studying whether GLP-1 drugs could help adolescents with metabolic ovarian conditions, and the unexpected brain connectivity findings emerged from that work [1]. The sample size — 13 participants — is too small to draw population-level conclusions, but the findings in a developing brain are a red flag that warrants larger studies.

Pregnant women. The FDA warns against using semaglutide during pregnancy, and animal studies have shown thyroid C-cell tumors [22]. A study found that most Australian women of reproductive age prescribed GLP-1 drugs were not using effective contraception [23]. Separately, GLP-1 drugs may accelerate metabolism of oral contraceptives, reducing their efficacy [23]. The so-called "Ozempic babies" phenomenon — unexpected pregnancies among GLP-1 users — has drawn media attention, but the neurological effects of fetal GLP-1 exposure remain essentially unstudied [22].

People with pre-existing neurological conditions. Patients with depression, anxiety disorders, eating disorders, or neurodegenerative conditions may respond differently to drugs that alter reward circuitry and brain connectivity. The semaglutide trial for cognitive dysfunction in major depressive disorder [8] represents one of the first attempts to study a neurologically vulnerable population, but the vast majority of GLP-1 safety data comes from metabolically focused trials that excluded patients with significant psychiatric comorbidities.

What Comes Next

The scientific community is at an inflection point. Researchers have established that GLP-1 drugs produce measurable brain changes. They have not established whether those changes are clinically meaningful for the average patient, let alone for vulnerable subgroups.

Several questions will determine how this story unfolds. First, whether the FDA will require neuroimaging or cognitive outcome data as part of post-market commitments for GLP-1 drugs — a step it has not yet taken. Second, whether large-scale prospective studies (not the small pilot studies conducted so far) will confirm or refute the preliminary findings. Third, whether the anecdotal reports of emotional flattening and anhedonia will be borne out by rigorous, blinded assessment.

The stakes are high. GLP-1 drugs have produced genuine, large-magnitude benefits for obesity and type 2 diabetes — conditions that themselves carry serious neurological consequences including increased dementia risk, stroke, and peripheral neuropathy. The benefit-risk calculation is not GLP-1 brain effects versus no brain effects; it is GLP-1 brain effects versus the brain effects of untreated metabolic disease.

As Shapiro told reporters about her salience network findings: the data surprised her, and she does not yet know what it means [1]. That honest uncertainty — replicated across dozens of research groups worldwide — is the most accurate summary of where the science stands.