Five peptides studied in cognitive and brain research: evidence, limits and open questions

Evidence guide · Reviewed 1 August 2026 · Laboratory research only

Peptide research in the brain covers far more than “cognitive enhancement.” Semax and Selank are investigated as short regulatory peptides. Davunetide was developed from an activity-dependent neuroprotective protein sequence. PACAP is an endogenous neuropeptide with broad signalling effects. GLP-1 receptor agonists connect metabolic signalling with current neurological research. Their evidence bases are different, and several well-designed trials have produced negative or mixed results.

01 · Short

The short answer

No peptide can be called a proven general cognitive enhancer. Semax and Selank have mechanistic, preclinical and limited human research, but international clinical evidence is incomplete. Davunetide reached a large controlled trial and failed its primary endpoints. PACAP remains mainly a mechanistic and preclinical field. GLP-1 medicines have active brain research, with mixed clinical findings that cannot be transferred to unapproved catalogue materials.

That is why a useful comparison focuses on the question each peptide can test. Brain research may measure receptor signalling, neuronal survival, synaptic plasticity, functional connectivity, task performance, imaging biomarkers or disease progression. A positive change in one layer is not proof of improvement in another.

02 · “Brain

Why “brain health” is too broad for a scientific claim

Brain health can refer to attention, working memory, mood, sleep, vascular function, neuroinflammation, neurodegeneration or recovery after injury. These outcomes use different instruments and populations. Even the word cognition includes multiple domains: processing speed, executive function, episodic memory, working memory, language and visuospatial performance.

Peptide studies also face delivery and measurement problems. Many peptides are vulnerable to enzymatic degradation and may not reach the central nervous system in a predictable form. Blood measurements do not necessarily reflect brain exposure. Behavioural results in rodents can be sensitive to strain, stress, handling and task design. Small human studies may be underpowered or exploratory.

A credible review therefore labels the evidence level:

03 · Semax

1. Semax: an ACTH-derived research peptide

Semax is a heptapeptide based on the ACTH(4-7) sequence with a Pro-Gly-Pro extension. It has been studied in neuronal, animal and limited human research, particularly in literature from Russia. Proposed research areas include neurotrophic signalling, oxidative stress, metal interactions, neuronal survival and cognitive tasks.

One laboratory study characterised Semax interactions with copper(II) and reported reduced copper-induced cytotoxicity in two cell lines under the tested conditions. Other research has examined BDNF-related expression and neuronal signalling. These findings are mechanistic. They do not establish a broad clinical benefit or prove that any commercial preparation will produce the same effect.

A 2020 functional-connectivity study compared Semax, Selank and placebo in 52 healthy participants. Researchers measured resting-state fMRI before and after administration and reported group and condition differences involving connectivity between the right amygdala and right temporal regions. Functional connectivity is a useful biomarker, but it is not a direct measurement of memory improvement or long-term clinical benefit.

Semax also illustrates the need for regulatory caution. A forensic-analysis paper reported Semax and Selank in seized preparations and stated that, to the authors’ knowledge at that time, the compounds had not completed clinical trials. The FDA now lists Semax among bulk substances for which it has limited safety information and notes potential immunogenicity and peptide-impurity concerns in compounded products.

Best research question: Does an analytically confirmed Semax preparation alter a preselected neuronal or network endpoint in a validated model?

Main gap: Limited, geographically concentrated and heterogeneous human evidence does not support a general “cognitive enhancement” promise.

04 · Selank

2. Selank: stress-circuit and neuropeptide research

Selank is a synthetic heptapeptide derived from the immunomodulatory peptide tuftsin. Research discussions often focus on anxiety-related behaviour, neurotransmitter systems, gene expression and cognitive or stress endpoints. Online marketing frequently calls it a nootropic or anxiolytic. Those labels can imply proven treatment effects that the broader evidence base does not justify.

The same 2020 fMRI study that evaluated Semax also examined Selank. Changes in resting-state connectivity were reported, including patterns involving the amygdala and temporal cortex. The study is valuable because it moves beyond a simple self-report. It remains a small exploratory study with short observation windows, and its imaging findings should not be converted into claims about treating anxiety or improving memory.

Preclinical Selank work has explored immune-neural interactions and expression of genes related to neurotransmission. Those mechanisms may generate testable hypotheses. A gene-expression change is not automatically beneficial, and a behavioural result in a stress model is not equivalent to a diagnosed human condition.

Analytical identity is important because small peptides can be confused with salts, modified forms or degraded products. A laboratory should confirm sequence-related identity and mass, document content and impurities, and use stability data relevant to the planned model. “Selank acetate” and a differently described Selank material should not be assumed identical without documentation.

Best research question: Does a defined Selank material alter a specified stress-circuit, transcriptional or behavioural endpoint relative to appropriate controls?

Main gap: Human safety, exposure and effectiveness data are not sufficient for broad personal-use claims.

05 · Davunetide

3. Davunetide: why negative trials matter

Davunetide, also known as NAP, is an eight-amino-acid peptide derived from activity-dependent neuroprotective protein. It attracted interest through preclinical work involving microtubule stability, tau biology and neuronal protection. Unlike many catalogue peptides, davunetide advanced into a large multicentre clinical trial.

The phase 2/3 trial in progressive supranuclear palsy randomised 313 participants to davunetide or placebo for 52 weeks. The treatment groups did not differ on the primary clinical endpoints. The investigators concluded that davunetide was not effective for progressive supranuclear palsy. Nasal adverse events were also more common in the davunetide group.

This negative result is scientifically useful. It shows that a plausible mechanism and encouraging preclinical findings can fail when tested against meaningful clinical outcomes. It also demonstrates why evidence reviews must include negative studies. Selecting only positive cell and animal papers would give readers a distorted view of peptide translation.

Davunetide should not be described as a failed peptide in every context. A failed trial answers a defined question about a molecule, disease, formulation, route, duration and endpoints. It does, however, block the claim that preclinical microtubule findings were enough to predict effectiveness in that disease.

Best research question: Which translational assumptions caused promising preclinical signals not to produce clinical benefit in the PSP trial?

Main gap: The central clinical programme did not meet its primary effectiveness endpoints.

06 · PACAP

4. PACAP: powerful biology, difficult translation

Pituitary adenylate cyclase-activating polypeptide, or PACAP, exists mainly as PACAP38 and PACAP27. It signals through PAC1, VPAC1 and VPAC2 receptors and participates in stress, autonomic, vascular, endocrine and neural processes. Brain research has examined neuronal survival, inflammation, synaptic function and injury responses.

Classic preclinical work reported that PACAP38 reduced hippocampal neuronal death in a rat forebrain-ischaemia model. That result helped establish PACAP as a neuroprotection research target. Later work has explored downstream cAMP, MAPK and PI3K-related pathways. Yet broad receptor distribution can complicate translation because a peptide may influence vascular, endocrine and sensory systems as well as the intended neural target.

PACAP is not a simple cognition candidate. It is also implicated in migraine biology and stress responses. A mechanism that appears protective in one injury model may create unwanted effects in another context. Stability, receptor selectivity, peripheral actions and delivery to the relevant brain compartment all matter.

For an institution, PACAP research is often most useful as a systems-biology problem: how receptor subtype, timing, tissue and disease state change the response. It is not evidence for casual personal experimentation.

Best research question: Which receptor, cell type and timing window explain a PACAP-related effect in the chosen neural model?

Main gap: Broad signalling and delivery challenges limit direct translation from animal neuroprotection to a general human cognitive claim.

07 · GLP-1

5. GLP-1 receptor agonist peptides: metabolic signalling meets neuroscience

GLP-1 receptor agonists are established medicines for specific metabolic indications, but their possible effects in neurological conditions are a separate research question. Receptors and signalling relevant to energy balance, inflammation and neuronal function have prompted trials in Alzheimer’s disease, mild cognitive impairment and psychiatric populations.

The clinical picture is mixed. An early randomised study of liraglutide in Alzheimer’s disease reported an imaging signal involving cerebral glucose metabolism, while cognitive scores did not change. A later phase 2b trial in 204 participants with mild to moderate Alzheimer’s disease found no significant difference in its primary cerebral glucose-metabolism outcome. A small proof-of-concept exenatide trial in mild cognitive impairment found no significant between-group effect on its primary cognitive measure.

Other 2025 and 2026 studies continue to examine semaglutide, imaging and cognition in selected populations. Those studies should be read according to design. A single-arm imaging study cannot establish causality. A trial in people with a defined metabolic or psychiatric condition does not prove cognitive enhancement in healthy people.

There is also a product-identity boundary. Evidence for approved liraglutide, exenatide or semaglutide products cannot validate an unapproved catalogue material called “GLP-3.” GLP-3 is not a universal scientific molecule name. On Emirates Peptides it is a catalogue convention associated with triple-receptor research, but that convention must not be generalised to literature or other suppliers.

Best research question: Does a named, regulated GLP-1 medicine change a predefined neurological endpoint in a specified clinical population?

Main gap: Mixed outcomes and strong confounding from metabolic change make “cognitive enhancer” an unjustified category claim.

08 · Comparison

Comparison table

09 · Read

How to read a cognitive peptide paper

Start with the molecule. Record the full sequence, modification, salt form and mass. If the paper uses a natural neuropeptide but the catalogue uses a fragment, do not merge the evidence. Check whether the method confirms identity and whether the tested preparation is described well enough to reproduce.

Then identify the endpoint. “Neuroprotection” may mean cell survival after a laboratory insult. “Memory” may mean performance in one animal task. “Connectivity” means a statistical relationship between signals in brain regions. “Cognition” in a clinical trial should be tied to a validated test and a prespecified analysis.

Look at the comparator and masking. Cell experiments need vehicle and toxicity controls. Animal behaviour studies benefit from blinded scoring and randomisation. Human trials require adequate allocation, masking where possible, a prespecified primary endpoint and handling of missing data. A before-and-after change without a control group can reflect practice effects, expectation or time.

Finally, check whether the conclusion matches the data. If a clinical trial did not meet its primary endpoint, secondary or subgroup signals are hypothesis-generating. If a study measures a biomarker but not performance, the article should not claim cognitive improvement.

10 · Quality

Quality requirements for brain-peptide research

Neural models can respond strongly to impurities, endotoxin, solvent conditions and concentration errors. A certificate of analysis should be batch-linked and should separate identity, purity and content. Mass spectrometry supports molecular mass; chromatography characterises the main peak and impurities under the stated method; peptide-content testing addresses how much actual peptide is present. None of these alone proves biological activity.

Where relevant, laboratories should assess endotoxin and bioburden, verify solution compatibility, document adsorption to labware and define stability during the experiment. A material that degrades during a long assay may expose cells to a changing mixture. Repeated freeze-thaw cycles, light, oxidation and surface interactions can affect some peptides.

Brain research also requires strong governance. Work involving humans needs ethics and regulatory approval. Animal studies need institutional review and welfare standards. Research-use labelling does not bypass those obligations.

11 · Questions

Frequently asked questions

Conclusion

The most important finding from cognitive peptide research is not a winner. It is the distance between a molecular idea and a dependable human outcome. Semax and Selank have interesting but limited translational evidence. Davunetide shows the value of a definitive negative trial. PACAP demonstrates the complexity of broad neuropeptide systems. GLP-1 research shows how mixed results can emerge as a field moves from metabolism into neurology.

A responsible 2026 article keeps those differences visible. It names negative results, separates biomarkers from cognition, identifies the exact peptide and refuses to convert research materials into personal recommendations. That approach is more useful to researchers and more trustworthy to readers.

13 · Practical

Practical takeaway for research teams

Select the cognitive domain before selecting the peptide. A network-connectivity study, neuronal-survival assay and working-memory task need different hypotheses and controls. Record whether the outcome is mechanistic, behavioural or clinical and do not substitute one for another. Include negative comparators, verify brain-relevant exposure where possible and preregister the primary analysis for confirmatory work. The literature becomes much easier to interpret when each result is labelled by molecule, model, endpoint, time and evidence layer.

14 · References

References

1. Panikratova YR, et al. Functional connectomic approach to studying Selank and Semax effects. Doklady Biological Sciences. 2020.
2. Tabbì G, et al. Semax, an ACTH4-10 peptide analog with high affinity for copper(II) ion and protective ability against metal induced cell toxicity. Journal of Inorganic Biochemistry. 2015.
3. Vanhee C, et al. The occurrence of putative cognitive enhancing research peptides in seized pharmaceutical preparations. Drug Testing and Analysis. 2020.
4. Boxer AL, et al. Davunetide in patients with progressive supranuclear palsy: a randomised, double-blind, placebo-controlled phase 2/3 trial. Lancet Neurology. 2014.
5. Uchida D, et al. Prevention of ischemia-induced death of hippocampal neurons by pituitary adenylate cyclase activating polypeptide. Brain Research. 1996.
6. Gejl M, et al. In Alzheimer’s disease, 6-month treatment with GLP-1 analog prevents decline of brain glucose metabolism. 2016.
7. Mullins RJ, et al. A pilot study of exenatide actions in Alzheimer’s disease. 2019.
8. Hölscher C, et al. Long-acting exenatide does not prevent cognitive decline in mild cognitive impairment. 2024.
9. Edison P, et al. Liraglutide in mild to moderate Alzheimer’s disease: a phase 2b clinical trial. 2025.
10. US Food and Drug Administration. Bulk drug substances that may present significant safety risks.