If you only want the short version. Five compounds are sold under this heading, and they are five different kinds of thing.
One is a mixture taken from pig brains. Three are copies of things the body already makes. One is built to imitate a drug from the 1970s.
Four of the five come from Soviet and Russian labs. That shapes what evidence exists, and what language it is written in.
Sorted by human evidence, the order runs Cerebrolysin, then Semax and Selank, then Noopept, then Dihexa. Dihexa has no human trials at all. The papers describing how it works were pulled in 2025, after a university found the data had been faked.
Five things wearing one label
Cerebrolysin
Also sold as FPF-1070. Not a peptide. It is a preparation of low molecular weight peptides and amino acids derived from porcine brain, given by injection, and widely used for stroke in Russia, Eastern Europe and parts of Asia.1
Semax
Sold as Semax and as N-Acetyl Semax Amidate, and derived from the ACTH 4-10 fragment. A synthetic analogue of a piece of ACTH, the hormone that drives the adrenal stress response, modified so it lasts longer.
Selank
Also listed as TP-7. A synthetic analogue of tuftsin, a short immune-signalling peptide the body makes.
Noopept
Known as omberacetam, and as GVS-111 in the original development literature. A dipeptide ester designed at a Russian institute to reproduce the effects of piracetam in a smaller, more potent molecule.
Dihexa
Also listed as PNB-0408. The outlier of the group, and the only one not from a Soviet or Russian lab. It came from a laboratory at Washington State University, built by modifying angiotensin IV, a fragment of the blood pressure system that turns out to affect memory in rats.
The grouping is a marketing convenience. What these five share is a buyer, not a chemistry.
What that buyer is spending is worth stating once. A monthly supply of any of these runs from tens to a few hundred dollars, indefinitely, on the expectation of thinking more clearly. Not one of them has been tested for that in a healthy person. Everything below is an attempt to work out what you would actually be paying for.
| Origin | Route | What was studied | Human trials | US status | |
|---|---|---|---|---|---|
| Cerebrolysin | Austria, porcine brain extract | Injection | Acute ischaemic stroke | Multiple randomised trials | Not approved |
| Semax | Russia, ACTH 4-10 analogue | Nasal spray | Stroke, optic nerve conditions | Yes, mostly Russian-language | Not approved. Recommended for the 503A list July 2026 |
| Selank | Russia, tuftsin analogue | Nasal spray | Anxiety | Yes, mostly Russian-language | Not approved |
| Noopept | Russia, piracetam-derived dipeptide | Oral | Vascular and post-injury cognitive impairment | Yes, small and mostly Russian | Not approved |
| Dihexa | United States, angiotensin IV analogue | Oral or transdermal | Nothing in humans | None | Not approved. Before the compounding committee by February 2027 |
None of the five was studied for cognitive enhancement in healthy adults.
Sorted by evidence, which reorders everything
The list vendors present is usually ordered by how exciting the mechanism sounds. Ordered by what has been tested in people, it looks different.
Cerebrolysin: the most tested, and it failed the test
It has been through multiple randomised trials in acute stroke, which puts it far ahead of everything else here. The Cochrane review of that evidence found no clinical benefit, no effect on deaths, and an increase in non-fatal serious adverse events, with a risk ratio of 2.39 across three trials and 1,335 participants.1 The reviewers went further than usual and argued against running more trials, on the grounds that recruiting patients into a study offering no potential benefit would be unethical.1
That conclusion is contested. A 2021 meta-analysis pooled 2,202 patients across twelve randomised trials and found no statistically significant safety differences against placebo.2 A 2025 review of fourteen trials reported no excess adverse events, and criticised Cochrane for what it included.2 The 2021 authors contacted the manufacturer of Cerebrolysin for additional evidence during their search, which is disclosed in the paper and is the kind of thing a reader should weigh.2
Two systematic reviews of the same drug reaching opposite conclusions on safety is the honest state of this evidence. What nobody claims is that it helps with everyday cognition in healthy people, which is what it is sold for online.
Semax and Selank were approved before the evidence was published
Both are registered medicines in Russia and approved nowhere else. Both have human trials behind them, most published in Russian, few meeting the standards Western regulators expect.
Neither was registered for cognitive enhancement. Semax is used in Russia for stroke and for conditions of the optic nerve. Selank was developed and registered as an anxiolytic, meaning a treatment for anxiety. Neither was approved on the basis of making a healthy person think more clearly, and no trial in either literature set out to test that.
Our comparison of the two goes through the approval timeline and the individual studies, including the Gusev stroke work and the Zozulia trial.
One pattern carries over to the rest of this page. Russian approvals rest on a different evidentiary standard and a different review process, so they do not tell you what an FDA approval tells you.
Noopept was studied in patients, and is sold to everyone else
Its most-cited result is a 2009 study by Neznamov and Teleshova. They tested Noopept against piracetam in patients with mild cognitive disorders, arising from vascular disease and from head injury.5 A more recent Russian trial randomised 150 patients with vascular cognitive impairment across three treatment groups over 45 days.5
Two clinical studies puts Noopept a long way ahead of Dihexa. Both enrolled people with diagnosed impairment. Nobody has published a substantial trial in healthy adults looking for sharper everyday focus, which is what the compound is sold for.
The chemistry sets it slightly apart from the rest of the group. It is a dipeptide ester built to reproduce the effects of piracetam, a 1970s nootropic, in a smaller and more potent molecule.
Dihexa: the retracted papers, and what is left
Every efficacy result for Dihexa comes from cultured neurons and rats.3 No human trial of Dihexa itself has been completed.
The famous claim, that it is roughly ten million times more potent than BDNF, comes from a genuine measurement in a cell culture assay counting new dendritic spines. It says nothing about cognitive potency in a person.3
What happened next is the part vendors leave out, and it is worth following in order.
A laboratory at Washington State University developed Dihexa from angiotensin IV and published the papers explaining how it worked, through a growth factor pathway called HGF and its receptor. Those papers were the mechanism. Everything anyone says about why this compound should improve cognition traces back to them.
A spin-out company formed to commercialise the work. A doctoral student from the same laboratory became its chief executive. The company raised substantial money, took a Dihexa prodrug called fosgonimeton into late-stage Alzheimer's trials, and in 2024 those trials missed their primary and key secondary endpoints.3
The chief executive had already resigned three years before that, in 2021. An investigation had examined the images in her doctoral dissertation, and in papers she co-authored with her supervisor.3
Then in April 2025 the Journal of Pharmacology and Experimental Therapeutics retracted two of those papers. The notices give the reason plainly. The university investigated, and specific figures were found to contain falsified or fabricated data.4
So the position today is that a compound with no human trials is sold on the strength of a mechanism whose foundational papers have been withdrawn for fabricated data, after the one attempt to test the idea in people failed. Animal behavioural results from other groups survived the retractions, so nobody has disproven anything. The mechanism is simply unsupported, which is a worse thing to be buying than a compound that failed honestly.
The version sold is not the version studied
Semax is sold online almost entirely as N-Acetyl Semax Amidate, listed as NA Semax or NASA Semax. Selank is commonly sold in an amidated form too.
Those are modified molecules. An acetyl group on one end and an amide on the other are added to slow breakdown, and the result is a different compound from the one in the Russian trial literature. Our compound library holds Semax and N-Acetyl Semax Amidate as separate entries for that reason.
Nobody has run a head-to-head trial establishing that the modified version does what the original did. The reasoning behind the modification is sound, and no evidence supports the substitution. That shape recurs across this field, where the studies belong to one molecule and the market sells another. The same problem sits at the centre of the TB-500 story our healing pair guide covers.
What "approved in Russia" is doing in every listing
Vendors lean on this phrase for Semax, Selank, Noopept and Cerebrolysin, and it carries less than it appears to.
A registration means a national regulator accepted a dossier. It does not mean the trials were randomised against placebo, powered adequately, or published somewhere a reader can check. For several of these compounds the accessible published record is thin against the length of time they have been in clinical use, and the trials that exist are frequently small.
So the question that gets you somewhere is whether you can read the study that persuaded the regulator.
Where they stand legally
None is FDA approved. Cerebrolysin is not approved in the United States and is not available through a lawful domestic route.
Two are moving through the compounding process, in opposite directions. Semax was among the six peptides an FDA advisory committee voted to recommend for the 503A bulks list in July 2026, against the written advice of the agency's own scientists. Dihexa goes before the same committee by the end of February 2027. Our coverage of the July vote and of the February slate explains what these recommendations are worth, which is less than a vendor will tell you, since formal addition requires rulemaking that has not happened.
What the group has in common
Every one of them is sold for cognitive enhancement in healthy adults, and not one of them has been tested for that. The trials, where they exist, studied stroke, cognitive impairment, anxiety disorders and dementia. Those are populations with something to recover.
The route differs across all five, which changes which of the practical guides applies. Cerebrolysin is injected. Semax and Selank are used almost entirely as nasal sprays. Noopept is taken by mouth. Dihexa is usually sold as an oral or transdermal preparation.
Only one of them, then, involves the reconstitution and rotation questions our injection guides cover. The nasal route brings a different problem, which is that a spray has no measurable unit. A syringe gives you a number even when the vial is unreliable, and a squeeze gives you nothing to check.
And the evidence problem here is structural. Four of the five come from a research tradition whose output is largely untranslated, and the fifth has had its foundations retracted. Neither situation gets fixed by a vendor citing a study title.
What is known about harm
The evidence question and the safety question have different answers here, and the safety answer is thinner.
Cerebrolysin has the only signal drawn from controlled trials, and it is the non-fatal serious adverse event finding described above, which two later meta-analyses dispute.12
Semax and Selank have decades of clinical use in Russia and remarkably little published adverse event reporting to show for it. Absence of reports is not evidence of safety. It usually means nobody was collecting systematically, which is a different statement from a clean record.
Noopept's documented problem is one of product quality. An analysis of over-the-counter cognitive supplements found products containing omberacetam at up to four times pharmaceutical doses, with as many as four unapproved drugs in a single product.5 Someone buying it is frequently not taking only it.
Dihexa's concern is theoretical and serious. The growth factor pathway it is proposed to act on is also a pathway involved in cancer, and no human study has looked.3
Nobody is running post-marketing surveillance on any of these. The adverse events on record are the ones users noticed and chose to mention.
What to do with this
Nothing here says these compounds do nothing. Four of the five have human trials, and a few of those trials found something. The gap is between what was studied and what is being sold, and it is the same gap every time.
Three things follow from that, and they are the useful part of the page.
Ask what population the study enrolled. For every compound here except Dihexa, the answer is people with a diagnosed problem: stroke, vascular cognitive impairment, head injury, anxiety. Someone without the condition that was studied is outside what the trial can speak to.
Ask whether you can read the study. For most of this literature the answer is no, because it is untranslated, unindexed, or both. A vendor citing a title is not the same as evidence you can check, and the site's whole method is refusing to treat those as equal.
Ask which molecule was tested. The nasal spray on sale is frequently a modified version of the compound in the papers, and the modification has never been tested against the original.
For anyone already using one of these, the practical steps are unglamorous. Know what form you actually bought. Keep to one compound at a time, so any effect can be attributed to something. And treat a subjective improvement over a fortnight with suspicion, because expectation is powerful and none of these has been tested against a placebo in someone like you.
Common questions
Which of these has the best evidence?
Cerebrolysin, by volume of trials, and the largest review of those trials concluded it does not work for the condition it was tested in.1 For everyday cognitive enhancement, none of them has direct evidence.
Is Dihexa dangerous?
Nobody knows, because no human safety study exists.3 The specific concern raised in the literature is that the growth factor pathway it is proposed to act on is also involved in cancer, and that has not been studied in people.
Does the retraction mean Dihexa does not work?
The published account of how it works was withdrawn after a finding of fabricated data.4 Behavioural results in animals from other papers were not all directly implicated, so the accurate statement is that the mechanism is unsupported and the compound is untested in humans.
Are Semax and Selank the same thing?
No. Same lab, same length, different targets and different purposes. Our head-to-head comparison covers the differences.
Is Cerebrolysin a peptide?
Not a single peptide. It is a mixture of low molecular weight peptides and amino acids obtained from pig brain tissue.1
Why is so much of this research in Russian?
Four of the five were developed in Soviet or Russian institutes, where the work was published domestically. Language is only part of it. Study design conventions and registration standards also differ, which is why the two kinds of approval are not comparable events.
Can I take these together?
Nobody has tested any combination of them in people. Vendors sell stacks anyway.
Are any of these legal to buy?
They are sold as research chemicals, which is a statement about the sale and not about safety or legality of use. Our guide on what research use only means covers the distinction.
Sources
1: Ziganshina LE, Abakumova T, Hoyle CHV. Cerebrolysin for acute ischaemic stroke. Cochrane Database of Systematic Reviews, 2020, updated 2023. Systematic review, abstract and plain-language summary read at source. Describes Cerebrolysin as a mixture of low-molecular-weight peptides and amino acids derived from porcine brain, widely used for acute ischaemic stroke in Russia, Eastern Europe, China and other Asian and post-Soviet countries. Found no benefit, no obvious effect on deaths, and moderate-certainty evidence of an increase in non-fatal serious adverse events, RR 2.39, 95% CI 1.10 to 5.23, across 3 trials and 1,335 participants, with a more prominent increase in the 30 mL for 10 days subgroup, RR 2.87. The review states that the authors advocate for no more trials with Cerebrolysin on the grounds that recruiting patients into a study offering no potential benefit would be unethical.
2: Strilciuc S, et al. Safety of Cerebrolysin for Neurorecovery after Acute Ischemic Stroke: A Systematic Review and Meta-Analysis of Twelve Randomized-Controlled Trials, 2021; and a 2025 systematic review and meta-analysis of fourteen randomised controlled trials. Peer-reviewed meta-analyses, abstracts read at source. The 2021 analysis pooled 2,202 patients from twelve randomised trials and registered non-statistically significant differences between Cerebrolysin and placebo throughout main and subgroup analyses. Its methods section discloses that the authors contacted the producer of Cerebrolysin to provide additional evidence and references. The 2025 review reports no excess adverse events, no significant increases in serious adverse events, mortality or haemorrhagic transformation, and criticises the Cochrane review for including non-randomised and outdated studies. Author affiliations and funding for both were not examined beyond the disclosed manufacturer contact.
3: Published profiles of Dihexa covering its origin, preclinical evidence base, potency claim, and commercial history. Commercial and community sources, corroborated across five that agree on every material fact. Dihexa is N-hexanoic-Tyr-Ile-(6)-aminohexanoic amide, developed from angiotensin IV in the laboratories of Joseph Harding and John Wright at Washington State University. All efficacy data are preclinical, in cultured hippocampal neurons and rats, with no completed human trial of Dihexa itself. The claim of roughly ten million times the potency of BDNF derives from a cell culture spinogenesis assay and describes molar potency at one narrow endpoint. The prodrug fosgonimeton, developed by the spin-out that became Athira Pharma, missed primary and key secondary endpoints in the LIFT-AD Alzheimer's trial reported in September 2024. The chief executive, a co-author on the compromised papers, resigned in October 2021. The HGF/c-Met cancer concern is reported consistently as theoretical and unstudied in humans.
4: Retraction notice to "Mimics of the Dimerization Domain of Hepatocyte Growth Factor Exhibit Anti-Met and Anticancer Activity", Journal of Pharmacology and Experimental Therapeutics, Volume 392, Issue 4, April 2025; and the companion retraction notice to "Development of Angiotensin IV Analogs as Hepatocyte Growth Factor/Met Modifiers", same journal and issue. Publisher retraction notices. The notice states that the article was retracted at the request of the Editor, that following an investigation by Washington State University specific figures were found to contain falsified and/or fabricated data, and that the journal has retracted the article.
5: Neznamov GG, Teleshova ES. Comparative studies of Noopept and piracetam in the treatment of patients with mild cognitive disorders in organic brain diseases of vascular and traumatic origin. Neuroscience and Behavioral Physiology 2009;39:311-321; a 2026 Russian randomised comparison of omberacetam against piracetam/cinnarizine and aminophenylbutyric acid in 150 patients with vascular cognitive impairment over 45 days; and published analysis of unapproved drugs in over-the-counter cognitive supplements. Peer-reviewed studies and analysis. The 2009 study is consistently described as the strongest clinical evidence for the compound, conducted in patients with diagnosed mild cognitive impairment of vascular and traumatic origin rather than in healthy adults, and as small and old by current standards. The supplement analysis reports that over-the-counter cognitive enhancement products may contain multiple unapproved drugs including omberacetam and aniracetam, at up to fourfold greater than pharmaceutical dosages and with as many as four unapproved drugs in individual products.