Peptides in Cognition & Sleep Research

Published August 2026 · Patriot Labs Research Team · ~13 min read

What this category actually is

A category tag on a catalog is an organising convenience, not a claim about chemistry. These four peptides are filed together because the published work on all of them uses central-nervous-system models — brain tissue, neuronal cultures, rodent behaviour. That is the whole of what they share.

Structurally and historically they could hardly be further apart. Semax and Selank came out of the same Soviet and then Russian programme at the Institute of Molecular Genetics, built on one design idea: take a short fragment of a natural human peptide and add a Pro-Gly-Pro tail at the C-terminus to slow enzymatic cleavage. DSIP is a Swiss discovery from the 1970s, isolated from blood rather than designed, and its status as a real endogenous factor has never been settled. Pinealon belongs to a fourth tradition — the Khavinson school of short peptide bioregulators, which proposes a mechanism unlike anything in the other three.

So the honest way to survey the category is one compound at a time, with the state of the evidence attached to each — and note in advance that the compound with the largest clinical corpus also has the largest replication problem. If the vocabulary is unfamiliar, the introduction to peptides covers sequence notation and why chain length matters.

Semax — an ACTH fragment with a proline tail

Semax is the heptapeptide Met-Glu-His-Phe-Pro-Gly-Pro. The first four residues are ACTH(4–7), a fragment of the adrenocorticotropic hormone sequence within the melanocortin family; the last three are a synthetic Pro-Gly-Pro extension. Filippenkov and colleagues, describing the compound in a 2020 rat transcriptome study, state the design rationale plainly: the C-terminal PGP tripeptide "was included to ensure the resistance of Semax to peptidases." The parent research target was the ACTH(4–10) region, which is why the peptide is usually introduced as an ACTH(4–10) analogue even though only four of those residues survive in it.

That resistance claim is worth checking rather than repeating. A 2006 degradation study by Zolotarev and colleagues incubated Semax with rat basal forebrain cell cultures and plasma membranes and found that splitting away of the N-terminal Met-Glu and the C-terminal Gly-Pro, with formation of pentapeptides, were the predominant processes — and that glial and neuronal cells produced different product patterns. The tail slows degradation; it is not itself inert.

The pathway it is studied for is neurotrophin expression. Agapova and colleagues followed Ngf and Bdnf transcript levels in rat hippocampus and frontal cortex after intranasal Semax and reported rapid, long-lasting activation of both genes, with the pattern differing between the two brain structures. Filippenkov's group later ran a full transcriptome in a rat transient middle cerebral artery occlusion model, identifying 394 differentially expressed genes at 24 hours — 191 up, 203 down — with inflammation-associated transcripts suppressed and neurotransmission-associated transcripts activated, and Bdnf among the genes altered under peptide exposure but not by ischaemia–reperfusion alone.

Now the part that vendor copy leaves out. Semax has a Russian clinical literature, and it has not been reproduced in Western trials. The Alzheimer's Drug Discovery Foundation's independent Cognitive Vitality evaluation concludes that "published literature of well-conducted studies is lacking," notes there are few published English-language human studies, and describes the reviewers as reliant on abstracts for much of the Russian-language material. A 2018 Russian meta-analysis by Shmonin and colleagues in the acute stroke setting screened 167 PubMed records and 197 records from the Russian elibrary database, narrowed to eight trials, and could include only three of them — 181 participants in total. Its authors' own recommendation is for future multicentre, adequately powered, double-blind studies using standardised designs.

So: a substantial preclinical molecular literature, a small and largely single-language clinical literature, no independent Western replication. Semax is supplied as a 10 mg lyophilised vial; the head-to-head chemistry against its sibling compound is covered in Semax vs Selank.

Selank — a tuftsin analogue on GABAergic endpoints

Selank is Thr-Lys-Pro-Arg-Pro-Gly-Pro, built the same way Semax is. The first four residues are tuftsin, a natural tetrapeptide located in the Fc domain of the immunoglobulin G heavy chain at residues 289–292 and described since its discovery as a phagocytosis-stimulating immune activator. The last three are the same Pro-Gly-Pro tail. Volkova and colleagues put it exactly that way: a synthetic analogue of endogenous tuftsin "elongated at the C terminus via the addition of three natural L-amino acids (Pro-Gly-Pro)." Note what that lineage implies — the parent molecule is an immunopeptide, not a neuropeptide, so any central signalling attributed to Selank is a property of the analogue rather than something inherited.

The pathway studied is GABAergic signalling. The most cited experiment is Volkova and colleagues' 2016 work in Frontiers in Pharmacology: thirty male Wistar rats, intranasal administration, frontal cortex tissue, a panel of 84 neurotransmission genes read at one and three hours. Forty-five transcripts changed significantly at one hour and 22 at three hours, including several GABA_A receptor subunit genes — Gabre and Gabrq moved sharply down at one hour and up at three — alongside dopamine and serotonin receptor transcripts. The authors report a positive correlation of r = 0.86 between the changes produced by Selank and those produced by GABA itself at one hour, and read the residual differences as consistent with allosteric modulation rather than direct GABA mimicry.

A companion study from the same group in 2017 is the more informative one, and it points the other way. In human IMR-32 neuroblastoma cells expressing functional GABA_A receptors, Filatova and colleagues found GABA alone altered 14 of 69 analysed genes and olanzapine altered 25 — while Selank alone produced no significant changes at all. Only in combination did a signature appear: co-application with GABA collapsed the altered-gene count from 14 to one, and co-application with olanzapine produced changes in 35. The authors conclude Selank "has no direct effect on the mRNA levels" of GABAergic genes in that system, and say plainly that they cannot yet explain the combination effects.

A 2008 report by Inozemtseva and colleagues in Doklady Biological Sciences addresses intranasal Selank and BDNF expression in rat hippocampus; we verified the citation but not the full text, so nothing beyond its subject is claimed here.

One framing note. Papers in this literature routinely open by describing Selank as anxiolytic. Those statements appear in introductions as citations to earlier Russian-language clinical work, not as findings the papers themselves generated — the data in the papers are transcript counts. Selank is supplied as a 10 mg lyophilised vial.

DSIP — a 1970s sleep factor that never resolved

Delta sleep-inducing peptide is the nonapeptide Trp-Ala-Gly-Gly-Asp-Ala-Ser-Gly-Glu, molecular weight around 850 daltons, isolated by the Schoenenberger–Monnier group in Switzerland from the cerebral venous blood of rabbits. Sources differ on date and conditions: the standard encyclopaedic account places the isolation in 1974 from rabbits in an induced sleep state, while Tukhovskaya and colleagues, who work on the peptide, date it to 1977 and specify that the donor blood was collected after low-frequency electrical stimulation of the intralaminar thalamic nuclei. Unlike the other three here, DSIP was not designed — it was hunted, the product of a mid-century programme that assumed a circulating humoral sleep factor existed and set out to purify it. That origin explains both its name and its problems.

DSIP's identity as a true sleep factor is contested, and has been for decades. The evidence normally required to call a molecule an endogenous signalling factor is largely missing. No precursor gene has been identified in rabbits, and neither has a receptor. The peptide is reported to have an in-vitro half-life of roughly 15 minutes against aminopeptidase-like activity, prompting proposals that it circulates bound to a carrier or within a larger precursor — but no structure or gene has been found for such a precursor either. Sequence-alignment searches have turned up a match to a hypothetical bacterial protein, raising the awkward possibility of a non-mammalian origin. And the sleep findings conflict: some studies report associations with slow-wave sleep promotion and REM suppression, others report no correlation, and several synthesised analogues have shown stronger effects than the native nonapeptide — a finding about the analogues, not about DSIP.

The most telling detail is who says so. Kovalzon and Strekalova titled their 2006 review "a still unresolved riddle," and Tukhovskaya's group quote that phrase approvingly in their own 2021 paper. Investigators actively publishing on a molecule rarely describe its central claim as unresolved unless it is.

Contemporary DSIP work has largely moved off the sleep endpoint. Tukhovskaya and colleagues used intranasal DSIP in a Sprague-Dawley rat middle cerebral artery occlusion model and reported improved rotarod latency over 21 days despite infarct volumes comparable to vehicle controls — a functional change without a corresponding tissue-sparing change, which the authors cannot mechanistically explain. Separately, Mu and colleagues in 2024 built a DSIP fusion construct with the Tat cell-penetrating sequence precisely because peptides of this kind "would otherwise struggle to effectively penetrate" the blood-brain barrier.

DSIP is supplied as a 15 mg lyophilised vial. The compound-specific guide is what DSIP is.

Pinealon and the bioregulator concept

Pinealon is the tripeptide Glu-Asp-Arg, usually written EDR. Three residues is short even by the standards of this page, and that shortness is the point of the theory attached to it.

What "peptide bioregulator" means as a research concept. The term belongs to the Khavinson school at the St Petersburg Institute of Bioregulation and Gerontology. Its central proposal, set out in a 2021 systematic review in Molecules, is that peptides of two to seven residues act through a mechanism unlike classical receptor pharmacology: they are said to penetrate into cell nuclei and nucleoli, to engage in sequence recognition at gene promoters through direct interaction with double-stranded DNA, and to influence DNA methylation status — to function, in short, as epigenetic regulators rather than as ligands for a surface receptor. Several are described as fragments of larger tissue-derived polypeptide preparations; EDR is presented in this literature as a tripeptide isolated from the preparation Cortexin.

That is a different kind of claim from the ones made for Semax or Selank, and it needs a different kind of evidence. For EDR the proposal is that the peptide enters cells, binds histone proteins or nucleic acids, interacts with the major groove of double-stranded DNA at the N7 and O6 atoms of guanine, and modulates MAPK/ERK signalling and the synthesis of downstream targets including caspase-3, p53, SOD2, GPX1, PPARA, PPARG and calmodulin.

The authors are unusually candid about the status of all this. The 2021 Molecules paper on EDR is written throughout in hypothesis language — "the EDR peptide is assumed to…", "it is possible that…" — states outright that "this hypothesis requires further investigation and experimental confirmation," notes that EDR's effect on several relevant kinases has not been studied, and characterises the protective mechanisms as assumed from molecular modelling rather than directly demonstrated. The companion systematic review concedes that "there are very few studies in this area."

The wet-lab results follow the same shape. A 2021 study in Pharmaceuticals compared EDR against a second tripeptide, KED, in 5xFAD transgenic mice. Both limited dendritic spine loss in hippocampal CA1 neurons, but KED restored mushroom spine density to control levels while EDR's effect was modest and varied by sex; the neuroplasticity trend for KED did not reach statistical significance. A 2022 review lists further EDR findings — spine restoration in hippocampal cultures under amyloid toxicity, promoter binding sites for SOD2, PPARα/PPARγ and CALM1, ERK1/2 activation — while detailing no mechanism for cellular entry or barrier crossing.

The replication problem is structural. The author lists across those four papers recur: Khavinson, Linkova, Ilina, Petukhov, one institute. This is a coherent body of work in peer-reviewed international journals that is nonetheless largely produced by a single research school and has attracted limited independent replication. That is not a claim of error — it is a statement about how much weight a single-group literature can carry, and the same caveat applies to the better-known compound from the same tradition. See what Epithalon is and the wider anti-aging category survey.

Pinealon is supplied as a 10 mg lyophilised vial; the compound guide is what Pinealon is.

The four side by side

This table summarises what each compound is and how solid its record is. It deliberately has no "effectiveness" column, because no study compares these four against one another on any endpoint.

The shared problem: getting into the brain

Every statement above about a peptide acting centrally carries an unstated premise: that the molecule reached the tissue where the effect is claimed. That premise is not free, and it is the single largest methodological weakness this category shares.

The blood-brain barrier is not a filter with a pore size. As Sánchez-Navarro and Giralt set out in their 2022 review of peptide shuttles, the barrier's endothelial cells are joined by tight and adherens junction proteins that restrict paracellular transport, and the barrier exhibits high proteolytic activity — so a peptide faces an exclusion problem and a degradation problem at the same interface. Passive transcellular diffusion favours lipophilic molecules with few hydrogen-bond donors and acceptors; short hydrophilic sequences carrying charged residues sit at the wrong end of that scale. The remaining routes are active — receptor-, transporter- and adsorptive-mediated transcytosis — and each requires a specific interaction that has to be demonstrated for a specific molecule, not assumed for a class.

This is why so much of the work above uses the intranasal route: material deposited on the olfactory epithelium can in principle reach the brain along olfactory and trigeminal nerve pathways without passing the vascular barrier at all. That pathway is real. It is also more complicated than the shorthand "bypasses the blood-brain barrier" suggests.

What has actually been measured for this class. Shevchenko and colleagues gave tritium-labelled Semax intranasally to rats and found radioactivity in brain tissue within two minutes, reaching 0.093% of the total introduced radioactivity per gram of brain, of which about 80% was intact Semax and the rest metabolites. That measurement establishes two things: label reaches brain quickly, and most of what is there at that early timepoint is still intact peptide. It does not separate direct nose-to-brain transport from nasal absorption into blood followed by uptake — the study does not make that distinction, and the two routes operate simultaneously.

The 2025 review by Drath, Richter and Feja is the most careful published account of what this route does and does not establish. The authors state that the exact transport mechanisms "are not fully understood." Olfactory transport can be fast — peak cerebrospinal fluid concentration for intranasal phenytoin at five minutes, a GLP-2 derivative in the trigeminal principal sensory nucleus at three — but the nasal mucosa is well vascularised, so material inevitably enters the systemic circulation too, which complicates attributing any brain measurement to the direct route.

The species problem is larger still. In rats and mice the olfactory epithelium makes up 40–50% of total nasal surface; in humans it is under 10%. One computational model cited in that review predicted nasally administered nanomaterials reach the mouse brain in an amount two orders of magnitude greater than that reaching the human brain, and the review concludes that extrapolating pharmacokinetics from laboratory animals to humans "is of limited validity."

The Pinealon literature is where this gap is widest. The reviews describe a tripeptide entering cells, then nuclei, then engaging specific promoter sequences — and the transport mechanism that would get a charged tripeptide from a peripheral administration site to a neuronal nucleus in vivo is not specified in the material reviewed here. The claim is not disproven; it is unaddressed, which is a different thing, and the distinction matters when reading any summary that presents the mechanism as settled.

What they have in common at the bench

Two points follow from their shortness. First, molar accounting: a fixed mass of a tripeptide contains far more molecules than the same mass of a nonapeptide, so cross-compound reasoning belongs on a molar basis rather than a milligram one. Second, identity verification carries more weight for obscure sequences than for familiar ones, because a short unusual sequence has fewer reference points in the analytical literature — which makes mass-spectrometric confirmation on the certificate of analysis correspondingly more useful. The COA testing guide covers what to look for; general handling for lyophilised material is in the storage guide and the reconstitution guide.

What the evidence does not establish

This section is the reason the guide is worth reading, so it is worth being blunt.

Nothing here establishes anything about cognition, mood, sleep or performance in people. The endpoints cited are transcript counts, transcriptome profiles, dendritic spine morphology, infarct volume and rodent motor tests — molecular and morphological measures in animals and cell lines. They are not clinical outcomes, and they do not become them by being restated in plainer language.

Gene expression is not protein, and protein is not function. The Semax neurotrophin work and the Selank GABAergic work both measure mRNA. The inferential distance from "a transcript moved at one hour" to "a pathway was functionally modulated" is substantial, and none of the studies cited here closes it.

The Semax clinical record is small and unreplicated outside Russia. The one meta-analysis retrieved could include three trials totalling 181 participants out of eight screened, and its authors' own recommendation is for adequately powered multicentre double-blind work. The independent evaluation concludes that well-conducted published studies are lacking and that much of the underlying material is inaccessible in English. Publication-language concentration is a form of selection bias, and citation count does not resolve it.

Selank produced no significant effect on its own in the cleanest system it was tested in. In IMR-32 cells expressing functional GABA_A receptors, Selank alone changed nothing measurable across the analysed gene panel. That negative result is part of the record and belongs alongside the positive rodent findings.

DSIP fails the standard tests for an endogenous factor. No precursor gene, no receptor, a short in-vitro half-life, conflicting sleep results, and a sequence match to a hypothetical bacterial protein that has never been ruled out. Results obtained with synthesised analogues are evidence about those analogues. Fifty years of work has not settled whether the native nonapeptide is a sleep factor at all.

The Pinealon mechanism is a modelled hypothesis from one research school. Its own authors describe it as assumed rather than demonstrated and state that it requires experimental confirmation. In the one head-to-head experiment retrieved, EDR was the weaker of two tripeptides and its effect varied by sex. No independent replication outside the originating group was located.

The route problem is unresolved for all four. Nothing retrieved establishes that intranasal delivery in a human would produce the brain exposure it produces in a rodent, and the difference in relative olfactory surface area predicts that it would not. Where systemic routes were used instead, no study cited here demonstrates that the intact peptide crossed the barrier.

No head-to-head comparison of these four exists. The table above is a survey of four separate literatures assembled by us for this guide, not a synthesis of any published comparison. It cannot be used to rank them, and it is not intended to.

The peptidase-resistance premise is weaker than usually stated. The Pro-Gly-Pro tail is described in the primary literature as included to confer resistance, but the one degradation study retrieved found the C-terminal Gly-Pro was itself among the predominant cleavage products. The tail buys time; it does not confer stability.

Frequently asked questions

What do Semax, Selank, DSIP and Pinealon have in common? Very little beyond the compartment they are studied in. They come from four unrelated research programmes, share no receptor family and no structural motif except shortness. Semax and Selank do share one design feature: both are natural fragments extended with a C-terminal Pro-Gly-Pro tail, a modification introduced by the same Russian institute to slow peptidase cleavage.

Is Semax the same thing as ACTH? No. Semax is Met-Glu-His-Phe-Pro-Gly-Pro — the ACTH(4–7) fragment joined to a synthetic Pro-Gly-Pro tail. It is a short synthetic analogue derived from the ACTH(4–10) region, not the intact hormone, and published work on it addresses gene-expression and transcriptome endpoints rather than adrenal signalling.

Why is DSIP's identity as a sleep factor contested? Because the usual evidence that a molecule is an endogenous signalling factor is missing. No precursor gene and no receptor have been identified, the peptide is reported to have a short in-vitro half-life against aminopeptidase activity, and published sleep findings conflict — some report slow-wave associations, others report none. Researchers active in the field have described it as an unresolved riddle.

What is a peptide bioregulator? A research concept associated with the Khavinson school in St Petersburg: the proposal that very short peptides of two to seven residues, several of them fragments of tissue-derived polypeptide preparations, act not as classical receptor ligands but by entering cells and nuclei and interacting with DNA and histones at specific promoter sequences. Pinealon (Glu-Asp-Arg) is described in that literature as a fragment of the polypeptide preparation Cortexin. The model rests substantially on molecular modelling and on work from one research group.

Does intranasal administration bypass the blood-brain barrier? Partly, and less cleanly than the phrase implies. Olfactory and trigeminal pathways are real and can be fast, but the nasal mucosa is well vascularised, so systemic absorption occurs alongside and confounds attribution. Published reviews state the exact mechanisms are not fully understood, and relative olfactory surface area in rodents is several times that in humans, which limits how far rodent pharmacokinetics extrapolate.

Is it approved for human use? No. None of these four peptides is an approved drug in the United States, and nothing in this guide describes use in a person. All Patriot Labs products are sold strictly for in-vitro research and laboratory use only, and are not for human or veterinary consumption.

All Patriot Labs products are sold strictly for in-vitro research and laboratory use only. Not for human or veterinary consumption. This guide is educational and describes peptide chemistry and published research in general terms; it is not medical advice, does not describe how to use any product, and the references cited do not constitute a product claim.