Sleep and cognition peptides: a complete guide

What sleep and cognition peptides are

Under the labels "nootropic peptides" and "sleep peptides" circulates a small group of molecules that share four traits: they are very short chains, three to nine amino acids long; they are credited with an effect on the central nervous system; they degrade in plasma within minutes; and their clinical literature is thin and concentrated in a small number of journals. The four most searched are DSIP (delta sleep-inducing peptide, international name emideltide), Semax, Selank and Pinealon.

Their origin explains much of what follows. DSIP was isolated in Basel in the 1970s from rabbit cerebral venous blood¹². Semax and Selank came out of Soviet and Russian academic programs; Semax was registered as a medicine in Russia and is available there as 0.1% and 1% nasal drops¹⁰. Pinealon is a tripeptide from the St. Petersburg group that works on short regulatory peptides²³²⁵.

For decades that evidence was never examined systematically outside its country of origin. That changed in July 2026, when the FDA's Pharmacy Compounding Advisory Committee (PCAC) evaluated seven peptides for the 503A list and published detailed briefing documents on two of the ones that matter here: Semax and emideltide³¹⁰²⁶. Those documents are today the most complete independent review available, and they are cited throughout this guide.

DSIP: the nonapeptide with no known receptor

DSIP is a nonapeptide with the sequence Trp-Ala-Gly-Gly-Asp-Ala-Ser-Gly-Glu. In the experiment that characterized it, its discoverers synthesized the peptide, five possible metabolic products, two analogs with exchanged amino acids and a related tripeptide, and infused them into the cerebral ventricles of rabbits under double-blind conditions. A total of 58 animals, including controls, were evaluated, and only synthetic DSIP specifically enhanced delta waves and spindles on the electroencephalogram¹.

The problem appears when you look for the mechanism. A 2006 review put the lack of progress down to the fact that the DSIP gene, its protein and a possible related receptor had never been isolated; it titled the paper plainly "a still unresolved riddle" and described the hypothesis of DSIP as a sleep factor as "extremely poorly documented and still weak"². Two decades later, the FDA still described its sleep-inducing effect as mediated by indirect opioid mechanisms, without direct interaction with opioid receptors³. An awkward finding adds to this: in twelve healthy volunteers, circulating DSIP peaked at 15:00 and bottomed at 01:00, was lower during REM sleep (P < 0.005) and slow-wave sleep (P < 0.05) than during wakefulness, and correlated closely with body temperature (r² = 0.66)⁴. That is not the pattern of a simple sleep hormone.

The pharmacokinetics are equally demanding: in four dogs the plasma half-life was 4.0 ± 0.7 minutes, and 2.9 and 2.0 minutes in a monkey and in rats⁸. Even so, a radiolabeled analog crosses the rat blood-brain barrier by a non-competitive mechanism⁹.

In humans, the two controlled insomnia trials are modest. One, double-blind and crossover, found that the significant differences versus placebo were already present at baseline and concluded that the improvement was "of little clinical significance"⁵. The other, double-blind with 16 patients, associated DSIP with higher sleep efficiency and shorter sleep latency, but the authors warned that the effects were weak and partly attributable to an incidental change in the placebo group, and concluded that short-term treatment "is not likely to be of major therapeutic benefit"⁶. The most striking series is a different one: 107 inpatients withdrawing from alcohol (n = 47) or opiates (n = 60), in which symptoms disappeared or improved markedly in 97% of opiate-dependent patients and 87% of alcoholics evaluated, although between 13% and 22% were not evaluable and there was no control group or blinding⁷. The full account of this molecule is in DSIP: the delta sleep peptide, fifty years on.

Semax: an ACTH fragment with a half-life of minutes

Semax is a heptapeptide: the ACTH(4-7) fragment joined to the Pro-Gly-Pro tripeptide. The FDA describes it as "a registered drug in Russia" available "as 0.1% and 1% nasal drops"¹⁰. In animal models it activates the transcription of neurotrophins and their receptors after cerebral ischemia¹³, and in rats the intranasal kinetics show that at two minutes only 0.093% of the administered radioactivity per gram is found in the brain, of which about 80% is intact Semax; the rest are metabolites, with Pro-Gly-Pro the main one¹⁴.

The published clinical studies are almost all in ischemic stroke and without randomization. In 1997, 30 treated patients were compared with 80 who received conventional therapy¹¹. In 2018, a series of 110 patients (43 men and 67 women) grouped by rehabilitation timing and Semax use reported higher plasma BDNF and better final Barthel index; it is not described as randomized or blinded either¹². Outside stroke, an open study in 27 patients with motor neuron disease concluded that Semax "does not influence either the course of CPD or the dynamics of clinical estimates", although it improved the total quality-of-life score²⁷.

Two neuroimaging studies in healthy volunteers deserve mention because they measure something objective. In one, 24 volunteers (14 on 1% intranasal Semax and 10 on placebo, mean age 43.9 ± 9.5 years) showed a greater volume of the rostral component of the default mode network¹⁵. In another, 52 healthy participants were scanned before and 5 and 20 minutes after receiving Semax, Selank or placebo, and differences in functional connectivity were found between the right amygdala and the right temporal cortex¹⁶. These are physiological signals, not clinical outcomes.

The FDA review is hard on three points. First: "We were not able to find pharmacokinetic studies in humans" by any route of administration. Second: in mice Semax potentiated amphetamine-induced striatal dopamine release, which the agency considers concerning because that increase in dopaminergic tone "is a response typically induced by drugs of abuse such as cocaine". Third: its antithrombotic properties "raise[] concern about the risk of bleeding". On effectiveness, the agency found "only two available references with insufficient details", which moreover demonstrated a lack of effectiveness¹⁰. Even so, the committee voted 8–5 in favor of including it for cerebral ischemia, migraine and trigeminal neuralgia²⁶.

Selank: a tuftsin analog and the GABA hypothesis

Selank is a heptapeptide analog of tuftsin. The mechanism its investigators propose is indirect: in cellular and animal models it changes the expression of genes involved in GABAergic neurotransmission, the same system benzodiazepines act on²⁰. A US pharmacology review groups it exactly there and describes it, together with phenibut, as "poorly studied Russian drugs with GABAergic mechanisms" sold to US consumers as dietary supplements²¹.

The indexed clinical trials share one design limitation: none has a placebo group. In 62 patients with generalized anxiety or neurasthenia, Selank (n = 30) was compared with medazepam (n = 32) and the anxiolytic effect was similar¹⁷. A second study compared Selank with phenazepam in 60 patients with phobic-anxiety and somatoform disorders¹⁸. A third added Selank to phenazepam in 40 patients versus 30 on phenazepam alone, and described an earlier onset of effect on the Hamilton scale and fewer benzodiazepine side effects¹⁹. Comparing against an active drug answers whether an alternative is equivalent, not whether it beats placebo; without that arm, spontaneous improvement and expectation stay inside the result.

Selank was not among the seven peptides the FDA advisory committee evaluated in July 2026²⁶, so there is no independent briefing document for it equivalent to those for Semax or emideltide.

Pinealon and short peptides: cellular and animal evidence

Pinealon is a tripeptide, Glu-Asp-Arg (EDR). Its evidence is of a different kind: in cultures of cerebellar granule cells, neutrophils and PC12 cells under oxidative stress, it dose-dependently restricted the accumulation of reactive oxygen species and reduced necrotic cell death, with a delayed time course of ERK 1/2 activation²³. In rats with gestational hyperhomocysteinemia induced by methionine loading, administration to the mothers was associated with better spatial orientation and learning in the offspring and with less accumulation of reactive oxygen species in isolated cerebellar neurons²⁴.

A 2020 review from the same group sets out the mechanistic hypothesis: the EDR peptide would enter the cell and bind histone proteins or ribonucleic acids, altering the expression of genes such as SOD2, GPX1 or caspase-3²⁵. It is an epigenetic hypothesis, not a demonstrated mechanism, and the review itself rests on earlier studies from the same program. In humans, what is indexed are small geriatric series published in Russian: one compared several geroprotective techniques in 110 people of different ages and concluded that the most pronounced effect on biological-age indicators came from combining Pinealon with another oligopeptide complex, with no randomization and no placebo group²⁸. In the indexed literature there are no randomized, placebo-controlled clinical trials of Pinealon; nor was it part of the FDA's 2026 review²⁶.

How they compare: design, size and comparator

Placed side by side, the four occupy different rungs of the same ladder, and none reaches the top:

• DSIP: it is the only one with double-blind placebo-controlled trials in humans, and those are precisely the trials that found no clear clinical benefit⁵⁶. Its most favorable series, in withdrawal, was open and uncontrolled⁷.
• Semax: it has the largest volume of real clinical use, thanks to its registration in Russia, but its stroke studies are comparative series without randomization or blinding¹⁰¹¹¹², and its cleanest objective data are neuroimaging in healthy volunteers, not outcomes¹⁵¹⁶.
• Selank: it has trials in patients with psychiatric diagnoses, but all against benzodiazepines and none against placebo¹⁷¹⁸¹⁹.
• Pinealon: it has no indexed randomized or placebo-controlled trials; its base is cellular and animal, with the odd open geriatric series^23242528.

There is one further pattern worth reading carefully: three of the four come from the same research ecosystem, and the teams that describe the mechanism are often the same ones publishing the clinical results. That does not invalidate the data, but it does explain why independent replication, in different populations and with different protocols, remains the missing piece.

Safety, characterization and purity

The 2026 regulatory discussion was as much about chemistry as about efficacy. For emideltide, the FDA concluded that it is not well characterized from a physicochemical perspective: inconsistent naming conventions that do not follow established standards, and the absence from the public literature of critical data on impurities, aggregates and endotoxins; for the acetate salt, the certificate of analysis provided "lacked specific tests" for those same parameters, and for the free base none was provided at all³. The agency made equivalent observations about Semax, including the lack of endotoxin data for injectable routes¹⁰.

The immunogenicity argument is the one least discussed outside the briefing documents and the one that carries most weight. A nine-amino-acid peptide "has the potential to be immunogenic", and that risk may be enhanced with injectable routes because of the potential for aggregation and for peptide-related impurities³. The stability, pharmacological activity and immunogenic properties of these peptides are also highly sensitive to the manufacturing process³¹⁰.

As for described adverse effects, the DSIP withdrawal studies reported transient headache, nausea and vertigo as common side effects, with cases of hypotension described as "progressive" after the second intravenous injection; the FDA adds that any adverse effect in that population is confounded by the withdrawal syndrome itself³. For Semax, the bleeding signal from its antithrombotic properties and the potentiation of dopamine release in mice are the two explicit warnings¹⁰.

The product-identity problem is not theoretical. A European analytical paper described two seized pharmaceutical preparations that contained Selank and Semax, peptides that, according to the authors, have not completed any clinical trials, and developed an LC-MS/MS method for ten of these compounds sold online²². In parallel, the pharmacology review cited above documents their sale as dietary supplements²¹. In that channel there is no guarantee that the contents match the label, and no control of impurities or endotoxins.

Open questions and what it would take to answer them

The honest reading of these four peptides is not that they have been refuted, but that most of what is claimed about them was never tested with the designs that would be required. What is missing is concrete:

• A receptor and a mechanism for DSIP. With no gene, no precursor and no receptor identified, any mechanistic explanation remains provisional².
• Human pharmacokinetics for Semax, which the FDA could not locate for any route of administration¹⁰; the only detailed kinetic data are in the rat¹⁴.
• Selank trials with a placebo arm, not only against benzodiazepines¹⁷¹⁸¹⁹.
• A randomized, placebo-controlled trial of Pinealon: what exists in humans are open, uncontrolled series^23242528.
• Published characterization data — impurities, aggregates, endotoxins — for the products actually in use, and not only for the active ingredient described in the literature³¹⁰.
• Independent replication outside the originating groups, with clinical outcomes and not only biomarkers or imaging¹²¹⁵¹⁶.

In the meantime, the practical question for a physician reading a claim about any of these peptides is always the same: what was the study design, what was it compared against, how many subjects were there, and who ran it? In this group of molecules, those four questions almost always separate the claim from the data. The shortest comparative summary of the first three is in Semax, Selank and DSIP: what is behind the nootropic peptides.