evipedia.ai
← Back

Dihexa for Cognitive Enhancement

Evidence Review created on 10/04/2026 using AI4L / Opus 5.5

Also known as: PNB-0408, MM-201, ATH-1001, Dihexa Acetate, N-hexanoic-Tyr-Ile-(6) aminohexanoic amide

Motivation

Dihexa is a small synthetic compound built from a fragment of a natural messenger molecule found in the brain. It was designed to survive digestion and reach the brain after oral administration. It is thought to strengthen a natural growth signal that helps nerve cells form new connections, which is why it has drawn interest as a way to sharpen memory and learning.

The compound was created at a university laboratory in the United States in the 2000s as a candidate treatment for dementia and was studied in laboratory animals. Only an injectable precursor that the body converts into dihexa has been tested in people, yet dihexa itself is now sold online and through some clinics as a brain-boosting compound, usually as a capsule or a cream. In 2025, a journal withdrew several papers from the developers’ laboratory because of image manipulation, which has drawn fresh scrutiny to the research behind it.

This review examines what is known about dihexa for improving memory and thinking in health-conscious adults: the animal and related human evidence, the known and theoretical risks, how it is used in practice, and how the research-integrity questions bear on the evidence.

Benefits - Risks - Protocol - Conclusion

A short selection of in-depth sources on dihexa and the hepatocyte growth factor (HGF, a natural growth protein that supports nerve-cell survival and connections) pathway it targets.

Only four items are listed: no other source met the depth and eligibility criteria, and most remaining dihexa content online consists of vendor or clinic marketing.

Andrew Huberman’s content mentions dihexa only in passing: in the October 2024 Huberman Lab episode with Dr. Craig Koniver, it is named as a peptide said to raise a brain growth protein that the FDA (US Food and Drug Administration) removed from pharmacy compounding (see the FDA bulk-substance safety list); this brief mention lacks the depth required for inclusion. No dihexa content was found from Rhonda Patrick, Peter Attia, Chris Kresser, Life Extension Magazine or Lifespan.io; web and on-site searches of each platform returned no article, episode or post discussing it.

Grokipedia

Dihexa

AI-written overview of chemistry, rat pharmacokinetics (absorption, distribution and clearance) and animal findings; it omits the 2025 retractions and misstates some trial registry numbers, so claims need checking against primary sources.

Examine

No Examine article on dihexa exists. Examine does not typically cover unapproved research compounds.

ConsumerLab

No ConsumerLab article on dihexa exists. ConsumerLab does not typically cover unapproved research compounds.

Systematic Reviews

One systematic review covers dihexa, and only as part of a broader class of animal studies.

No systematic review or meta-analysis addresses dihexa’s principal risk, theoretical tumor promotion, or any human outcome; the risk side is unrepresented.

Mechanism of Action

  • Origin: Dihexa is a chemically capped three-amino-acid fragment of angiotensin IV, a breakdown product of the blood-pressure hormone angiotensin II, modified to resist digestion and cross the blood–brain barrier.
  • Proposed primary pathway: The developers proposed dihexa binds HGF and helps it activate MET (Wright & Harding, 2015), driving growth of dendritic spines (small protrusions where nerve cells connect) and new synapses, which their 2013 study in rat neurons showed without testing HGF (McCoy et al., 2013). The developers founded M3 Biotechnology (later Athira Pharma); the 2013 paper carries a 2021 expression of concern.
  • Mechanism evidence after retraction: The key paper tying dihexa’s memory effects to HGF was retracted in 2025 (retraction notice). Independent researchers found a MET blocker abolished dihexa’s memory effect in rats (Martino et al., 2025).
  • Competing explanations: Older work attributed angiotensin IV effects to the AT4 receptor (an angiotensin IV binding site), which another group identified as IRAP (insulin-regulated aminopeptidase, a peptide-trimming enzyme), per the developers’ review (Wright et al., 2015); a third group attributed dihexa’s effects to anti-inflammatory PI3K/AKT signaling (Sun et al., 2021).
  • Pharmacology: In rats, half-life was 12.68 days after intravenous dosing, with extensive tissue distribution, brain entry and little breakdown by liver CYP enzymes (the main drug-processing enzymes) (McCoy et al., 2013). In people, dihexa formed from fosgonimeton, its injected inactive precursor (PubChem structure record), had a plasma half-life of about 1.5 hours in an Athira-run trial (Hua et al., 2022); selectivity beyond HGF is uncharacterized.

Historical Context & Evolution

  • Original intended use: Joseph Harding and Jay Wright at Washington State University studied angiotensin IV and memory from the early 1990s; early analogs worked only when injected into the brain, so dihexa was designed around 2007 as an oral dementia drug candidate (ScienceDaily repost of the university release).
  • Initial findings: The 2012–2013 paper reported reversal of drug-induced and age-related learning deficits in rats (McCoy et al., 2013); the university release claimed dihexa was seven orders of magnitude more potent than BDNF (brain-derived neurotrophic factor, a natural growth protein) in cell assays.
  • Path to health optimization: Oral activity and “synapse-building” claims led biohackers and peptide clinics to adopt it as a cognitive enhancer, mostly as creams or capsules, without human testing.
  • Integrity questions: In 2021, four papers from the laboratory, including the dihexa paper, received expressions of concern, after Athira’s chief executive, a co-author, was placed on leave over concerns that she altered images (Retraction Watch); a company investigation later found altered images (Retraction Watch, 2025). Three were retracted in 2025, including the HGF-mechanism paper (retraction notice); the original dihexa paper remains under concern, not retracted.
  • What changed since: Independent mouse and rat studies reported memory benefits (Sun et al., 2021; Martino et al., 2025), a Huntington’s disease (inherited neurodegenerative disorder) model found none (Wells et al., 2024), and its injectable precursor fosgonimeton failed its main human trial (Porsteinsson et al., 2025). Dihexa’s standing rests on animal data that are partly replicated and partly questioned.

Expected Benefits

High 🟩 🟩 🟩

No benefit reaches High: the only randomized human trials gave fosgonimeton, an injected precursor converted to dihexa in the blood, and found no significant cognitive benefit.

Medium 🟩 🟩

No benefit reaches Medium: no human outcome data exist for oral or skin-applied dihexa, and trials of its injected precursor count only as indirect evidence.

Low 🟩

Improved Memory and Cognition ⚠️ Conflicted

Indirect: trials used fosgonimeton, an injected precursor of dihexa. Dihexa reversed memory deficits in drug-impaired and aged rats and injured rats, but not in a Huntington’s model. Athira’s fosgonimeton improved P300 (a brain-wave speed marker) but failed the LIFT-AD trial. Healthy-adult claims are vendor-linked anecdotes. Net: unproven in humans.

Magnitude: In LIFT-AD, the difference from placebo at week 26 was −0.70 points on ADAS-Cog11 (a 0–70 cognitive impairment scale; standard error, a measure of uncertainty, 0.77; p = 0.35, meaning the difference is consistent with chance) and −0.08 on the GST (a combined cognition and daily-function score; standard error 0.10; p = 0.70); no 95% confidence interval (the range likely to contain the true effect) was reported.

Speculative 🟨

Increased Synapse Formation

Cell-culture and animal data only. Dihexa increased dendritic spines and synapses in rat memory-region neurons at extremely low concentrations (McCoy et al., 2013); the key confirmatory paper was retracted.

Reduced Brain Inflammation ⭕️ Not Central to Cognitive Enhancement

Animal data only. In Alzheimer’s model mice, dihexa lowered inflammatory signals and activation of glia (brain support cells) (Sun et al., 2021). This bears on neurodegeneration rather than enhancing healthy cognition.

Motor Recovery in Parkinson’s Models ⭕️ Not Central to Cognitive Enhancement

Animal data only, summarized in the developers’ review of rat Parkinson’s models (Wright et al., 2015). This bears on movement disorders, not cognition.

Hearing Hair-Cell Protection ⭕️ Not Central to Cognitive Enhancement

Animal data only. Dihexa protected zebrafish sensory hair cells from antibiotic toxicity in a study co-authored by the developers (Uribe et al., 2015). This bears on drug-induced hearing loss, not cognition.

Peripheral Nerve Repair ⭕️ Not Central to Cognitive Enhancement

Animal data only. With stem cells, dihexa injected into muscle improved rat limb function after sciatic nerve repair; a developer co-authored the study (Weiss et al., 2021). This bears on nerve injury, not cognition.

Benefit-Modifying Factors

  • Genetic polymorphisms: In LIFT-AD, carriers of APOE ε4 (a gene variant that raises Alzheimer’s risk) showed larger, non-significant differences with fosgonimeton (Porsteinsson et al., 2025); no genetic data exist for dihexa itself.
  • Baseline cognitive status: Dihexa was tested almost only in impaired animals (drug-induced amnesia, old age, brain injury, Alzheimer’s models); any benefit in cognitively healthy adults may be smaller or absent.
  • Baseline biomarkers: No biomarker predicting response is known. In LIFT-AD, participants with more advanced impairment showed larger, non-significant differences from placebo with the injected precursor (Porsteinsson et al., 2025).
  • Sex: No sex difference appeared in the injured-rat study (Martino et al., 2025); the aged-rat study used both sexes without separate analysis; no human data exist.
  • Pre-existing conditions: Benefits appeared in models of drug-induced amnesia and aging (McCoy et al., 2013), Alzheimer’s disease (Sun et al., 2021) and concussion-like injury (Martino et al., 2025), but not in a model of Huntington’s disease (Wells et al., 2024).
  • Age: 24-month-old rats improved at the same oral dose as young rats (McCoy et al., 2013); whether this applies to adults over 60 is unknown.

Potential Risks & Side Effects

High 🟥 🟥 🟥

No risk reaches High: human adverse-event data come only from trials of dihexa’s injected precursor, with no trial, cohort or case report of oral or skin-applied dihexa.

Medium 🟥 🟥

No risk reaches Medium: the only human safety data come from dihexa’s injected precursor, which counts as indirect evidence.

Low 🟥

Injection-Site Reactions and Adverse-Event Discontinuation

Indirect: seen with fosgonimeton, an injected precursor converted to dihexa, given by daily subcutaneous (under-the-skin) injection. In LIFT-AD, adverse-event discontinuations were mostly injection-site reactions; serious events were balanced (Porsteinsson et al., 2025); ACT-AD, SHAPE and a long-term open-label extension posted similar reactions. Relevant only to injected dihexa.

Magnitude: Discontinuation due to adverse events was 14.2% with fosgonimeton versus 4.6% with placebo; serious adverse events were 4.2% versus 6.9%.

Eosinophilia and Allergic-Type Reactions

Indirect: seen with dihexa’s injected precursor fosgonimeton. In LIFT-AD, eosinophilia (a rise in allergy-related white blood cells) appeared only with fosgonimeton, and angioedema (deep swelling of skin or airways) was among drug-related serious events (Porsteinsson et al., 2025); ACT-AD and the open-label extension also posted eosinophilia.

Magnitude: Eosinophilia occurred in 7.1% with fosgonimeton 40 mg and 7.5% with 70 mg versus 0% with placebo; serious angioedema suspected to be drug-related occurred in 2 participants on fosgonimeton versus none on placebo.

Tumor Promotion

Indirect: a cancer review ties sustained MET activation to tumor growth. In the open-label extension of fosgonimeton, dihexa’s injected precursor, 4 of 423 participants had serious cancers; causality is unknown without controls. Developer safety studies reportedly found no tumors (blog post); regulators found no human data (regulator’s safety list).

Magnitude: 4 of 423 participants had serious cancer events (colon, kidney, lung, prostate) over up to 173 weeks of fosgonimeton; no control group.

Speculative 🟨

Contaminated or Mislabeled Products

Product testing only, with no human outcome data. Online cognitive-enhancement supplements held undeclared unapproved drugs and inaccurate doses (Cohen et al., 2021). Research-chemical dihexa lacks pharmaceutical quality control, making similar problems plausible.

Accumulation with Repeated Dosing ⚠️ Conflicted

Rat plasma half-life was 12.68 days (McCoy et al., 2013). In a randomized human trial, dihexa formed from injected fosgonimeton cleared within 24 hours, without accumulation (Hua et al., 2022). Net: accumulation appears unlikely.

Headache, Fluid Retention and Attention Changes

Isolated user reports compiled in a practitioner blog describe headache, fluid retention, hyperfocus and attention problems (blog post). No controlled data exist, and causality is unverified.

Risk-Modifying Factors

  • Genetic polymorphisms: Inherited activating MET mutations or known cancer-predisposition syndromes would theoretically amplify the tumor-promotion concern; no genetic data on dihexa safety exist.
  • Baseline biomarkers: Abnormal liver or kidney tests before use make any new abnormality hard to attribute; no biomarker predicts dihexa toxicity.
  • Sex: No sex-specific risk data exist. Pregnancy carries unknown risk because growth-factor signaling is active in fetal development and no reproductive data exist.
  • Pre-existing conditions: Current or past cancer, or precancerous lesions, raise the theoretical tumor-promotion concern; skin sensitivity may matter for injected or topical use.
  • Age: Cancer incidence rises with age, so the theoretical tumor-promotion concern weighs more for adults over 60, who also carry more undiagnosed early cancers.

Key Interactions & Contraindications

No dedicated human interaction study of dihexa exists; every interaction below is inferred from mechanism except the entry on dementia drugs that block acetylcholine breakdown, which rests on trial subgroup data with its injected precursor.

  • MET inhibitors (cancer drugs that block the receptor dihexa is proposed to activate; capmatinib, tepotinib, crizotinib, cabozantinib): Avoid (theoretical). They would cancel dihexa’s effect, and their use implies a cancer in which MET activation is undesirable.
  • Other cancer therapies (chemotherapy such as cisplatin, targeted drugs such as osimertinib, immunotherapy such as pembrolizumab): Avoid (theoretical). Growth-factor activation could counteract tumor control; coordination with the treating oncologist applies.
  • Acetylcholinesterase inhibitors (dementia drugs that block acetylcholine breakdown; donepezil, rivastigmine, galantamine): Caution. In trials of dihexa’s injected precursor, these drugs appeared to weaken effects on cognition and blood markers, prompting LIFT-AD to stop enrolling users (Porsteinsson et al., 2025); combined use may blunt benefit.
  • Renin-angiotensin drugs (blood-pressure medications: lisinopril, losartan): Monitor (theoretical). Dihexa derives from an angiotensin fragment; effects on blood pressure or these drugs are unstudied, so home blood-pressure checks after starting detect unexpected changes.
  • Over-the-counter medications: No interaction identified. Caution (theoretical) with over-the-counter stimulants (caffeine tablets) used in focus stacks, because combined effects on sleep, heart rate and anxiety are untested; separating introductions by a week aids attribution.
  • Supplements with additive neurotrophic aims (lion’s mane, Hericium erinaceus; Semax; other “synaptogenic” peptides): Monitor (theoretical). Additive growth-signal stimulation is untested and could amplify unknown effects; adding one agent at a time allows attribution of benefits and adverse events.
  • Other interventions (growth-factor therapies such as becaplermin or mecasermin, other peptide injections): Monitor (theoretical). Overlapping growth-factor stimulation could add to the tumor-promotion concern; spacing courses apart limits overlap.

Populations who should avoid Dihexa:

  • People with current or past cancer or known precancerous conditions (theoretical, MET activation); trials of its injectable precursor fosgonimeton excluded a malignant tumor within 3 years before screening (LIFT-AD record)
  • People receiving MET inhibitors or other cancer therapy (theoretical)
  • People with known inherited MET-activating mutations or cancer-predisposition syndromes (theoretical)
  • Pregnant or breastfeeding women (no reproductive data; the FDA found no human exposure data, FDA bulk-substance safety list)
  • Children and adolescents (no data in developing brains)

No dihexa-specific source gives numeric thresholds; the 3-year malignancy window comes from related-compound trial exclusion criteria.

Risk Mitigation Strategies

Doses and timings below follow common practice unless cited.

  • Cancer screening before use: Completing age-appropriate screening (colonoscopy, mammography, skin check) before starting reduces the chance of stimulating an undetected tumor, addressing the tumor-promotion risk.
  • Short, defined courses: Limiting use to 2–4 week cycles with equal or longer breaks caps cumulative exposure, addressing the tumor-promotion concern and the accumulation suggested by rat data.
  • Low starting dose: Starting at the bottom of the practitioner range (8 mg daily, blog post) limits exposure while tolerance is assessed, addressing headache and overstimulation reports.
  • Independent lot testing: Requiring an independent laboratory certificate (HPLC, a purity test, plus mass spectrometry for identity) for each batch reduces the contamination and mislabeling risk.
  • Avoiding injection of unregulated product: Non-sterile research chemicals given by subcutaneous injection risk infection and injection-site reactions; non-injected forms avoid this risk.
  • Stop rules: No source gives a dihexa-specific stop threshold; a new lump, unexplained weight loss, or abnormal liver or kidney results prompt stopping and medical evaluation, addressing tumor and organ-toxicity concerns.
  • Symptom diary: Logging headache, sleep, attention and mood daily detects user-reported adverse effects early and separates them from expectation effects.

Therapeutic Protocol

Timing, cycling and titration parameters without a citation reflect common practice.

  • No tested human dose: No human trial has tested oral or skin-applied dihexa; doses below come from practitioners, animal studies or an injected precursor, not clinical evidence of benefit.
  • Practitioner range: 8–45 mg daily, compiled from user reports by Jay Campbell, who promotes a vendor (blog post); transdermal cream on the inner forearm is the route he reports as most popular.
  • Clinic transdermal use: Daniel Stickler is quoted in the same blog post prescribing dihexa cream for learning goals; no dose is given.
  • Developer animal dose: 2 mg/kg daily by oral gavage (tube feeding) fully reversed drug-induced deficits in rats (McCoy et al., 2013); conversion to a human dose is not validated.
  • Precursor human dose: Fosgonimeton 40 mg by daily subcutaneous injection in LIFT-AD (Porsteinsson et al., 2025) converts in plasma to ATH-1001 (Hua et al., 2022); fosgonimeton is phosphate-tagged dihexa (PubChem structure record); equivalent oral or skin doses are unknown.
  • Time of day: In animals, dosing preceded learning sessions by minutes (McCoy et al., 2013); morning dosing is common practice to avoid possible sleep disruption.
  • Half-life: Rat plasma half-life was 12.68 days intravenously (McCoy et al., 2013); in humans, dihexa formed from injected fosgonimeton had a plasma half-life of about 1.5 hours (Hua et al., 2022).
  • Single versus split dosing: A single daily dose is common practice and was used in all fosgonimeton trials (Porsteinsson et al., 2025); no study has compared single and split schedules.
  • Genetic polymorphisms: APOE ε4 status may modify response, based on a subgroup trend with the injected precursor (Porsteinsson et al., 2025); no genotype-guided dosing exists.
  • Sex differences: Fosgonimeton showed no sex effect on blood levels (Hua et al., 2022); no sex-specific dihexa dosing exists.
  • Age: Aged rats responded to the same dose as young rats (McCoy et al., 2013); older adults face a higher baseline cancer risk, which weighs on any decision to use.
  • Baseline biomarkers: A baseline cognitive test score anchors any later judgment of response; no biomarker guides dose.
  • Pre-existing conditions: Cancer history excludes use (theoretical); liver or kidney disease adds uncertainty because clearance in humans is unknown.

Discontinuation & Cycling

  • Short-term use: User reports describe roughly 14-day cycles (blog post); no data support long-term or lifelong use.
  • Withdrawal effects: None are known; discontinuation has never been studied.
  • Tapering: No tapering protocol exists or appears needed; in humans, dihexa formed from injected fosgonimeton cleared from blood within 24 hours (Hua et al., 2022).
  • Cycling: Cycles of 2–4 weeks with equal or longer breaks are common practice; no evidence shows cycling preserves efficacy, and the rationale is limiting cumulative exposure.

Sourcing and Quality

  • No pharmaceutical-grade product: Dihexa is not an approved drug; it is sold mainly as a “research use only” chemical without good manufacturing practice (GMP, the regulated production standard) oversight.
  • Compounding status: FDA once placed dihexa acetate in Category 2 (significant safety risks); its nomination is now withdrawn (FDA bulk-substance safety list). FDA’s interim policy covers only Category 1, so removal confers no compounding eligibility (FDA 503A policy; Frier Levitt legal summary).
  • What to look for: A lot-specific certificate from an independent laboratory showing purity (HPLC), identity (mass spectrometry), and, for anything injected, sterility and endotoxin (bacterial toxin) testing.
  • Formulations: Capsules, powder and transdermal creams are sold; oral absorption was shown only in rats, and skin absorption has never been measured.
  • Brands: No brand or pharmacy can be named as reputable, because no product has published independent testing; vendor endorsements from influencers often carry affiliate discount codes.

Practical Considerations

  • Time to effect: Rats improved within 3–8 days of daily dosing during training (McCoy et al., 2013); the human time course is unknown.
  • Common pitfalls: Treating rodent results or press claims (“seven orders of magnitude more powerful than BDNF”, university release) as human evidence, stacking several compounds at once, and overlooking the retractions behind the mechanism.
  • Regulatory status: Dihexa is not approved by the FDA for any use; FDA found no human exposure data (FDA bulk-substance safety list).
  • Access: Legal pharmaceutical access is essentially absent; supply comes from unregulated online vendors, which shifts quality risk onto the user.
  • Funding incentives: Dihexa is cheap to make and has no commercial sponsor or insurer coverage, so neither manufacturers nor payers have a financial incentive to fund human trials; the evidence gap partly reflects this structure.

Interaction with Foundational Habits

  • Sleep: Direction unknown; no study exists. Synapse strengthening is consolidated during sleep, so sleep loss could plausibly blunt any benefit; morning dosing is common practice to avoid possible stimulation at night.
  • Nutrition: No known interaction; food effects on absorption are untested, and rats received dihexa as a saline suspension (McCoy et al., 2013). No nutrient depletion is known.
  • Exercise: Possibly potentiating; exercise raises BDNF, another growth protein supporting synapses, but no study has combined exercise with dihexa, and no blunting effect is known.
  • Stress management: None known; no data exist on cortisol or stress responses. Anti-inflammatory effects seen in mice (Sun et al., 2021) have not been linked to stress physiology; logging stress and sleep alongside cognitive tests helps separate stress effects from any compound effect.

Monitoring Protocol & Defining Success

Before starting, a baseline record establishes reference points: a validated cognitive test, a liver panel including ALT (alanine aminotransferase, a liver enzyme), a kidney panel including creatinine, seated blood pressure, and completion of age-appropriate cancer screening. Because no human safety data exist for oral or skin-applied dihexa, these tests are safety checks rather than validated markers of benefit, and the cognitive test is the only marker the intervention is expected to change.

Ongoing monitoring follows common practice: cognitive testing and blood pressure at 2–4 weeks and at the end of each cycle, and liver and kidney panels at the end of each cycle, then every 6–12 months during repeated cycles. Any new lump, unexplained weight loss or abnormal result is a reason to stop and seek evaluation. Success is defined as a consistent improvement on the same cognitive test, beyond practice effects, without adverse findings.

Biomarker Optimal Functional Range Why Measure It? Context/Notes
ALT 7–56 U/L (standard reference range) Safety check: unknown liver effects; abnormal stops use Ranges vary by laboratory; fasting not required; usually paired with AST (aspartate aminotransferase, another liver enzyme)
Creatinine 0.6–1.3 mg/dL (standard reference range) Safety check: unknown kidney effects; abnormal stops use Ranges vary by laboratory and sex; usually paired with eGFR (estimated glomerular filtration rate, a kidney-function estimate)
Blood pressure No established target; track change from own baseline Safety check: angiotensin-derived compound with unknown effect Seated after 5 minutes’ rest, same time of day, average of two readings
Cognitive test score No established target; track change from own baseline Expected to change: the outcome of interest The same validated test (e.g., MoCA, Montreal Cognitive Assessment, or a computerized battery), taken at the same time of day, keeps results comparable; repeat testing inflates scores

Qualitative markers:

  • Memory in daily tasks (names, appointments, recall of reading)
  • Focus and ability to sustain demanding work
  • Mental fatigue and stamina
  • Sleep quality and onset
  • Headache, mood changes or restlessness

Emerging Research

  • No ongoing dihexa trials: As of October 4, 2026, ClinicalTrials.gov lists no trial of dihexa itself and no recruiting or active trial of its precursor or any related HGF modulator for cognition; for health-focused adults, no human dihexa data are expected soon.
  • LIFT-AD, completed: Phase 2/3 trial of fosgonimeton in mild-to-moderate Alzheimer’s disease (NCT04488419), 554 enrolled, completed 2024; no significant benefit, which the authors attribute partly to a mostly mild population whose placebo group did not decline over 26 weeks (Porsteinsson et al., 2025).
  • ACT-AD, completed: Phase 2 fosgonimeton trial (NCT04491006), 77 enrolled, primary endpoint P300 latency, completed 2022; posted efficacy results give baseline measures only, with no between-group comparison; posted adverse events, led by injection-site reactions and eosinophilia, are reported in the Risks section.
  • SHAPE, terminated: Phase 2 fosgonimeton trial in Parkinson’s disease dementia or dementia with Lewy bodies (a dementia with movement problems and hallucinations) (NCT04831281), 28 enrolled, terminated 2023; posted efficacy results give baseline measures only; injection-site reactions appear in the Risks section; falls were similar with placebo (3 of 9) and fosgonimeton (6 of 19).
  • Open-label extension, terminated: Phase 2/3 extension giving fosgonimeton 40 mg for up to 173 weeks after LIFT-AD (NCT04886063), 423 enrolled, terminated October 2024 after LIFT-AD failed; posted adverse events, without a control group, include injection-site reactions (213 of 423), eosinophilia (20 of 423) and 4 serious cancers, reported in the Risks section.
  • ATH-1105 phase 1, completed: Oral single and multiple ascending-dose safety study (NCT06432647) of ATH-1105, an Athira HGF modulator developed for ALS (amyotrophic lateral sclerosis, a motor-neuron disease) (Berthiaume et al., 2024); 80 enrolled, completed November 2024; no results posted.
  • Independent mechanism replication: A MET blocker abolished dihexa’s memory effect in injured rats (Martino et al., 2025), partly restoring support for the retracted mechanism and strengthening the case.
  • Negative animal data: Dihexa failed to protect rats in a Huntington’s model induced by a toxin that disrupts cellular energy production, unlike the memory-deficit models in which it worked (Wells et al., 2024), weakening claims of broad neuroprotection.
  • Open questions: Human pharmacokinetics of oral or skin-applied dihexa, chronic toxicology and cancer-safety data are absent; such data would change the risk picture more than further rodent memory studies.

Conclusion

Dihexa is a synthetic fragment of a natural brain messenger, designed to reach the brain after oral use and to strengthen a growth signal that helps nerve cells form connections. For health-focused adults seeking sharper memory and learning, its appeal rests on animal results: memory loss caused by drugs, old age or brain injury was reversed in rodents, and independent laboratories have partly confirmed these findings, though one animal model showed no benefit.

No human study has tested dihexa as it is sold. The only human trials injected a precursor that the body converts into dihexa; it was well tolerated apart from injection reactions and occasional allergic-type reactions but did not improve thinking or daily function in people with Alzheimer’s disease. Claims of benefit in healthy adults come from personal reports alone.

The risks are mostly unknown rather than disproven. The growth signal dihexa amplifies also drives many cancers, dihexa lingers for days in animals but cleared within a day in people, and products come from unregulated sellers with no quality guarantees.

The evidence base carries notable conflicts of interest. Most early studies came from the inventors, who founded the company that developed its injectable precursor, and several of their papers were withdrawn after images were found to be altered; the main human trials were company-run, and much practical guidance comes from promoters who sell or recommend products. Overall, dihexa remains an animal-stage compound whose promise for cognitive enhancement in healthy people is untested and whose long-term safety is unknown.

Top - Benefits - Risks - Protocol