Pharmacological Developments in Cognitive Enhancement
Linas JuozenasShare
Intelligence Unleashed · Evidence-led guide
Pills, Profiles and Precision What cognitive pharmacology can—and cannot—do
Medicines can restore function, relieve symptoms and sometimes improve a particular cognitive task. That is not the same as making a healthy person broadly more intelligent. Here is how to separate credible treatment, early research and precision care from the “smart pill” story.
The credible future is narrower—and more useful—than a universal smart pill: define a real clinical problem, choose an intervention for the right population, measure daily function as well as test scores, and keep benefits, harms, consent and privacy visible.
Start with the right question
“Does it improve cognition?” is too vague. A useful question names the ability, the person, the context, the time horizon and the outcome that matters.
Cognition is not one dial. It includes wakefulness, sustained attention, impulse control, working memory, long-term learning, flexibility, motivation and the ability to use knowledge in daily life. A drug can improve one measure, leave another unchanged and make a third worse. It can also improve the performance of someone with a disorder without offering the same benefit to a healthy, rested person.
This is why the language matters. Treatment aims to reduce symptoms or impairment in a defined condition. Restoration aims to recover function lost through illness, injury or sleep disruption. Enhancement aims to move a healthy person beyond their ordinary baseline. These goals can overlap at the edges, but they do not share an automatic evidence base.
A result obtained in adults with ADHD or schizophrenia is evidence about that population and indication—not a shortcut to claims about intelligence in healthy students or workers.
State
How awake, anxious, tired or motivated someone is right now. A temporary state change can alter a test without changing underlying ability.
Skill
What practice has made more fluent or accurate. Faster task completion during drug exposure does not establish learning that transfers later.
Health and function
Whether symptoms, independence, work, study or relationships improve in a way that matters and lasts. This is usually the most consequential level.
Ask which domain actually changed
| Domain | What a study may measure | What the result does not prove |
|---|---|---|
| Wakefulness | Reaction time, lapses or subjective sleepiness | Deeper understanding, creativity or better judgment |
| Attention | Accuracy or persistence on a sustained task | Long-term retention or flexible problem-solving |
| Working memory | Holding and manipulating a small amount of information | Broad intelligence or useful everyday performance |
| Learning | Recall after practice or a delay | Transfer to different material, settings or months later |
| Executive control | Inhibition, switching or planning tasks | Consistently better real-world decisions |
| Daily function | School, work, self-care or social functioning | That the same intervention helps every person or context |
Read the status—and climb the evidence ladder
“Available,” “prescribed,” “in a trial” and “sold online” describe different levels of review. None of them alone tells you how large, durable or relevant a cognitive benefit is.
Defined condition
A regulator has reviewed evidence for a specific product, population and indication. Approval is not a blanket endorsement for other uses.
Clinical judgment
A licensed clinician may prescribe an approved medicine for another purpose where lawful. That purpose has not automatically passed the same approval review.
Question under study
A registered trial tests a hypothesis under a protocol. Participation is not proof of efficacy, and many promising candidates fail.
Marketing is not approval
Supplement and “nootropic” claims can reach the market without the premarket effectiveness review required for medicines in the United States.8
The evidence ladder
No rung guarantees the next. Brain target engagement is not a cognitive outcome; a statistically significant task score is not necessarily a noticeable benefit; and a short-term benefit cannot establish long-term safety. The most credible claim states exactly where on this ladder the evidence ends.
Practice effects
Repeating a cognitive test can improve scores even without an effective drug. Good studies use suitable controls, alternate forms and planned analyses.
Baseline and context
Sleep loss, illness, stress and starting ability can change the direction or size of an effect. Average results can hide wide variation.
Trade-offs
More persistence is not always better judgment. Gains in a narrow task can arrive with insomnia, anxiety, appetite change or overconfidence.
What existing medicines show
Modern pharmacology can meaningfully treat specific disorders. Evidence for turning rested, healthy people into broadly stronger thinkers is far weaker and more domain-specific.
Caffeine offers the familiar lesson: reducing sleepiness can improve vigilance, yet being more awake does not necessarily make an argument more accurate or a memory more durable. Prescription stimulants can reduce ADHD symptoms and improve functioning for appropriately assessed patients. Wakefulness medicines have defined medical uses. Their therapeutic value does not convert them into universal upgrades.
A meta-analysis of randomized trials in healthy, non-sleep-deprived adults found small effects in some cognitive domains for methylphenidate and modafinil, with no overall cognitive effect for d-amphetamine. The pattern was selective rather than a broad rise in intelligence, and evidence for meaningful academic or workplace gains remained limited.4
| Example | Credible use or finding | What should not be inferred | Important context |
|---|---|---|---|
| Caffeine | Can increase alertness and reduce perceived fatigue | Broad intelligence, reliable creativity or compensation for chronic sleep loss | Tolerance, timing, anxiety and sleep disruption can change the net effect |
| Prescription stimulants | Established treatments for ADHD when clinically indicated | That nonmedical use is safe, fair or broadly performance-enhancing | Cardiovascular, psychiatric, sleep, appetite, misuse and dependence risks require professional assessment |
| Wakefulness medicines | Can treat excessive sleepiness in specific diagnosed conditions | That wakefulness equals learning or makes sleep biologically optional | Indication, interactions and individual risks matter |
| Centanafadine | FDA-approved in July 2026 as Simtriyo for ADHD in specified adult and paediatric populations | An indication for “executive-function disorder” or proof of healthy-person enhancement | It is a CNS stimulant with boxed warnings and other labelled risks3 |
| Disease-specific therapies | May slow decline, relieve symptoms or support function in a defined illness | A transferable benefit in people without that illness | Diagnosis, stage, goals and risk tolerance shape clinical value |
Nonmedical sourcing adds a separate danger. Pills obtained outside licensed prescribing and pharmacy systems may be counterfeit or contain unexpected substances. A 2024 CDC health advisory specifically warned about counterfeit-pilling risks and cited prescription-stimulant misuse among college students.11
The centanafadine lesson
A treatment for ADHD
Simtriyo (centanafadine) is approved in the United States for ADHD in adults and in children aged six years and older who weigh at least 20 kilograms. The approval was supported by randomized Phase III trials using ADHD symptom scales.3
A general executive-function upgrade
The approved indication is not “executive function disorder,” and success in people with ADHD does not establish a broad cognitive benefit for healthy users. The correct description keeps population, endpoint and safety label attached.
The experimental pipeline: mechanisms, not miracles
Drug development is a sequence of tests, not a countdown to enhancement. Negative late-stage trials are especially valuable because they show where an appealing biological story did not become a clinical benefit.
Evidence reset: several forecasts once circulated with this article could not be verified in trial registries, regulator records or peer-reviewed literature. The examples below are limited to traceable programmes and their actual status as of September 2026.
Iclepertin
Iclepertin was studied for cognitive impairment associated with schizophrenia—not as a healthy-person enhancer. Across three randomized Phase III CONNEX trials, 1,835 treated participants were included and the drug did not significantly improve the primary cognitive endpoint compared with placebo, although it was generally well tolerated.1
Lesson: strong target rationale and encouraging earlier signals do not guarantee clinically meaningful cognition gains.
TAK-041
TAK-041, also known as NBI-1065846, was developed as a selective GPR139 agonist. It is not a GABAB positive allosteric modulator. Mechanistic and early clinical research should not be represented as proof of a general cognition benefit.2
Lesson: verify the mechanism, trial registry, planned endpoint and result before repeating a pipeline claim.
LSD microdosing
Blinded research does not support presenting microdosing as an established cognitive enhancer. In one six-week randomized trial, participants receiving LSD reported feeling more creative, but objective creativity measures showed no acute or durable benefit; anxiety also contributed to withdrawals.5
Lesson: subjective experience, expectancy and objective performance are different outcomes.
Where research is looking
| Target family | Clinical question | Why it interests researchers | Why caution remains |
|---|---|---|---|
| Monoamine systems | ADHD, mood, fatigue and wakefulness | Dopamine, norepinephrine and serotonin shape attention, motivation and arousal | Benefits and adverse effects depend on condition, baseline and exposure; misuse risk can be material |
| Glutamate and plasticity | Cognitive symptoms across psychiatric and neurological illness | Glutamatergic signalling is central to learning and synaptic change | Complex networks resist simple “more is better” logic; iclepertin shows translation can fail |
| Cholinergic pathways | Memory symptoms in defined disorders | Acetylcholine contributes to attention and memory | Symptom benefit in disease does not imply durable enhancement in healthy people |
| Orexin and arousal | Narcolepsy and disorders of excessive sleepiness | Orexin systems help regulate sleep and wakefulness | Restoring wakefulness is not equivalent to raising intelligence |
| Metabolic, immune and protective pathways | Condition-specific decline or dysfunction | Brain function depends on vascular, immune and metabolic health | Broad mechanisms make surrogate outcomes and causal claims easy to overstate |
| Psychedelic-related plasticity | Primarily psychiatric treatment research | Acute and longer-term changes may interact with psychotherapy and context | Blinding, expectancy, legality, psychiatric risk and uncertain cognitive transfer matter |
A pipeline chart should be dated. Record the registry identifier, sponsor, phase, indication, comparator, primary endpoint and latest result. If a compound, company or programme cannot be traced to primary records, it should not be presented as an imminent breakthrough.
Precision pharmacology is a clinical process—not a DNA horoscope
Useful personalization asks how a particular person, condition and treatment interact. It combines clinical history with evidence; it does not turn a consumer profile into an “ideal nootropic” prescription.
Pharmacokinetics
What the body does to a medicine: absorption, distribution, metabolism and elimination. Age, liver or kidney function, other drugs, smoking and some genetic variants can change exposure.
Pharmacodynamics
What the medicine does to the body and brain: intended effects, adverse effects and the relationship between exposure and response. The same concentration can affect people differently.
Clinical context
Diagnosis, symptoms, sleep, cardiovascular and psychiatric history, pregnancy, substance use, other treatments and personal goals can matter more than a novel data stream.
Follow-through
Precision includes a baseline, an outcome worth improving, safety monitoring, a review date and a plan to continue, adjust or stop. Personalization without follow-up is only segmentation.
What pharmacogenetics can genuinely contribute
For some gene–drug pairs, a genotype can inform metabolism, exposure, adverse-event risk or therapeutic strategy. The FDA’s association table makes two limits explicit: inclusion does not necessarily mean testing is recommended before prescribing, and many listed associations have not been evaluated for whether testing improves clinical outcomes.6
That is a long way from predicting who will gain intelligence from a pill. Most cognitive outcomes are influenced by many genes and many environmental factors. A panel can be analytically accurate yet clinically unhelpful for the claim being made. Direct-to-consumer tests also differ in evidence and coverage, and the FDA advises discussing clinically relevant results with a qualified professional.7
“This variant may affect how this specific drug is processed.”
The claim names a gene–drug pair, evidence source, consequence and role of clinical judgment.
“Your genome reveals your perfect cognitive enhancer.”
The claim jumps from limited associations to a broad outcome, ignores context and usually lacks prospective clinical validation.
Natural, personalized and AI-guided are not safety categories. A product can be natural and harmful, personalized and poorly validated, or algorithmic and confidently wrong. Ask what was measured, how the recommendation was validated and whether it changes care for the better.
AI and digital biomarkers: support the decision, do not automate the person
Wearables and apps can organize observations and reveal patterns. They do not automatically measure brain function, establish causality or justify day-to-day psychotropic dose changes.
Sleep estimates, heart-rate patterns, movement, phone interaction and reaction-time tasks can become useful signals when their measurement properties and context are understood. But a proxy can drift when a device changes, when a person is ill, when routines change or when the training population differs from the user. Missing data are often meaningful too: the people least able to engage may disappear from the dashboard.
Regulators therefore evaluate biomarkers and drug-development tools for a specified context of use. Qualification for one purpose does not validate every use of the same signal.9 An algorithm that supports a clinician’s review is also a different intervention from a consumer service that autonomously changes a prescription.
A safer decision loop
- Organize symptom and side-effect reports.
- Prompt adherence or a scheduled review.
- Compare change with a documented baseline.
- Surface a pattern for clinician and patient discussion.
- Maintain an audit trail of data and decisions.
- Treat a consumer wearable score as a direct neural measurement.
- Assume correlation reveals what caused the change.
- Alter prescriptions autonomously from noisy daily data.
- Hide uncertainty behind one proprietary “readiness” number.
- Use mood or performance data for unrelated workplace decisions.
The personalization paradox
More intimate data can make a service feel tailored while also increasing surveillance and security risk. Genetic data are persistent; mood, sleep and cognitive-performance traces can expose vulnerability; workplace or school collection introduces power imbalances. Even a technically accurate inference may be inappropriate for a different purpose.
Purpose
What exact decision requires the data—and could less sensitive information answer it?
Access
Can an employer, insurer, school, vendor or data broker see or infer from it?
Control
Can the person decline, correct, export and delete their data without losing essential care?
Afterlife
How long is it retained, what trains on it and what happens if the company is sold?
Safety, pressure and cognitive rights
Enhancement choices do not occur in a vacuum. Rules, incentives and unequal bargaining power can turn an apparently personal decision into a condition of participation.
Autonomy
People need understandable information and a genuine ability to refuse. “Optional” use is not fully voluntary when grades, shifts, promotion or team acceptance depend on it.
Fairness
Access, risk and competitive pressure will not be distributed evenly. Institutions should improve workload and accessibility before normalizing chemical competition.
Privacy
Biological and behavioural data need strict purpose limits. A health signal should not quietly become a productivity score or an eligibility filter.
Treatment for a diagnosed condition is healthcare—not cheating. A fair policy protects access and accommodation while addressing nonmedical coercion, diversion and unsafe sourcing.
For schools and universities
- Do not stigmatize prescribed treatment or require disclosure beyond legitimate accommodation needs.
- Address impossible workload, sleep-hostile scheduling and competitive pressure.
- Provide confidential health support and evidence-based substance education.
- Separate academic integrity from disability and medical care.
For employers
- Never make enhancement use an explicit or implicit condition of productivity.
- Keep health data out of routine performance management.
- Design shifts, staffing and recovery time so medication is not used to patch unsafe work.
- Protect workers who seek treatment, accommodation or a human review.
Responsibility follows power. If an institution creates the pressure, selects the platform or benefits from the data, it cannot place all risk on the individual. Vendors, prescribers, schools, employers and regulators each have obligations that a consent screen cannot erase.
Check any “smart drug” claim in six moves
Use this framework on a product page, press release, podcast or research headline. If the claim cannot survive all six questions, reduce your confidence.
The P-P-S-O-D-O check
Make the claim carry its missing context.
A credible answer names what was tested and where the uncertainty remains. A marketing answer changes the subject.
Green flags
- A specific indication and relevant population
- A registered protocol and prespecified outcome
- Absolute benefits as well as harms and withdrawals
- Results that distinguish task scores from daily function
- Independent replication or converging evidence
- Clear regulatory status and conflict-of-interest disclosure
- Privacy limits, human oversight and a stopping plan
Red flags
- Guaranteed IQ, memory or productivity gains
- “Clinically proven” without a traceable study
- A mouse, biomarker or survey result presented as human efficacy
- Only relative percentages, testimonials or before-and-after screenshots
- Secret blends, vague “proprietary AI” or unsupported gene scores
- No discussion of interactions, withdrawals or negative findings
- Urgency, subscription lock-in or autonomous dosing claims
A headline translated
| The headline says | Ask for | A responsible translation may be |
|---|---|---|
| “Boosts memory by 30%” | Which test, baseline, comparator, absolute change and follow-up? | “A small study found a difference on one recall task immediately after exposure.” |
| “Personalized to your DNA” | Which variants, which drug, clinical validity and better patient outcomes? | “The report includes variants that may affect metabolism; its ability to choose a superior treatment is uncertain.” |
| “AI optimizes your dose” | Regulatory status, input quality, validation, clinician role and rollback? | “Software organizes self-reported and wearable data; autonomous prescription changes are not established.” |
| “Natural nootropic stack” | Exact ingredients, identity testing, interactions, trial evidence and adverse events? | “A supplement blend is being marketed without medicine-style premarket effectiveness review.” |
A responsible research and clinical roadmap
Replace speculative approval dates with standards that can be checked now. Progress is a better evidence system, not merely a faster pipeline.
Guardrails before scale
Human review
A named professional remains responsible for consequential medication decisions, with enough information and authority to disagree.
Monitoring and remedy
Adverse events, model drift and unequal outcomes are monitored; people can report harm, correct data and obtain review.
Non-digital routes
Essential care remains accessible to people who cannot or do not wish to use wearables, apps or genetic testing.
Purpose limits
Clinical data are not reused for employment, insurance, advertising or model training without a lawful, meaningful basis.
Independent scrutiny
Methods, conflicts, failure rates and subgroup results are available to regulators and qualified external evaluators.
Post-market learning
Approval or launch begins real-world monitoring; it does not end the obligation to update evidence and warnings.
A practical reader guide
If concentration, memory or mental stamina feels different, start by defining the problem—not by shopping for the most sophisticated-sounding product.
- Describe the change. What is difficult, when did it begin, how often does it occur and what daily activities are affected? “Brain fog” can describe many different experiences.
- Check the foundations and causes. Sleep, stress, depression, anxiety, pain, infection, nutrition, substance use, hormonal changes and existing medicines can affect cognition. New, sudden or worsening symptoms deserve appropriate clinical attention.
- Verify the product’s status. Look for the exact active ingredient, approved indication, regulator record, trial registry and evidence in a population like you—not a borrowed claim from another disorder.
- Review the whole list. Share prescriptions, over-the-counter drugs, supplements, caffeine and other substances with a qualified clinician or pharmacist so interactions and duplication are visible.
- Agree on an outcome and review. Know what improvement would matter, what harms to watch, when to reassess and what would lead to stopping. Do not change or combine prescription treatment based solely on an app or consumer test.
Questions for a clinician
- What conditions or existing medicines could explain this change?
- What is this treatment approved for in my country?
- What benefit should I realistically notice in daily life?
- What common and serious harms matter for my history?
- How will we monitor benefit, and when will we reconsider?
Questions for a test or app vendor
- Was this exact recommendation prospectively validated?
- Which regulator reviewed the claimed use?
- How often is the model wrong, and for whom?
- Who sees my genetic, mood, sleep or performance data?
- Can I delete it, opt out of training and still access care?
Seek urgent medical help for sudden confusion or a sudden neurological change. This article cannot assess symptoms, diagnose a condition or determine whether a medicine is appropriate for an individual.
Frequently asked questions
Short answers to the claims most likely to blur treatment, performance and precision medicine.
Is any pill proven to make healthy people broadly smarter?
No medicine has established a broad, durable increase in general intelligence for healthy people. Some substances can affect alertness or particular tasks under particular conditions. Those effects are not interchangeable with better judgment, deeper learning or long-term capability, and they carry trade-offs.
Does improved focus mean improved learning?
Not necessarily. Focus can increase time on task, but learning also depends on comprehension, retrieval, feedback, rest and transfer. A study should test what can be recalled or applied later—ideally without the drug present—not only how productive participants felt.
Can a DNA test choose my best nootropic?
No validated consumer DNA profile can identify an ideal intelligence-enhancing drug. Some pharmacogenetic results can inform the handling or risk of specific medicines, but their clinical role is gene–drug and context specific. They belong alongside diagnosis, other medicines, organ function, preferences and monitoring.
Are supplements safer because they are natural?
No. “Natural” does not establish identity, purity, dose, benefit or safety. Supplements can cause adverse effects and interact with medicines; in the United States, they do not undergo the same premarket approval process as medicines. Discuss all products you use with a qualified clinician or pharmacist.8
Is off-label prescribing the same as approval for enhancement?
No. Off-label prescribing may be a legitimate clinical decision, depending on law and context, but it means the regulator has not approved that product for that particular use. Ask what evidence supports the decision for your condition and what uncertainty remains.
Can a wearable reliably tell when a prescription dose should change?
A wearable may contribute a signal, but most consumer scores are noisy proxies influenced by device algorithms and context. Changing a prescription requires more than a readiness score. It should follow appropriate clinical guidance, validated measures and an agreed monitoring plan.
Did iclepertin prove that its target was useless?
No. A negative programme answers a narrower question about a particular compound, population, protocol and set of outcomes. It weakens the claim that this intervention produced the hoped-for benefit, but it does not make every future approach to the broader biology impossible. That is why negative trials are informative rather than embarrassing.
Is microdosing an evidence-based creativity enhancer?
Current blinded evidence does not establish it as one. Expectancy is difficult to control because participants may notice effects, subjective feelings can differ from objective scores, and studies are often small or short. Legality and medical risk also vary.
Evidence and further reading
Primary studies and official guidance behind the major claims. Drug status, labels and guidance can change; verify current information in your jurisdiction.
- Keefe et al.: Iclepertin CONNEX Phase III trials Three randomized trials of cognitive impairment associated with schizophrenia; no significant primary cognitive benefit versus placebo.
- Dvorak et al.: Discovery of TAK-041 Primary pharmacology paper identifying TAK-041 as a potent and selective GPR139 agonist.
- FDA prescribing information for Simtriyo (centanafadine) Official indication and safety label; see also the July 2026 approval announcement and Phase III summary.
- Roberts et al.: Meta-analysis of stimulant cognitive effects in healthy adults Randomized controlled evidence on modafinil, methylphenidate and d-amphetamine in healthy, non-sleep-deprived adults.
- Multimodal creativity assessments following LSD microdosing Randomized, placebo-controlled analysis separating subjective creativity from objective task outcomes; see also the parent six-week trial for cognition, mood and safety findings.
- FDA: Table of Pharmacogenetic Associations Official explanation of supported gene–drug associations, their limits and the continuing role of clinical judgment.
- FDA: Direct-to-Consumer Tests Evidence, oversight, interpretation and privacy considerations for consumer tests.
- NIH Office of Dietary Supplements: What You Need to Know Federal overview of supplement safety, interactions, claims and the U.S. regulatory distinction from medicines.
- FDA: Drug Development Tool Qualification Programs Why biomarkers and other tools are qualified within a specific context of use.
- FDA: Artificial Intelligence-Enabled Medical Devices Official device landscape and regulatory information; inclusion on a list is not endorsement of unrelated consumer claims.
- CDC Health Alert Network: Disrupted prescription stimulant access and counterfeit-pill risk Public-health advice on licensed prescribing, pharmacy sourcing, misuse and counterfeit pills.
- National Institute on Aging: Cognitive health and older adults Overview of health, medicines and lifestyle factors that can affect cognitive function.
- WHO: Ethics and governance of artificial intelligence for health International guidance on human autonomy, transparency, accountability, equity and sustainable AI in health.
Medical disclaimer: This article is for general educational information and does not provide diagnosis, treatment, prescribing or emergency advice. Do not start, stop, combine or change any prescription medicine or supplement because of this article, a wearable, an app or a consumer genetic test. Decisions should be made with a qualified healthcare professional who knows your health history and local rules. If symptoms are sudden, severe or urgent, seek local emergency care.
The useful version of precision
A better question—not simply more personal data
Cognitive pharmacology is most credible when it begins with a defined difficulty and ends with a meaningful, monitored outcome. It is least credible when a receptor story, a one-off task score or a stream of wearable data is presented as an intelligence dial.
The goal is not continuous chemical competition. It is safer, more effective care: treating genuine impairment, learning from negative trials, matching evidence to the person in front of us and protecting the right to think, work and learn without hidden pressure or surveillance.