What is a nootropic?

Giurgea did not just coin a word; he proposed a strict definition. To count as a true nootropic, he argued, a substance should ^[2]:

1. enhance learning and memory; 2. help the brain hold on to learned behaviour under conditions that would normally disrupt it, for example low oxygen or electroconvulsive shock; 3. protect the brain against physical and chemical injury; 4. improve the efficiency of the cortex's higher-level control processes; and 5. do all of that with very few side effects, extremely low toxicity, and without the sedation or stimulation of ordinary psychoactive drugs.

Criterion five is the one that does the real work, and it is the one modern usage quietly drops. A stimulant borrows tomorrow's alertness and charges interest on the loan; Giurgea's idea was of something that changes the machinery of cognition itself while being close to inert in every other respect. Piracetam genuinely fits that description. A review covering 407 papers found no affinity at the adrenergic, muscarinic, serotonin, dopamine, adenosine, opioid, GABA, benzodiazepine or glutamate receptors that psychoactive drugs normally act on, and a strikingly low toxicity with no serious side effects ^[4].

Criteria two and three are also more specific than they look, and in the same direction. Both are about damaged or stressed brains, not healthy ones. Giurgea's model systems were hypoxia, electroconvulsive shock and chemically induced amnesia. That bias runs through the entire field to this day: effects are largest where function is already impaired and smallest in rested young adults, which is exactly the population that buys most of the products.

One honest complication, because it explains why the definition failed to hold the line. The criteria were never turned into a test anyone could run: no threshold for how much learning has to improve, no agreed panel of injury models, no toxicity ceiling, and no requirement for a human trial at all. The definition works well as a statement of ambition and badly as a filter.

Modern usage sorts these compounds into loose families, and the families are worth learning because they genuinely do different things. Nothing in the table below shares a mechanism with anything else in it; the only thing they have in common is the shelf they are sold on.

Two things fall out of that table. The first is that the label is doing no work at all: the useful question is never whether something is a nootropic, but what this specific compound does, how strong the human evidence is, and what it costs in risk. The second is more uncomfortable. The best evidenced entries are the ones nobody markets with the word, because they already have an approved indication or sit in a coffee cup.

For how members of different families are combined in practice, and which combinations have a rationale rather than a story, see nootropic stacking. For the transmitter systems underneath all of them, see neurotransmitters 101.

Strip the marketing away and almost the whole category pulls on one of about six levers. Knowing which one a compound uses predicts its side effects, its tolerance behaviour and its interactions far better than the word nootropic ever will.

Cholinergic supply. Acetylcholine sets attention and memory encoding, and rather than carrying the message it changes how the network responds to everything else ^[21]. Precursors work upstream of the receptor: oral citicoline raises plasma choline in a dose-dependent way for five to ten hours, and raises uridine at the same time ^[19]. See the cholinergic system and choline sources compared.

Catecholamine tone. Dopamine and norepinephrine set effort, salience and alertness. Stimulants and eugeroics work here, and so does the reason their benefits are so slippery to measure: raising arousal improves a dull task and degrades one that is already at ceiling.

Glutamate and AMPA signalling. Fast excitatory transmission carries learning itself. Aniracetam potentiates AMPA receptor responses in slice work, which is where the ampakines began as a drug class. See NMDA, glutamate and memory.

Membranes, blood flow and metabolism. Piracetam's best supported action is physical rather than receptor based: it interacts with membrane phospholipids and restores fluidity in aged or damaged membranes, with knock-on effects across several transmitter systems ^[20]. Mitochondrial supports sit in the same territory; see mitochondria and cellular energy.

Stress axis damping and neurotrophic signalling. Most adaptogen claims, unpacked, are claims about cortisol rather than about cognition directly; see the HPA axis and cortisol. BDNF and synaptic plasticity are the most fashionable proposed mechanism in the field and among the least well demonstrated in living humans; see BDNF and neuroplasticity.

Notice what is absent from that list. There is no lever marked intelligence. Every mechanism on it changes a state: arousal, effort, encoding, blood flow, stress tolerance. States are the honest unit of measure here, and a compound that reliably shifts one is doing something real; it is simply not doing the thing the marketing implies.

Evidence quality across this space varies by orders of magnitude, and the variation is not random. It tracks almost perfectly with whether a compound ever had a pharmaceutical sponsor with a regulatory filing to defend.

Three problems make cognitive enhancement unusually hard to measure honestly. Ceiling effects: a rested young adult already performs near the top of most laboratory tasks, so there is little room to improve, and studies that find room have often chosen a task the participants happened to be bad at. Test choice: many standard instruments were designed to detect impairment in patients rather than gains in the healthy, and they are insensitive at the top of their range, which is why the same drug looks inert on a simple battery and effective on a harder one ^[8]. Publication bias: in the stimulant meta-analysis, the working-memory and long-term-memory effects were both qualified by evidence of it, which is why the authors' own conclusion is that the effect on healthy cognition is probably modest overall ^[9].

A fourth problem is specific to this field and worth understanding properly, because it is the source of most of the disagreement. Global impression is not cognition. Pooled across nineteen double-blind placebo-controlled studies, clinical global impression of change favoured piracetam with a clear odds ratio ^[6]. The Cochrane review of the same territory pooled the same kind of data, reached the same conclusion about global impression, and then reported that the evidence of effects on cognition and other specific measures was inconclusive; its bottom line was that the published literature did not support clinical use ^[5]. Both statements are true at once. Somebody feeling better, and their clinician agreeing, is a real finding. It is not evidence of a cognitive effect, and treating the two as interchangeable is how a drug ends up with a reputation its trial record cannot carry.

The pattern in the table below is worth stating plainly: the effects that survive good methodology are small, state-dependent, and largest in people who are impaired, sleep deprived, stressed, older or deficient. That is not a reason to dismiss the category. It is a reason to expect a nudge rather than a step change, and to treat any source promising otherwise as one that has not read the primary trials.

This is the part with immediate practical consequences. No regulator anywhere recognises nootropic as a class. A product carrying the word can contain a licensed medicine in one country, an unapproved drug in another, and something that is a lawful supplement ingredient in neither.

Piracetam is the sharpest illustration. It is a prescription medicine across much of Europe, where its strongest licensed use is cortical myoclonus rather than anything cognitive ^[20], while in the United States it is neither an approved drug nor a permitted dietary supplement ingredient. That has not kept it out of products sold as supplements. An analysis of five brands marketed as cognitive enhancers detected piracetam in four of them, with declared quantities that did not match the contents of the bottle ^[16].

A follow-up study went further and is worth reading in full. Ten products labelled as containing omberacetam (noopept), aniracetam, phenylpiracetam or oxiracetam were purchased and analysed by mass spectrometry. Between them they contained five unapproved drugs, including phenibut, vinpocetine and picamilon, several of which were not declared on the label at all. Where a quantity was printed, it was wrong 75% of the time, and a consumer following the serving instructions could take as much as four times a normal pharmaceutical dose, of up to four unapproved drugs at once, without knowing any of them were present ^[17].

The honest way to read a label in this category follows from that. The ingredient list is a claim, the stated dose is a claim, and the absence of an ingredient from the label is not evidence of its absence from the capsule. Independent assay is the only thing that settles either question. See research chemicals and harm reduction for how to handle the rest of that problem.

Usage is also lower than the coverage implies. A German survey of 1,035 pupils and 512 university students found lifetime use of prescription stimulants specifically for cognitive enhancement at 1.55% of pupils and 0.78% of students, with illicit stimulant use for the same purpose slightly higher again ^[18]. Estimates elsewhere vary widely, and the fair summary is that outside student populations nobody has a reliable number at all ^[10].

Four questions get you most of the way with any new compound, and they are the questions this wiki tries to answer on every entry. What is the proposed mechanism, and has it ever been shown in a human rather than a slice or a rodent? What is the best controlled human study, and what did it actually measure? In whom does the effect appear: healthy, impaired, sleep deprived, deficient? What is the exposure: dose, half-life, and what happens after months rather than one afternoon? See half-life and bioavailability for why the fourth question changes the answer to the other three.

That fourth question is where the field is thinnest, and the gaps are worth naming rather than leaving silence to imply they are filled.

Chronic use in healthy people is essentially unstudied. Nearly every positive result above came from a single dose or a few weeks of dosing. What daily use of a wakefulness agent, a cholinergic or a racetam does to a healthy brain over years has no dataset behind it in either direction, reassuring or otherwise. Tolerance falls in the same gap: a compound that works on day one and not on day sixty has an effect worth naming, and few trials run long enough to see it. See tolerance, dependence and down-regulation.

Mechanism remains unsettled even for the oldest members. Three decades after piracetam reached the market, a review of 407 papers concluded that no generally accepted mechanism of action had emerged for the class ^[4]. Further work has refined the candidates without settling the question.

Some effects may run the wrong way. Modafinil studies using basic paradigms have reported impairment on divergent creative thinking alongside improvement on convergent executive tasks ^[8]. Narrowing attention has a cost as well as a benefit, and almost no standard battery is built to measure the cost.

Individual variation is large and mostly unexplained. Baseline performance, sleep debt, diet, habitual caffeine intake and metabolic genotype all shift the result, which is why group averages this small translate badly into a prediction about one person.

None of this is medical advice. The category contains a small number of real findings, a large number of plausible mechanisms that have never been tested in a person, and a marketing layer that treats the two as equivalent ^[3]. Asking what a study measured, rather than what its abstract concluded, is most of what it takes to tell them apart.