Learning & Memory

Last reviewed 5 October 2026

What it is

Learning is the process of taking in something new and changing behavior because of it. Memory is what stays. Researchers divide the work into encoding (the intake), consolidation (the slow stabilizing that follows, often during rest or sleep) and retrieval (getting the material back out), and they divide memory by kind. Working memory holds a few items for seconds, episodic memory holds personal events, semantic memory holds facts, and procedural memory holds skills, such as a typing sequence or a musical passage.

Two more words need pinning down. Plasticity is a biological term: neurons and their connections can physically change. It is measured in cells and in animal brains. Learning, as this page uses the word, is a change in what a person can do, shown by a task. The first can happen without the second, and much of the excitement about psychedelics and learning moves from one to the other without a bridge.

Asking people how good their memory is turns out to be a weak guide to how they perform. A meta-analysis of 107 studies in healthy adults found a correlation of r = 0.15 between people's confidence in their memory and their actual memory performance. According to its abstract, the link was stronger for ratings made about the task at hand than for general beliefs about one's own memory, so the figure for general beliefs is lower than 0.15 [1]. In 53 studies of normally aging older adults the correlation was smaller still, r = 0.062, and the authors cautioned against using general memory beliefs as a substitute for testing [2]. Evidence for young healthy adults in particular is thin: no meta-analysis of this question in young healthy adults turned up in PubMed and Europe PMC, and PsycINFO could not be searched. One cross-sectional study of 125 people aged 18 to 28 found no link between a memory questionnaire and objective memory tests after correction for multiple comparisons. Satisfaction with one's memory tracked low mood and anxiety instead, and in a sample this size "no link" means no large link [31]. A Bayesian meta-analysis of a questionnaire on everyday cognitive failures, drawn from mixed groups that included students and several clinical samples, found correlations of about -0.06 to -0.19 with executive-function tasks and did not test memory itself [30]. Together these suggest a weak link, which for healthy young adults remains an inference and not a finding, so the studies below that count most are built on tasks.

Metanoia describes itself this way: Metanoia is a California 501(c)(3) public charity advancing the science of natural psychedelics for personal growth and human flourishing in healthy adults. Not treating illness. Expanding potential. Learning and memory are one of twelve dimensions of human capacity within that mission. The question for this page is narrow. In healthy adults, with controlled designs and real tasks, what has been shown about natural psychedelics and the ability to learn and remember? So far, very little has been shown, and the best-quantified result points toward impairment while the drug is active.

What is known

The short version. In rodents, psychedelics change brain structure and speed the extinction of a conditioned fear. In people, the best-quantified result is acute: while the drug is active, reaction time slows and some memory tasks get worse in proportion to dose. Two small controlled studies of LSD, a semi-synthetic compound, report better scores on a memory or motor-learning task a day later. Each is a single study: in the first, the gains have p values between 0.018 and 0.044 inside a wide battery of tests, and in the second the effect has several possible explanations besides learning. Repeated low-dose use has been tested against placebo several times and has not improved cognition. Nothing found for this page shows a durable gain in learning or memory in healthy adults from any natural psychedelic.

Plasticity in cells and animals

In a 2018 study, most of the psychedelic compounds tested increased the growth of neuronal branches, and some the tiny spines where neurons connect, in cultured rat cells [3]. In living animals the picture was narrower: only DMT was tested in adult rats, where it raised spine density in the prefrontal cortex a day later and increased the electrical signals at synapses, and LSD and DOI were tested in living fruit flies only. Ibogaine itself did nothing in the cultured cells, though its breakdown product did, and the messenger RNA for BDNF did not rise [3]. A 2023 study in adult male mice found that psilocybin, LSD, ketamine and ibogaine each reopened a window that normally closes after youth, during which the animals learn the reward value of social contact. The window stayed open for different lengths of time, roughly matching how long the drugs' effects last in people, a match drawn from published literature across five compounds with no statistical test reported. Ketamine's effect was gone by one week, psilocybin's by three weeks, and raising the LSD dose did not extend the window [4].

One kind of rodent learning has been tested many times, fear extinction. An animal learns that a tone predicts a mild shock, then hears the tone without the shock until it freezes less. Extinction adds a competing memory and leaves the old fear in place, so it counts as a learning model. In mice given psilocybin before extinction training, freezing fell faster during training in both studies and was lower when tested days later, but the dose that produced the lasting effect was the middle of three doses in one study [5] and only the highest dose in the other, which used male mice [6]. The lasting effect is therefore dose-sensitive, and the two laboratories did not agree on which dose worked. Giving the drug before the fear was learned, or right after extinction ended, changed nothing, so the drug seemed to need a live learning episode to act on [5]. Whether the serotonin receptor linked to hallucinations is required is disputed. In one group's work, blocking the 5-HT2A receptor erased the lasting benefit, while blocking 5-HT1A affected only the return of fear in a new setting [5]. Another group found the benefit survived blocking 5-HT2A, but its report was read in ahead-of-print form and did not measure the head-twitch response that would confirm the block [7]. The blocking drugs also changed freezing on their own [5] [7]. A further study in male mice found that extinction improved in animals stressed beforehand, though recall of the extinction memory after chronic stress did not reach significance (p = 0.094), and the effect appeared when the drug was in the body during training [8].

In a preprint of a calcium-imaging study in the retrosplenial cortex of mice, psilocybin quieted cells active during fear and recruited new cells active during extinction, but only in the subset of mice that turned out to respond well, a group defined after the fact by how much their fear faded. By the third extinction session and one month later the treated and control groups did not differ. The peer-reviewed version could not be accessed, so this rests on the preprint [9].

According to its abstract, a 2026 synthesis of 177 experiments in mice and rats concluded that psychedelics appear to enhance fear extinction across a broad range of doses, with lower doses generally linked to larger effects (MDMA the exception) and limited evidence for nonlinear dose-effect patterns. It noted that the interval between treatment and recall testing may matter, cautioned that the pooled patterns were sensitive to a few influential data points, and found too little LSD data to test its hypotheses for that compound [10].

Read plainly, this literature says that in rodents psychedelics can make a particular kind of emotional re-learning go faster. It does not say whether healthy adults would learn vocabulary, an instrument or a route any faster.

What happens in the acute window in people

A 2025 meta-analysis pooled 13 studies and 42 effect sizes on psilocybin's acute effects on attention and executive functions in healthy volunteers [11]. Reaction times slowed by a large amount (Hedges g = 1.13, 95 percent confidence interval 0.57 to 1.70). Accuracy did not change reliably (g = -0.45, interval -0.93 to 0.03, which includes zero). The slowing grew with dose. The studies disagreed with each other, there was clear evidence of publication bias (the effect stayed significant after an adjustment), and the authors rated the risk of bias as moderate to high, mainly because of blinding problems and a lack of preregistration [11]. It covers attention and executive tasks around the drug's peak and says nothing about learning over days.

For memory itself, a 2021 review pulled together experimental, observational and qualitative work on classic psychedelics. Its abstract says that impairment on memory tasks rises with dose, with none at low doses and more at higher ones. The pattern held across spatial and verbal working memory, semantic memory and episodic memory for non-personal material. The same review notes that these drugs make autobiographical memories feel more vivid, which is a different phenomenon from accurate performance on a memory test. Only the abstract of this review was accessible [12].

After the acute effects: two controlled studies in healthy volunteers

Two randomized, placebo-controlled crossover studies of LSD in healthy volunteers tested thinking and learning after the drug had worn off. LSD lies outside the natural compounds Metanoia studies. It appears here because these are the controlled studies of post-acute learning that this review found. Both used single sessions.

A Brazilian study of 24 healthy volunteers, mean age 35, tested a battery of tasks 24 hours after dosing. Compared with placebo, immediate recall of a complex drawing, a consolidation measure on an object-location memory task and phonological verbal fluency were all better, with p values between 0.018 and 0.044 for the recall and consolidation gains inside a wide set of tests, and the sources available do not make clear whether the analysis corrected for the number of comparisons. Delayed recall of the same drawing, a word-list learning test and the object-location task scored in points showed no difference. The Wisconsin Card Sorting Test, a measure of cognitive flexibility, was worse, with more perseveration. The authors called the work exploratory. Only the abstract of this paper was accessible, and these details come from it and from a systematic review that tabulated the study [13] [32].

A Swiss study published in June 2026 tested 45 healthy people, 43 of whom were analysed, in a double-blind crossover, 15 of them with earlier classic psychedelic use [14]. The main learning measure was a sequence-typing task done with the non-dominant hand the day after dosing. Learning during practice and the first test block, ten minutes later, did not differ between conditions. What differed was improvement across a rest period. In the second test block, 80 minutes after practice ended, the main text reports typing 348 milliseconds faster than at the end of practice after LSD and 262 milliseconds faster after placebo, a 32.8 percent greater gain, with no change in accuracy. The raw averages in the supplement are 348 and 254 milliseconds, so the size of the gap depends on how it is computed [14]. The authors describe this as the first demonstration of improved motor learning after a psychedelic and set out why it is hard to interpret: more fatigue and low mood in the LSD condition the next day, no measure of motivation, attention or arousal, no correlation with any brain measure, a single time point that cannot show whether a gain lasts, and a power calculation based on a brain-stimulation measure and not on the typing task [14].

One week later in the Swiss study, perceived stress was lower and one of the two subscales of a cognitive flexibility questionnaire was higher. Those are self-reports, and all participants correctly identified the real LSD dose. The authors describe baseline scores as near the ceiling for flexibility and near the floor for stress, a description this page cannot confirm from the supplement's averages [14]. The study also reported one serious adverse event, a hospitalization for a depressive episode two weeks after a dose, classed as possibly related. The participant had had symptoms before enrolling and was left out of the analysis, and five other participants still reported mild adverse effects such as anxiety at the last visit [14]. The two studies disagree about flexibility, a worse score on a task in one and a better score on a questionnaire in the other.

Two laboratory procedures meant to index human long-term potentiation produced no overall potentiation in the Swiss study, and blood BDNF, a growth-related protein used to track plasticity in animals, did not change at any time point in serum or plasma [14]. An earlier placebo-controlled study had found plasma BDNF higher over six hours after the lowest and the highest of three low LSD levels, but not after the middle one. It involved 24 recreational users of psychedelics, and only 5 of 23 had complete blood data [15]. The two studies do not give a consistent picture, and the Swiss authors point out that blood BDNF mostly does not originate in the brain [14]. Peripheral BDNF has not proved a reliable window onto brain events, and the animal plasticity findings have not been confirmed in humans with these tools.

A 2026 systematic review of placebo-controlled studies of LSD, DMT and psilocybin counted 32 studies, 31 in healthy people, 14 of which measured memory. Memory was weaker in some studies, mainly with psilocybin during the acute effects, better in one, and unchanged in others, depending on the task, the amount and the timing. Several consolidation tests showed no difference, including one study of repeated low LSD doses in 80 healthy men and one in adults aged 55 to 75. The authors list small samples, imperfect blinding and a focus on acute effects as limits [32].

A second 2026 systematic review looked at changes lasting a day or more in executive function, attention and decision-making after classic psychedelics, MDMA and ketamine, drawing on 67 articles. Psilocybin showed the most consistent improvement, particularly in cognitive flexibility and attention, and LSD and DMT showed no consistent subacute changes. Improvements appeared more often in psychiatric populations than in healthy ones, which may reflect repair of an existing deficit rather than a rise above normal functioning, and the authors flagged selective reporting. Only the abstract was accessible, and episodic learning is not among its domains [16].

A further 2026 study followed 28 healthy adults with no psychedelic experience through a first psilocybin session, single-blind, with a fixed order of sessions and no preregistration. A month later participants made slightly fewer errors on a cognitive flexibility task, with no change at other stages of the task. Episodic memory was not tested, and the EEG test of visual plasticity did not work as designed [33]. This is a flexibility result, so it does not change the central statement of this page about learning and memory.

Low-dose repeated use against placebo

Anecdotal reports claim that repeated low doses sharpen thinking [17] [18], but controlled data are less kind. A self-blinding study asked people who already planned to take small repeated amounts to prepare capsules at random, with a placebo group, and to report online over a ten-week study that included four weeks of dosing [17]. Of 1,630 people who signed up and 240 who started, 191 finished. Every self-reported psychological measure improved in the active group. The placebo group improved on some of them, mindfulness and paranoia, but not on well-being or life satisfaction, and no measure differed significantly between the groups. On six online cognitive tasks, the composite score did not change in the active group over the four weeks and did not differ from placebo. Participants correctly guessed their capsule 72 percent of the time against 63 percent expected by chance [17]. The authors concluded that the benefits many people report are consistent with expectation. Most participants took LSD (about 61 percent) and psilocybin mushrooms were a minority (about 24 percent, 57 of 240), so the result is mostly not about a natural compound [17].

A laboratory study looked closely at psilocybin-containing mushroom material in 34 people who were planning to start low-dose use [18]. Each person spent one week on the mushroom material and one week on a placebo made from an edible mushroom, double blind. About 75 percent of weeks were correctly unblinded. There was no significant benefit on creativity, attention and inhibition tasks or on well-being scales. Where uncorrected differences showed up, they pointed toward worse performance, and brain recordings showed a drop in theta power. The authors note that healthy volunteers may have been at ceiling, that the potency of mushroom material varies, and that two sessions a week cannot show cumulative effects [18].

A 2026 systematic review and meta-analysis of low-dose studies in healthy adults covered 24 studies and 3,681 participants, only six of them in pooled analyses, and the pooled outcomes were depression, anxiety and stress, not cognition. According to the abstract, a narrative summary showed subjective and brain-activity effects in the randomized trials but minimal impact on cognition, creativity or sustained mood, and observational studies pointed to improvements likely shaped by expectation or lifestyle. Only the abstract was accessible [19].

Long-term use

One study followed 127 regular ritual users of ayahuasca, a plant brew, and 115 controls drawn from other religious groups, testing both at baseline and a year later [20]. Users scored better on the Stroop test, the Wisconsin Card Sorting Test and a working memory task, and the authors found no evidence of cognitive impairment. It is not a trial. The groups were not matched for premorbid intelligence, participants were self-selected, many outcomes were tested, and fewer people returned for the second assessment [20]. The result speaks against lasting harm to the measured abilities in this group. It cannot show that ayahuasca improved anything.

What is not known

The central gap is simple. No controlled study found for this page shows that a natural psychedelic, given once or repeatedly, produces a lasting improvement in learning or memory in healthy adults. The searches ran in PubMed on 1 October 2026 and could have missed a recent paper.

Several questions sit under that gap. The first is whether a gain the day after dosing lasts, and whether it is learning at all. The one controlled finding on offline motor learning has not been replicated, was measured once, and has rival explanations in fatigue, arousal and motivation [14]. The day-after visuospatial gain in 24 people is exploratory and rests on p values between 0.018 and 0.044 [13]. The second is whether anything transfers from animals. Reopened critical periods and faster fear extinction are findings in mice and rats, the human tools meant to index plasticity did not work as designed in two recent studies, and blood BDNF has not tracked behavior [14] [33]. In mice the drug helped only when extinction training was happening at the same time [5], and whether human learning needs a matching learning episode has not been tested in healthy adults.

Expectation is a persistent problem. Blinding failed fully in the Swiss study and substantially in the low-dose studies [14] [17] [18], and any effect resting on self-report is hard to separate from belief. The type of memory also matters, since autobiographical vividness, task-based episodic memory, working memory and skill learning have been measured in different studies and should not be added together. Ceiling effects may limit what healthy volunteers can show [18], and a review of cognitive change found improvements more often in psychiatric groups, which suggests repair rather than enhancement [16]. The human post-acute learning studies found here used LSD, whole mushroom material appears only in the low-dose placebo study and in no study of learning over days, and windows of 24 hours to one week leave open what happens over months, in either direction.

A self-rating of memory tracks performance only weakly, even in young adults as far as the limited evidence goes [30] [31]. What Metanoia asks of this dimension follows from the evidence above: task-based outcomes in healthy adults, followed over windows longer than a day, with expectation recorded. This page reports no Metanoia study.

Compared with other paths

For learning and memory, other routes have far more evidence behind them. The comparison below mixes populations and outcomes, so the effect sizes should not be lined up as if they were measured on one scale. Several sources in this comparison could be read only as abstracts, as the source list marks, so their figures are as the abstracts report them.

Sleep

Sleep is the best-supported biological contributor to memory. A large narrative review, written by the authors of the leading theory, describes it as a state in which newly learned material is reactivated and moved into longer-term storage. Much of the evidence for the mechanism comes from animals, most human studies in the review are small, and the evidence in people for a role of REM sleep in declarative memory is described as inconsistent [21]. Meta-analyses of at least one night without sleep found that losing sleep after learning lowers later memory (Hedges g = 0.277) and that losing sleep before learning lowers the ability to take in new material (g = 0.621). After an adjustment for publication bias the figures fell to g = 0.166 and g = 0.463, average statistical power was 14 percent in the first analysis and 55 percent in the second, and the samples were mostly healthy young adults [22]. These figures describe what is lost without sleep, so they do not say how much extra sleep would improve memory in people who sleep normally.

Physical activity

A meta-analysis of cardiovascular exercise found that a single bout improved short-term memory (standardized mean difference 0.26) and long-term memory (0.52), while longer training programs had a small effect on short-term memory (0.15) and none on long-term memory (0.07, not significant) [23]. A meta-regression of 80 randomized trials in healthy people found a small overall benefit of long-term exercise on cognition [24]. In a trial of 120 previously inactive older adults aged 55 to 80, analysed among those who finished, a year of aerobic walking enlarged the anterior hippocampus by about 2 percent while a stretching and toning comparison group lost about 1.4 percent. Spatial memory scores improved in both groups with no difference between them, and blood BDNF did not differ [25]. That result is in older adults and does not carry over automatically to young ones.

Spaced and tested practice

A synthesis of 317 experiments on verbal recall, with 85 percent of the data from young adults, found that spreading study across time beats cramming, and that the best gap depends on how long the material must be remembered. Most experiments tested retention within a day, and only about 4 percent tested retention beyond a month [26]. A meta-analysis found that testing oneself beats restudying, with recall tests giving larger benefits than recognition tests [27]. A monograph reviewing ten study techniques rated practice testing and distributed practice as high utility and rereading and highlighting as low [28].

Meditation and training

A second-order meta-analysis combined six meta-analyses of working memory training in adults and found a small overall improvement (standardized mean difference 0.335, 95 percent confidence interval 0.223 to 0.447). The samples included some people younger than 18 or older than 55, and some clinical groups. Only adaptive working memory training was clearly better than control (0.370), while mindfulness (0.248, interval -0.009 to 0.506), resistance training (0.350, interval including zero) and gamified training (0.210) were not, and no method was significantly better than another, though with only six meta-analyses the comparison had little power. The authors said real-life usefulness may be minimal and did not test whether gains carry over to other abilities [29]. The Mindfulness page on this site covers that dimension separately. Psychotherapy is not a standard way to improve memory in people with no clinical memory complaint, so no comparison with therapy is offered.

Where psychedelics stand

Set against these, the psychedelic route is the least tested and the only one where the best-quantified result is a temporary decline. Psychedelics are supported mainly by animal findings and by a handful of small human studies whose results are open to other explanations.

History and tradition

This page makes no claim about traditional use and learning. None of the sources read for it examined that question, and tradition would not substitute for controlled data here. The ayahuasca study above [20] concerns members of modern religious groups and says nothing about historical practice.

The book on all twelve dimensions   Back to the twelve dimensions