Nootropics for memory and learning: what the evidence supports

A supplement can show an effect on one kind of memory test and none on another, so we grade five memory outcomes separately.

Working memory is holding and using information for a few seconds, measured with span tasks such as Digit Span [8] or backward digit span [10].

Episodic memory is memory for specific events and pairings, measured with tests such as the Paired Associate test [8] or word-list and story recall [34].

Delayed recall is how much of a studied list a person can still recall after a delay [7].

Memory as an outcome pools several memory measures into one domain, as in the creatine and ginseng meta-analyses [9][12].

Global cognition is measured with screening scales such as the Mini-Mental State Examination (MMSE) [17] or clinical scales such as ADAS-cog [25].

Many memory trials enroll older adults or people with a diagnosis, and their findings may not apply to healthy younger adults. We name the population in every summary.

Sleep. In two meta-analyses of studies from 1970 to 2020 on at least one full night without sleep, sleep loss after learning impaired later memory (Hedges' g = 0.277), and sleep loss before learning had a larger effect (g = 0.621) [1]. The authors found signs of publication bias and very low statistical power in most studies [1].

Exercise. An umbrella review of 133 systematic reviews (2,724 randomized controlled trials, 258,279 participants of all ages, including clinical groups) found that exercise improved memory (standardized mean difference, SMD = 0.26), more in children and adolescents than in adults [2].

Testing yourself. Recalling studied material can itself be a strong learning event, known as the testing effect, and in a meta-analysis recall tests gave larger benefits than recognition tests [3]. Across 222 classroom studies with 48,478 students, quizzing raised achievement to a medium extent (g = 0.499) [4].

Spacing study sessions. A synthesis of 317 verbal recall experiments found that the gap between study sessions giving the best retention grew as the time until the final test grew [5].

Each summary names the outcome its grade covers and covers effects only; side effects, interactions and groups who need extra care are in each compound's monograph.

Bacopa

A systematic review of six 12-week randomized controlled trials in adults without dementia found that bacopa extracts, at 300-450 mg a day in the studies, improved 9 of 17 free-recall tests, with little evidence in other cognitive domains [6]. A 2026 network meta-analysis of 29 RCTs in healthy adults (2,107 participants) found that low-dose bacopa outperformed placebo for delayed memory, though the authors noted that the lack of direct comparisons may limit the evidence [7]. See the bacopa monograph.

Citicoline

In a 12-week randomized, double-blind trial of 100 healthy adults aged 50-85 with age-associated memory impairment, citicoline improved episodic memory (Paired Associate test) and a composite memory score more than placebo [8]. Both were secondary outcomes; the full report lists only these two tests as differing from placebo, not the primary outcome, Spatial Span [8]. Kyowa Hakko Bio, which makes the product tested, funded the trial and helped design it, and two authors were its employees [8]. See the citicoline monograph.

Creatine

A meta-analysis of 8 randomized trials in healthy people found a small memory benefit compared with placebo (SMD = 0.29, with high variation between trials), concentrated in adults aged 66-76 (SMD = 0.88), with no significant effect in people aged 11-31 (SMD = 0.03) [9]. In a preregistered crossover trial of 123 healthy adults taking 5 g a day for 6 weeks, the effect on backward digit span only bordered on significance (p = 0.064) [10]. See the creatine monograph.

Ashwagandha

A 2026 meta-analysis of 20 RCTs (1,249 participants) reported improved memory with ashwagandha compared with placebo (SMD = 0.52) [11]. The pooled trials mixed healthy, physically active and special populations [11]. See the ashwagandha monograph.

Panax ginseng

A 2024 meta-analysis of 15 RCTs (671 participants) found a small effect on memory (SMD = 0.19) but no effect on overall cognition, attention or executive function, and it pooled healthy people with several clinical groups [12]. An earlier Cochrane review (searched in 2009) of 9 trials, 8 of them in healthy participants, found a lack of convincing evidence of cognitive enhancement [13]. See the Panax ginseng monograph.

Omega-3

A meta-analysis of 25 RCTs in adults without dementia, including people with mild cognitive impairment, found no effect of long-chain omega-3 fats on global cognition (Hedges' g = 0.02) and a mild memory benefit (g = 0.31), which the authors described in older adults [14]. A meta-analysis of 10 placebo-controlled trials in healthy people, people with cognitive impairment and people with Alzheimer's disease found no effect on composite memory (g = 0.04) and no benefit for working memory [15]. See the omega-3 monograph.

Huperzine A

In a double-blind, matched-pair trial in 34 pairs of junior middle school students who complained of poor memory, 4 weeks of huperzine A gave a higher memory quotient than placebo (115 versus 104) [16]. In people with Alzheimer's disease, a clinical population, a meta-analysis of 20 RCTs found higher MMSE scores at 8 to 16 weeks, but most trials had a high risk of bias [17]. See the huperzine A monograph.

Magnesium

In a 6-week randomized, double-blind trial of 100 adults aged 18 to 45 with self-reported dissatisfied sleep, magnesium L-threonate (Magtein, 2 g a day) improved the NIH Total Cognition Composite more than placebo (p = 0.043), with larger effects on working and episodic memory [18]. The trial was funded by Threotech, which supplied the product and helped design the study [18]. A 2024 meta-analysis of 3 RCTs and 12 cohort studies concluded that the trial evidence was insufficient to draw conclusions on magnesium supplements [19]. See the magnesium monograph.

Holy basil

In a 30-day placebo-controlled trial of 40 healthy adults, 300 mg a day of holy basil leaf extract gave faster reaction times and fewer errors on the Sternberg working-memory test than placebo [38]. Both groups improved, but placebo gains leveled off after day 15 [38]. See the holy basil monograph.

PQQ

In a 12-week randomized, double-blind trial of 64 healthy Japanese adults aged 40 to under 80 who felt more forgetful, PQQ disodium salt (21.5 mg a day) improved composite and verbal memory scores more than placebo [39]. Ryusendo, which makes the product tested, funded the trial, and its researchers ran it [39]. See the PQQ monograph.

L-tyrosine

A systematic review of 15 studies in healthy adults, all but one using a single test session, reported that in the cognitive studies a single dose before testing counteracted drops in working memory under demanding conditions such as extreme weather or cognitive load [20]. The authors called for larger trials lasting multiple weeks [20]. See the L-tyrosine monograph.

Phosphatidylserine

In a 12-week placebo-controlled trial of 149 adults with age-associated memory impairment, phosphatidylserine from bovine cortex (100 mg three times a day) improved learning and memory tests more than placebo, and a subgroup analysis suggested that people who scored lower before treatment were most likely to respond [21]. The first author was affiliated with Memory Assessment Clinics, Inc. [21]. A 12-week trial of soy-derived phosphatidylserine (300 or 600 mg a day) in 120 adults over 57 with age-associated memory impairment found no differences on any outcome [22]. See the phosphatidylserine monograph.

Lion's mane

In a 16-week randomized, double-blind trial of 30 Japanese adults aged 50 to 80 with diagnosed mild cognitive impairment, a clinical population, lion's mane raised scores on a cognitive function scale more than placebo [23]. The first author worked at the Mushroom Laboratory of Hokuto Corporation [23]. In a single-dose crossover trial of 18 healthy adults aged 18 to 35, there was no significant effect on composite global cognition [24]. Two authors were employed by Beyond Alcohol Ltd., which uses lion's mane in its products and contributed towards the publication fee [24]. See the lion's mane monograph.

Saffron

A meta-analysis of 4 RCTs in people with Alzheimer's disease or mild cognitive impairment, a clinical population, found better scores than placebo on ADAS-cog and CDR-SB [25]. The authors judged the evidence insufficient for any recommendation [25]. See the saffron monograph.

B vitamins

A Cochrane review of 14 trials in cognitively healthy people (27,882 participants, mostly over 60 with a history of heart or blood vessel disease) found that folic acid, vitamin B12 or B6, alone or combined, probably have little or no effect on global cognition up to 5 years [26]. In the 3-year FACIT trial of 818 adults aged 50 to 70 with raised homocysteine, 800 µg of folic acid daily gave a better change in memory than placebo, a secondary endpoint [27]. Pooled data from 4 homocysteine-lowering trials (1,340 people) showed no significant effect on memory or other cognitive domains [37]. See the B vitamins monograph.

Acetyl-L-carnitine

All 11 trials in a Cochrane review were in people with Alzheimer's disease, a clinical population, and the review found no evidence of benefit on cognition [28]. It did find a benefit on clinical global impression, which the authors said may be due to chance given the number of comparisons [28]. See the acetyl-L-carnitine monograph.

Glycine

A double-blind, randomized crossover trial of a branded glycine product, Bioglycin, in healthy students and middle-aged men found better retrieval from episodic memory but no effect on focused or divided attention [29]. A crossover trial of a single 0.8 g/kg dose in healthy young adults found no significant effects on working memory, declarative memory or attention [30]. See the glycine monograph.

Uridine

The controlled memory trials summarized here tested Souvenaid, which combines uridine monophosphate with omega-3 fats, choline, vitamins and other nutrients [32]. In the 36-month LipiDiDiet trial of 311 people with prodromal Alzheimer's disease, a clinical population, partly funded by Danone Nutricia Research, memory declined less than with placebo, but only 81 participants had 36-month data eligible for efficacy analysis [32]. A Cochrane review of three trials in 1,097 people with Alzheimer's disease found it probably results in little or no difference in global or specific cognitive functions; two of the trials were funded by the manufacturer [31]. See the uridine monograph.

Cocoa flavanols

In COSMOS-Web, 3,562 older adults were randomly assigned to cocoa extract (500 mg of cocoa flavanols a day) or placebo, and the prespecified primary endpoint, a memory improvement in all participants after 1 year, was not statistically significant [33]. COSMOS was supported by a grant from Mars Edge, a segment of Mars, which donated the study pills, and two COSMOS-Web authors were Mars employees [33]. Among participants in the lower thirds of habitual diet quality or flavanol intake, the authors reported that the intervention restored memory [33]. In COSMOS-Mind (2,262 participants, mean age 73), cocoa extract had no effect on global cognition, and the paper states that Mars did not contribute to the trial's design or conduct [34]. See the cocoa flavanols monograph.

5-HTP

In a single-blinded 12-week trial of 30 cognitively healthy older adults, with an untreated control group and no placebo, MoCA screening scores rose in the 5-HTP group, but the change did not differ from the control group [40]. Cognition was a secondary outcome [40]. See the 5-HTP monograph.

Vitamin D

A preregistered meta-analysis of 24 trials (7,557 participants, mean age 65.21) found a small effect on global cognition (Hedges' g = 0.128) but none on specific cognitive domains [41]. In DO-HEALTH, a trial of 2,157 adults aged 70 or older co-funded by companies including DSM Nutritional Products, vitamin D had no effect on MoCA scores [42]. See the vitamin D monograph.

Gotu kola

Pooling the two trials of gotu kola alone (54 participants) found no significant effect on working memory (SMD 0.19) [43]. In a 12-week trial of 60 older adults with mild cognitive impairment, a clinical population, adding gotu kola to exercise gave no extra cognitive benefit [44]. See the gotu kola monograph.

Lemon balm

In a 96-week randomized, double-blind trial of 323 older adults with subjective or mild cognitive impairment, lemon balm extract made no significant difference to cognitive measures compared with placebo [45]. One subgroup difference was reported, in participants without hypertension [45]. See the lemon balm monograph.

Ginkgo

A meta-analysis of 13 randomized trials in healthy individuals found an effect on memory close to zero (d = -0.04) [35]. In the GEM trial, 3,069 adults aged 72 to 96, with normal cognition or mild cognitive impairment, took 120 mg of ginkgo extract twice daily or placebo, and over a median of 6.1 years annual rates of change did not differ in any cognitive domain, including memory [36]. See the ginkgo monograph.

Artichoke extract

In aged mice, luteolin in the diet for 4 weeks improved spatial working memory, with no effect in young adult mice [46]. As of November 11, 2026, we found no randomized trials of artichoke extract or luteolin on memory in healthy adults. See the artichoke extract monograph.

Celastrus paniculatus

A 2026 systematic review states that randomized controlled trials in humans are absent [47]. A 2023 review also noted the absence of any reported clinical studies [48]. See the Celastrus paniculatus monograph.

Supplements for studying applies this evidence to exam preparation, and How to read a supplement study explains terms such as effect size. Every grade here also appears in the evidence matrix; how we grade evidence sets out the rubric.

The newest syntheses here are 2026 reviews of bacopa and ginkgo [7], ashwagandha [11] and Celastrus paniculatus [47]. When new research changes a memory grade, the change is listed on what's new.

What is the best nootropic for memory?

Grades apply to one compound and one outcome, so no single supplement is best for every kind of memory. Sleep, exercise and self-testing are each supported by meta-analyses [1][2][4]. Among supplements, a systematic review of six 12-week trials found bacopa improved 9 of 17 free-recall tests in adults without dementia [6].

Do memory supplement results in older adults apply to younger people?

Not necessarily. Creatine's pooled memory effect was concentrated in adults aged 66-76 and absent in people aged 11-31 [9], and the citicoline and phosphatidylserine trials enrolled older adults with memory complaints [8][21].