The efficacy of nutritional phytochemical compounds in improving cognition

Abstract

The incidence of cognitive disorders is rising, and medications to address cognitive impairments are limited. Phytochemical compounds have had some success in improving cognitive outcomes in preclinical studies. However, the evidence surrounding the implementation of these compounds routinely in (1) improving cognition in healthy individuals or (2) as prophylaxis and treatment against the cognitive sequelae of neurological insult has not been collated. To address this, the evidence base surrounding the impact of phytochemical agents on cognitive function in healthy individuals was reviewed. Web of Science, PubMed, PsycNET, and AYUSH Research Portal were systematically searched for double-blind and single-blind randomized controlled trials assessing the efficacy of Curcuma longa, Bacopa monnieri, Ocimum tenuiflorum, Camellia sinensis, Centella asiatica, Ganoderma lingzhi, and Rosmarinus officinalis on cognitive function. Seven studies met inclusion. These examined the impact of B. monnieri, O. tenuiflorum, or C. sinensis. There was no benefit of B. monnieri or C. sinensis on processing speed, attention, working memory, language, psychomotor function, or overall cognitive performance. O. tenuiflorum improved reaction times on an executive function task. Further, more methodologically robust trials of longer duration are recommended to investigate the efficacy of phytochemical supplementation for cognition.

Significance Statement.

This study addresses the effectiveness of various naturally occurring molecules in improving cognitive functions in healthy adults. It evaluated whether compounds, including Bacopa monnieri, Ocimum tenuiflorum, and Camellia sinensis, can improve memory, processing speed, and executive functions. While no significant cognitive improvements in healthy individuals were observed for most of the molecules we reviewed, O. tenuiflorum did demonstrate marginal benefits in reaction times during an executive function task. Overall, these phytochemicals demonstrated limited efficacy across cognitive domains, including memory and attention. This study consolidates the evidence based on phytochemicals for cognitive enhancement, highlighting a need for more robust, methodologically sound trials to determine if these natural compounds hold promise in cognitive therapeutics, particularly for populations with cognitive impairments.

INTRODUCTION

There is a rising incidence of cognitive disorders throughout the lifespan, ranging from neurodevelopmental conditions^1,2 and acquired neurological injury³ to neurodegenerative diseases later in life.⁴ These conditions place significant health and socioeconomic demands on society, with costs of care expected to exceed approximately $13bn by 2031.^5,6 As such, there is a growing demand and bioethical imperative for research into therapeutic interventions and prophylactic measures to prevent these conditions.⁷ It is evident that safe and effective cognitive enhancers have the potential to benefit both healthy and cognitively compromised individuals, as well as society as a whole.⁸ However, while acetylcholinesterase inhibitors, for example, demonstrate effects on cognitive performance in diagnoses of cognitive impairment,⁹ concerns remain surrounding their risk and side-effect profile.^9,10

Phytochemical compounds, often marketed as dietary supplements,¹¹ are widely available to adults and are frequently promoted for cognitive and brain health.¹¹ Many of these compounds are derived from herbs and spices that have been used in traditional medical systems for centuries. Traditional uses commonly emphasize cognitive support, with formulations reportedly intended to enhance memory, learning ability, and concentration. In addition to their historical use, several of these phytochemicals are now incorporated into commercially available products explicitly advertised for brain health, cognitive performance, or mental clarity, reflecting sustained public and scientific interest. Accordingly, this review prespecified phytochemicals that meet 3 criteria: long-standing traditional association with cognition, biological plausibility supported by preclinical evidence, and progression into human research or commercial cognitive-health formulations, including Curcuma longa L. (curcumin), Bacopa monnieri (L.) Wettst. (BaMo; Brahmi), Centella asiatica (L.) Urb. (gotu-kola), Ganoderma lingzhi (Curtis) P. Karst. (reishi), Salvia rosmarinus Spenn. (rosemary), Ocimum tenuiflorum L. (OcTe; holy basil), and Camellia sinensis (L.) Kuntze (CaSi; tea plant, as in green or black tea). Several of these compounds have demonstrated cognitive effects in preclinical models, either in isolation or in combination, and some have progressed to clinical trial evaluation.^12–19 For example, B. monnieri and C. sinensis (in the form of green tea) may assist mild cognitive impairment and memory in older populations. In a meta-analysis of 9 studies on older people who were healthy or had mild memory impairment, 140–600 mg/day B. monnieri improved cognitive performance, particularly speed of attention, over 12–24 weeks of supplementation.²⁰ Furthermore, a systematic review from 2017 on the impact of green tea on cognitive function identified 4 randomized controlled studies demonstrating benefits in memory and attention in healthy participants and patients with mild cognitive impairment, and 5 observational studies showing benefits of green tea consumption in adults older than 65 years.²¹ There was no significant cognitive effect from oral ingestion of the green tea extract polyphenol epigallocatechin gallate (EGCG) in young healthy volunteers although cerebral blood flow parameters did change.²² These studies suggest that the combined effects of compounds in green tea may be more effective than the administration of separate compounds and that the demonstrable effects of these phytochemicals on cognition in young healthy adults are perhaps less likely.^21–23

Neurobiologically, the consumption or administration of various phytochemicals has been shown to mechanistically mimic or complement traditional approaches to the management of cognitive impairment; for example, through interaction with acetylcholine esterase enzymes^16,24–27 and modulation of oxidative stress and reactive oxygen species.^28,29 Based on an expanding body of preclinical evidence exploring the biological basis of how various phytochemicals may exert their action on cognition, as well as the increasing medical and societal needs for effective treatment, the number of clinical trials reporting outcomes relating to phytochemicals and their impact on cognition has grown.

While the evidence surrounding green tea and its relevant constituents is relatively well summarized, the impact that other herbs and phytochemical compounds may have on cognition remains to be systematically reviewed. Furthermore, we have yet to establish compelling evidence in support of their routine use in (1) improving cognitive function in healthy individuals or (2) as prophylaxis and treatment against the cognitive sequelae of neurological insult. Here we report results from a systematic review investigating the efficacy and adverse effects of some commonly researched phytochemical compounds in improving cognitive performance in adults of working age who are healthy or who have nonprogressive neurocognitive impairments.

METHODS

A comprehensive literature search to identify all relevant double-blind and single-blind randomized controlled trials (RCTs) assessing the efficacy of C. longa, BaMo, C. asiatica, Ganoderma lingzhi, Salvia rosmarinus, OcTe, and CaSi on cognitive function in healthy or nonprogressive cognitively compromised populations was performed.

The Preferred Reporting Items for Systematic Reviews and Meta-Analysis reporting guidelines were used to document outcomes. Details of the literature search and data extraction are reported in the supplementary methods.

Search Methods

Web of Science (1864-2019), PubMed (1809-2019) and PsycNET (1967-2019), and AYUSH Research Portal were searched on January 5, 2024. Text-based terms were compiled based on the botanical names of the selected phytochemicals (using the Medicinal Plant Naming Service, MPNS³⁰) and accepted alternative terms; the MPNS is a working list of all known plant species produced as a collaboration of the Royal Botanic Gardens, Kew and the Missouri Botanical Garden. In addition, experts in the field were consulted to contribute additional text-based terms to ensure comprehensive search.

Inclusion Criteria

Inclusion criteria were oral administration of C. longa, C. asiatica, Ganoderma lingzhi, Salvia rosmarinus, BaMo, OcTe, and/or CaSi for a minimum of 3 weeks; double-blind and single-blind and quasi-RCTs; healthy participants with no significant medical history and documentation of educational difficulties; and clinical participants with a formal diagnosis of a nonprogressive/nondementing neurological condition made by an appropriately qualified medical professional putting them at risk of cognitive compromise as defined by accepted criteria, such as the ICD-10.

Exclusion Criteria

Exclusion criteria included the following: any study that did not adopt a blinded RCT or quasi RCT methodology; inclusion of participants of non-working age (e.g., younger than 18 or older than 65 years); studies using non-oral administration (e.g., intravenous); and studies exploring a dosing schedule that is less than 3 weeks.

Example Search String

The example search string was (((((CaSi OR green tea OR C. sinensis OR thea sinensis) OR (Ocimum tenuiflorum OR Ocimum sanctum OR holy basil OR holy basils OR holy basil leaf OR anisodorum OR villicaulis) OR (angustifolius OR chilensis OR communis OR eriocalyx OR flexuosus OR latifolius OR laxiflorus OR noeanus OR officinalis OR palaui OR pallescens OR prostratus OR rigidus OR rosemary OR Rosemary leaf extract OR Rosemary plant OR Rosmarinus angustifolius OR Rosmarinus angustissimus OR Rosmarinus clementei OR Rosmarinus eriocalyx OR Rosmarinus latifolius OR Rosmarinus laxiflorus OR Rosmarinus noeanus OR Rosmarinus nutans OR Rosmarinus officinali OR Rosmarinus officinalis OR Rosmarinus palaui OR Rosmarinus pallescens OR Rosmarinus prostratus OR Rosmarinus rigidus OR Rosmarinus subtomentosus OR Rosmarinus tomentosus OR Rosmarinus tournefortii OR Rosmarinus trogloditarum OR serotinus OR tenuifolius OR tournefortii) OR (Chondrocarpus asiaticus OR Chondrocarpus triflorus OR Glyceria asiatica OR Glyceria triflora OR monantha OR Hydrocotyle biflora OR Hydrocotyle brasiliensis OR Hydrocotyle brevipedata OR Hydrocotyle ficarifolia OR Hydrocotyle ficarioides OR Hydrocotyle inaequipes OR Hydrocotyle lurida OR Hydrocotyle nummularioides OR Hydrocotyle reniformis OR Hydrocotyle repanda OR Hydrocotyle sylvicola OR Hydrocotyle triflora OR Hydrocotyle tussilaginifolia OR Hydrocotyle uniflora OR Hydrocotyle asiatica OR Hydrocotyle OR Centella affinis OR Centella arbuscula OR C. asiatica OR Centella biflora OR Centella boninensis OR Centella bupleurifolia OR Centella capensis OR Centella chamissonis OR Centella cuneifolia OR Centella dregeana OR Centella dusenii OR Centella eriantha OR Centella filicaulis OR Centella floridana OR Centella glabrata OR Centella glochidiata OR Centella hederifolia OR Centella hermaniifolia OR Centella hermanniifolia OR Centella hirtella OR Centella homalocarpa OR Centella linifolia OR Centella macrocarpa OR Centella montana OR Centella renifolia OR Centella repanda OR Centella solandra OR Centella triflora OR Centella tussilaginifolia OR Centella ulugurensis OR Centella uniflora OR Centella verticillata OR Centella villosa OR Centella virgate OR asiatica OR crista OR capensis OR micrantha OR eriantha OR acuminata OR affinis OR cochlearia OR glabrata OR litoralis OR schlechteriana OR linifolia OR macrocarpa OR longifolia OR montana OR floridana OR villosa OR virgata OR Gotu Kola) OR (Brahmi Leaf OR 3-Glu-Rha-jujobogenin OR B. monnieri OR Anisocalyx limnanthiflorus OR Bacopa micromonnieri OR cuneifolia OR micromonnieri OR Bramia indica OR Calytriplex obovata OR Gratiola monnieri OR Gratiola portulacacea OR Gratiola tetrandra OR Habershamia cuneifolia OR Herpestis cuneifolia OR Herpestis micromonnieri OR Herpestis monnieri OR Herpestis procumbens OR Limosella calycina OR Lysimachia monnieri OR Monnieri cuneifolia OR Septas repens) OR (C. longa OR Tumeric OR Yellow Tumeric OR Curcumin OR Diferuloylmethane OR Amomum curcuma OR Curcuma brog OR Curcuma domestica OR vanaharidra OR Vana haridra OR Curcuma ochrorhiza OR Curcuma soloensis OR Curcuma tinctoria OR Stissera curcuma OR Stissera curcuma OR Haldi OR Hardi OR Pasupu OR Pasupu Kommulu OR Manjal OR Arishina OR Haldar OR Halaj OR Halud OR Haladar OR Halad OR Kumkum OR Zardchob OR Curcuma aromatic OR Curcuma kwangsiensis OR Curcuma phaeocaulis OR Curcuma wenyujin OR Tumeric rhizome OR turmeric root) OR (("Curcumin"[Mesh] OR "Curcuma"[Mesh] OR "bacopasaponin C"[Supplementary Concept] OR “C. asiatica extract”[Supplementary Concept] OR "Centella"[Mesh] OR "Reishi"[Mesh] OR "Rosmarinus"[Mesh] OR "Ocimum sanctum"[Mesh] OR "CaSi AR25"[Supplementary Concept] OR “CaSi polyphenone E”[Supplementary Concept] OR "Tea"[Mesh]))) AND (((randomized controlled trial[pt]) OR (controlled clinical trial[pt]) OR (randomized[tiab]) OR (placebo[tiab]) OR (drug therapy[sh]) OR (randomly[tiab]) OR (trial[tiab]) OR (groups[tiab])) NOT (animals[mh] NOT humans[mh])))) AND "Cognition"[Mesh]) AND (Cognition[Title/Abstract] OR Cognitive[Title/Abstract])

All the references identified by the searches defined above were downloaded into Covidence systematic review software (Veritas Health Innovation, 2023) and blind screened for inclusion by at least 2 reviewers (Alexander Marsh, AM; Darren Quelch, DQ; Marion Mackonochie, MM). Any discrepancies in screening were resolved by discussion and a fourth reviewer (Vivien Rolfe, V.R.) to make a final decision on any discrepancies. Full text review and data extraction were performed by A.M. and D.Q. The screening and selection process is reported in Figure 1.

Figure 1.

Preferred reporting items for systematic reviews and meta-analysis (PRISMA) flow diagram summarising identification of eligible studies.

Outcomes

Primary

Reliable change in cognitive performance on any objective, published, standardized psychometric test. When reliable change was not reported, means and SDs of cognitive endpoints of the intervention and control group were considered.

Secondary

1. Neurophysiological proxies of cognitive function, including functional magnetic resonance imaging (fMRI), electroencephalography, magnetoencephalography, and fludeoxyglucose positron emission tomography.

2. Performance of activities of daily living measured by validated rating scales (such as the Vineland³¹ or Adaptive Behaviour Assessment System^32-34).

3. Incidence and severity of adverse effects.

4. Death.

Data Analysis

Risk of bias was assessed according to Cochrane standards categorizing outcomes into high, low, or unclear risk of bias.³⁵ Separate meta-analyses were conducted for the different phytochemicals’ impact on the cognitive functions assessed. For each analysis, the effect size of the difference between the control and intervention group was computed. A random-effects or standard-effects meta-analytic model was employed based on heterogeneity among studies.^35,36 The degree of heterogeneity was assessed by visual inspection of forest plots and by examining the χ² test for heterogeneity. Heterogeneity was quantified using the I² statistic, with an I² value of 40% or more representing moderate levels of heterogeneity and leading to a random effects approach to analysis.³⁵ The summary effect sizes were computed using a maximum-likelihood estimator in RevMan 5.3.³⁷ The significance level was set at P < .05 (2-tailed).

RESULTS

Description of Studies

A total of 417 search results were retrieved following searches of PubMed, Web of Science, PsychNET, and the AYUSH Research Portal. Following duplicate removal (n = 53), title and abstract screening (n = 225), and full text eligibility exclusion (n = 32), 39 potentially relevant studies were identified (Fig. 1).

Seven studies met the inclusion criteria^16,38–43 (Table 1 for summary) and explored 3 of the herbs: C. sinensis, B. monnieri, and Ocimum tenuiflorum.

Table 1.

Summary of characteristics of included studies.

	Study	Age range	Length of Study Including Follow Up (days)	Population Description	Treatment	Control	Included Primary Outcomes	Included Secondary Outcomes
CaSi	De La Torre et al38	18-30	548	Patients with Down syndrome genetic variations (trisomy 21, partial trisomy, mosaic, or translocation)	n = 43 CaSi supplement containing 45% EGCG in 200 mg (dose weight dependent) + cognitive training for 12 months	n = 41 Rice flour in a 200 mg pill + cognitive training for 12 months	PS (Simple reaction test), Attention (Spatial Span Forwards and Digit Span Forwards), WM (Spatial Span Backwards, Digit Span Backwards), Language (Word Fluency, Language Comprehension - Token Test, Boston Naming Test), Psychomotor (Motor Screening Test), Memory (Paired Associates Learning, Visual Recognition Memory, Verbal Episodic Memory – Cued Recall Test), EF (Word Fluency, Inhibition Cats and Dogs, Tower of London, Wiegel Sort Test, Sternberg Test, Stroop Test)	Adverse Events Adaptive Functioning (ABAS Total Score) Neurophysiological proxy (resting-state fMRI, TMS)
OcTe	Sampat et al17	18-30	30	Healthy males	n = 20 70% ethanolic extract of OcTe, 300 mg once daily for 30 days	n = 20 Placebo (sucrose) was supplied by the Dabur Pharmaceutical (India) Ltd. and taken the same as directed for OcTe for 30 days	EF (Sternberg Test, Stroop Test)	Adverse Events
BaMo	Kumar et al39	19-22	42	Healthy medical students	n = 28 BaMo supplement containing 45% Bacopasaponins in 150 mg, twice daily with food for 45 days	n = 14 Unreported content of placebo. Same dose instructions as intervention	PS (Simple reaction test, Symbol Digit Modalities Test), Attention (Digit Span Forwards, Choice Reaction Test, Choice Discrimination Test), WM (Digit Span Backwards, Memory Span for Nonsense Syllables), Memory (Paired Associates Learning, Logical Memory Test), Psychomotor (Finger-tip tapping)	None
	Roodenrys et al40	40-65	c.134	Healthy volunteers	n = 37 BaMo supplement (KeenMind128), 300 mg for persons <90 kg, and 450 mg for persons >90 kg for 1 month	n = 39 Unreported content of placebo and dosing	Attention (Spatial Span Forwards and Digit Span Forwards),), WM (Spatial Span Backwards, Digit Span Backwards), PS (Coding Copy)	Adverse Events
	Sathyanarayanan et al41	35-60	84	Healthy volunteers	n = 33 Bacomind129 (BaMo extract), 225 mg twice daily for 12 weeks	n = 33 Placebo capsules contained equal quantity of starch along with the same excipient. in same size of hard gelatine capsule shell	Memory (Auditory Verbal Learning Test), EF (Stroop Test)	Adverse Events
	Stough et al42	18-60	84	Healthy volunteers	n = 23 BaMo supplement (KeenMind128), 150 mg twice daily taken for 84 days	n = 23 Placebo capsules identical in shape, color, smell, taste, and weight but content was not described	PS (Simple reaction test, Symbol Digit Modalities Test), Attention (Digit Span Forwards), WM (Digit Span Backwards), Language (Speed of Comprehension Test), Memory (Auditory Verbal Learning Test), EF (Trail Making B)	Adverse Events
	Stough et al43	18-60	90	Healthy volusnteers	n = 33 BaMo supplement (KeenMind128), 150 mg twice daily taken for 90 days	n = 29 Placebo capsules identical in shape, color, smell, taste, and weight but content was not described	PS (Simple reaction test)	Adverse Events

Abbreviations: BaMo, B. monnieri; OcTe, Ocimum tenuiflorum; CaSi, C. sinensis; PS, Processing Speed; WM, Working Memory; EF, Executive Function.

One study³⁹ focused recruitment on medical students aiming to evaluate the effect of cognitive-enhancing supplements in a presumably already high-functioning cognitive population. The OcTe study¹⁶ recruited only male participants though gave no reason for doing so. A single study reported outcomes among patients with Down syndrome using CaSi supplementation.³⁸ It is important to note that in this study CaSi was administered alongside cognitive training in both trial arms; observed effects cannot be attributed to CaSi in isolation and are interpreted cautiously.

All studies went through ethical review and obtained approval. All adopted a randomized, double blind, placebo-controlled, noncrossover, parallel trial design. Only 2 studies^38,41 conducted a power calculation to estimate sample size.

In total, 416 participants were included in the analysis of the 7 studies (BaMo: 5 studies; OcTe: 1 study; CaSi: 1 study). Two-hundred and seventeen were in treatment groups (CaSi n = 43; OcTe n = 20; BaMo n = 154) and 199 were in placebo groups (CaSi n = 41; OcTe n = 20; BaMo n = 138). The dosing and formulation of compounds varied across groups (Table 1). One study³⁹ gave additional dosing instructions to be taken with food as evidence suggested it may improve absorption.³⁹ One study had an adjuvant treatment of cognitive training exercises.³⁸

Overview of Neurocognitive Domains and Tests

Cognitive assessments varied significantly across studies (Table 1), and as such outcome data extraction for meta-analysis were limited. One study⁴⁰ adapted existing measures but gave limited description and when contacted did not have the requisite records describing how they altered the original measures. Subsequently, a Clinical Psychologist trained in the psychometric assessment of neurocognition verified what data to extract and combine for meta-analyses.

Domains Assessed

Processing speed, the rate at which simple cognitive operations are executed, was assessed using the Simple Reaction Test, Symbol Digit Modalities Test, Coding Copy, and Choice Reaction Test.

Attention, the capacity to maintain focus and process relevant stimuli, was measured with Digit Span Forwards, Spatial Span Forwards, the Choice Reaction Test, and the Choice Discrimination Test.

Working memory, the ability to hold and manipulate information over short periods, was evaluated using Digit Span Backwards, Spatial Span Backwards, the Sternberg Test, and Memory Span for Nonsense Syllables.

Memory, encompassing verbal and visual learning and recall processes, was examined with Paired Associates Learning, Visual Recognition Memory, Verbal Episodic Memory – Cued Recall, the Auditory Verbal Learning Test, and the Logical Memory Test.

Executive function, which supports inhibition, cognitive flexibility, and planning, was assessed using the Stroop Test, Trail Making Test Part B, Tower of London, Word Fluency, Inhibition Cats and Dogs, the Wiegel Sort Test, and the Sternberg Test.

Language was measured with the Boston Naming Test, Word Fluency, and the Language Comprehension – Token Test.

Psychomotor function, reflecting sensorimotor speed and coordination, was assessed through the Motor Screening Test and Finger-Tip Tapping.

Adaptive functioning (ABAS Total Score) and neurophysiological measures (resting-state fMRI and TMS) were included in 1 study as supplementary indicators.

Risk of Bias of Included Studies

Bias was low to moderate across all 7 studies (supplementary information Table S1; supplementary Figure S1). There was significant variation in the comparisons of baseline characteristics, and no studies explored all relevant factors (supplementary information Table S2). All studies checked compliance and had a minimum standard of at least 80% compliance, though compliance methods varied.

Effects of the Interventions on Cognition

Reliable change was prespecified as the preferred outcome; however, no eligible study reported reliable change estimates or sufficient data to derive them, necessitating reliance on end point means.

The C. sinensis study differed from the remaining trials in both population and co-intervention and was therefore interpreted separately and not combined with pooled analyses of healthy adult populations.

Overall meta-analytic data are presented in supplementary Figure S2.

Bacopa Monneiria

No statistically significant differences were found between BaMo and placebo on processing speed (PS) overall (95% CI, −0.16 to 0.36, P = .45), or on any individual tests (P > .05); attention overall (95% CI, −0.28 to 0.17, P = .65), or on individual tests (P > .05); working memory (WM) overall (95% CI, −0.32 to 0.16, P = .50) or on individual tests (P > .05); on a speed of comprehension test⁴² (95% CI, −0.48 to 0.67, P = .75); on the Finger-tip tapping assessment⁴² of psychomotor speed (95% CI, −0.20 to 0.45, P = .55); or on EF overall (95% CI −0.04 to 0.35, P = .11) or on individual tests (P > .05).

The combined Auditory Verbal Learning Test (AVLT) forgetting scores demonstrated significant heterogeneity (I² = 84%, chi² = 6.30, P = .01) and were not included in the overall meta-analysis for memory and explored separately. When separately meta-analyzed, 1 study⁴¹ demonstrated no significant differences between BaMo and placebo groups (95% CI, −0.60 to 0.84, P = .74) on AVLT forgetting scores, whereas the other study⁴² demonstrated a statistically significant benefit of taking 150 mg of BaMo, twice daily for 84 days on AVLT forgetting scores (95% CI, −2.73 to −0.47, P = .006).

No other significant differences were demonstrated between BaMo and placebo groups either overall (95% CI, −0.22 to 0.17, P = .80) or on individual tests (P > .05).

When all cognitive tests for BaMo were combined into meta-analysis, no statistically significant differences were found between BaMo and placebo overall (95% CI, −0.15 to 0.07, P = .46; I² = 0%, chi² = 17.70, P = .89; Figure 2).

Figure 2.

Forest plot comparing Bacopa monnieri and placebo across all cognitive tests investigated among the included studies.

Camellia Sinensis

In a study of healthy participants, no statistically significant differences were found between CaSi supplement and placebo on the following: a PS test (Simple Reaction Time Task Mean Latency³⁸; 95% CI, −0.64 to 0.22, P = .34); attention overall (95% CI, −0.34 to 0.26, P = .78), or on any individual tests (P > .05); on WM overall (95% CI, −0.21 to 0.40, P = .54) or on individual tests (P>.05); language overall (95% CI, −0.05 to 0.45, P = .11) or on individual tests of language (P > .05); on Motor Screening Test speed scores³⁸ (95% CI, −223.13 to 49.13, P = .21); on memory overall (95% CI, −0.04 to 0.35, P = .11) or on any of the individual memory tests (P > .05); or on EF overall (95% CI, −0.04 to 0.35, P = .11) or on any of the individual tests (P > .05).

When all cognitive tests for CaSi were combined into meta-analysis, no statistically significant differences were found between CaSi and placebo overall (95% CI, −0.05 to 0.13, P = .39; I² = 0%, chi² = 15.74, P = .73; Figure 3).

Figure 3.

Forest plot comparing Camelia sinensis and placebo across all cognitive tests investigated among the included studies.

Ocimum Tenuiflorum

In young healthy males, a statistically significant benefit was found in favor of OS extract over placebo on Stroop Inhibition Reaction Time (95% CI, −3.75 to −1.95, P < .00001; Sampath et al., 2015).¹⁶ Error rate scores were not analyzed as the data were not computed as per test manual instructions and therefore cannot be meaningfully interpreted.

Secondary Outcomes

Bacopa Monnieri

One participant⁴⁰ reported incidence of adverse gastrointestinal reaction to BaMo. One participant⁴¹ in the placebo arm reported back pain. One study⁴² reported adverse events were mainly mild and comparable between groups; no statistical analysis was performed but a greater percentage of nausea, dry mouth, and fatigue was experienced in the BaMo versus the placebo arm. One study⁴³ noted 5 withdrawals due to adverse events (BaMo n = 3, placebo n = 2) but did not provide details.

Camellia Sinensis

One study³⁸ examined adaptive functioning using the Adaptive Behavior Assessment System 2^nd edition.³³ There were no significant differences found between the CaSi and placebo groups (95% CI, −0.51 to 0.35, P = .71).

One study³⁸ examined the impact of CaSi on functional connectivity using MRI. They examined 18 participants with Down syndrome (CaSi n = 10, placebo n = 8) and demonstrated significant enhancement of regional functional connectivity with increases in the functional integration of cortical and subcortical distributed networks, including the frontal cortex; Wernicke's area; the precuneus, occipitotemporal cortices, somatosensory cortices, and basal ganglia in the CaSi versus placebo group, with resting-state functional MRI. In addition, following non-invasive TMS stimulation, cortical excitability normalization was seen in the CaSi versus placebo group. Of note, this study was not powered to assess the effect of CaSi on these outcomes. No deaths were reported. Adverse events were reported to be mild and not related to the treatment, with no statistically significant differences between CaSi and placebo groups.

Ocimum Tenuiflorum

No adverse events were reported.¹⁶ No other secondary outcomes were examined.

DISCUSSION

We sought to identify and evaluate the efficacy of various phytochemicals on cognition in adults of working age who are healthy or who have nonprogressive neurocognitive compromise. The secondary objective was to highlight the quality of research evidence available. Seven eligible studies were included; 1 for CaSi, 1 for OcTe, and 5 for BaMo. From a number of psychological outcomes, only 1 significant result was extracted for OcTe, namely, the speed at which an inhibitory control task can be performed.

Cognitive Effects

Bacopa Monnieri

BaMo demonstrated no benefit or detriment to PS, attention, WM, language, psychomotor, or EF within a heathy population. This is generally consistent with other reports.^44–46 However, despite similar dosing and supplement formulation, Kongkeaw et al., 2014 report improved reaction times from group with a greater age range compared with our study (18-74 years) and included those with “mild cognitive impairment” (MCI), which may explain the discrepancy.²³ The greater success in older and MCI populations may be due to the higher likelihood of pathology at such an age. The potential for ROS-related changes or loss of acetylcholine and protein aggregation associated with MCI may provide more opportunity for the mechanism of action of the compounds to take place. This may also be an artefact of the poor statistical methods applied.^47–49 Furthermore, a study investigating the impact of BaMo in a younger population (4-18 years) demonstrated improvements in WM (Kean et al., 2016); however, they also included studies that do not meet the requirements of an RCT or quasi-RCT,³⁵ failed to publish data on quality or bias assessment of studies and also included studies with poor statistical methods.^46,49–51

The effects of BaMo on memory function remain inconclusive because “rate of forgetting” data could not be included in the overall meta-analysis due to significant heterogeneity across studies. When rate of forgetting data were examined individually, inconsistent findings between the studies were found; one demonstrated a significant benefit,⁴³ while another showed no benefit.⁴¹ Neither of these studies compared differences in cognition between the BaMo and placebo group at baseline. Aside from rate of forgetting data, all other effects of BaMo on memory function were insignificant, which is generally consistent with previous reviews.^23,44,45 Two of the aforementioned reviews that included older and MCI populations found benefits in free recall, likely underpinned by similar reasons as those provided for reaction time. Further and more robust methodological exploration of the effect of BaMo on memory is likely required to resolve the inconsistent findings and clarify the validity of this review’s results of no effect.

Camellia Sinensis

CaSi appeared to demonstrate no significant effect on cognitive functioning within a Down syndrome population. This absence of benefit is consistent with a previous systematic review examining long-term use in both clinical and healthy populations.⁵² However, Camfield et al., 2014 did observe improved attention function⁵³ when investigating only short-term effects (hours) of dosing rather than routine use. Further, statistical analysis demonstrated the caffeine within the CaSi supplement used was most likely responsible for the improved attention result.⁵³ Caffeine has long-established short-term benefits on attention, though these effects are short lived, likely due to the pharmacokinetic and metabolic timeline of caffeine itself along with the temporal effects of caffeine on the gating of adenosine.^54,55 This short-term impact would explain the absence of significant effect in both this and the previous review,⁵² which examined longer-term outcome. This review revealed no other studies that looked at the effects of C. sinensis, specifically in the form of green or black tea, in a healthy adult population. Interpretation of these findings should remain cautious, as EGCG was administered alongside cognitive training in both trial arms; outcomes therefore cannot be attributed to C. sinensis or EGCG in isolation. Notably, no cognitive benefit was observed even under these conditions.

Ocimum Tenuiflorum

OcTe demonstrated a beneficial effect in the speed at which an inhibitory control task (the Stroop test) is completed. The OcTe study only presented means and SDs for speed of completion and not for the error rate or the computed result, which would be considered the gold standard. When reported in isolation, the speed score may not take into consideration active compromise of performance (reported via the error rate) to improve performance on speed; thus, little can be inferred from an improvement of speed in this task. No other significant benefits or adverse impacts on EF were seen for OcTe supplementation. No other cognitive domains were examined, highlighting lack of available data and limiting conclusions on OcTe’s efficacy as a cognitive enhancer.

Combined, the cognitive results of this review highlight the importance of meta-analysis; 6^{16,38–40,42,43} of the 7 studies examined concluded individually there was significant benefit for their respective phytochemicals on numerous aspects of cognition, whereas this review demonstrates limited benefit for cognitive function when effect sizes were calculated and results pooled.

Adaptive Function, Neurophysiology, and Adverse Events

De la Torre et al. examined the impact of CaSi on adaptive functioning.³⁸ While no significant effect was observed, this finding, to the author’s knowledge, is novel as no previous meta-analyses have explored this factor before.^52,53 Despite this, De la Torre et al. reported improved connectivity between cortical and subcortical structures that support several cognitive functions following CaSi use. In addition, normalization of excitability of frontal cortex following TMS stimulation was observed. These findings are consistent with previous electroencephalography findings within the frontal cortex⁵² and could potentially be explained by mechanistic evidence of CaSi’s neuromodulatory effect on monoamines.^56–58 Subsequent effects on behavior were not observed by this review, but neurophysiological change often precedes behavioral effect,^59–61 as such longer-term assessment may demonstrate benefit.

All compounds appeared to be well tolerated, with the majority of participants demonstrating good compliance (>80%) and with mild gastrointestinal problems being highlighted as the only notable consequence. However, this finding should be taken with caution as one study⁴¹ failed to report adverse event data and another demonstrated a near 50% attrition rate and did not report reasons, which may have been related to adverse events.

Limitations

This is the first review, to our knowledge, to examine the possible effects of “cognitive enhancing” phytochemicals in healthy individuals while also taking into consideration the importance of a variety of baseline characteristic comparisons and their potential impact on clinical outcome. None of the studies included were deemed to be of low quality or high risk of bias. However, the results of this study should be interpreted with caution due to the limited number of studies being included and therefore the overall small sample size. Overall, the uncertainty of risk of bias necessitates caution in the interpretation of these data.

There were significant flaws in the statistical comparisons made in all of the studies included. Only 1 study attempted to examine change from baseline,³⁸ while all other studies compared end-point data between groups. There was significant variation in the comparisons of baseline characteristics, and no studies explored all relevant factors, which could have an influence and introduce bias to end-point clinical outcome.⁶² One study did explore difference from baseline, though there was no accounting of statistical confounds, such as test–retest variability or regression to the mean,⁶³ potentially leading to the presentation of statistically significant results that are not clinically significant, as they may be due to statistical artefacts.^49–51

The cognitive outcomes varied significantly across studies, limiting data extraction. Some centers validated their own measures of function, while others adapted existing measures without description. This significantly reduces reliability of trials and makes meta-analysis complex, preventing the availability of good quality evidence.

Another limitation is the lack of biomarker data for cognition, evidence of mechanisms (e.g., reduction in biomarkers of OS), or validation of bioavailability of the drugs (e.g., serum samples). Only 1 study examined neurophysiological correlations (or biomarkers) of cognitive function through fMRI and TMS. Absence of these markers, alongside absence of effect on cognitive data, may help validate their lack of efficacy.

Some pooled analyses included multiple cognitive outcomes from individual studies, which may increase the risk of unit of analysis bias and overestimate precision. Although the pooled findings were nonsignificant and therefore unlikely to reflect false positive effects, results should be interpreted cautiously as exploratory summaries.

This review was not preregistered, which may increase the risk of unrecognized bias; however, the review question, eligibility criteria, outcomes of interest, and analytic approach were defined in advance and adhered to throughout and are fully described.

One included study employed adapted or incompletely described cognitive measures; although expert review was used to assess construct comparability and these outcomes did not drive pooled results, their inclusion may introduce measurement uncertainty.

No formal adjustment for multiple comparisons was applied due to the exploratory nature of the analyses and the heterogeneity of outcomes; however, given the predominantly null findings, the risk of type I error influencing the overall conclusions is likely to be limited.

Lastly, only 1 study assessed adaptive functioning; it has been consistently highlighted that changes in cognition that do not translate to adaptive behavior are of limited efficacy in improving quality of life,^{32–34,64,65} making this an essential secondary outcome in cognitive therapeutic studies.

Future Directions

Additional trials are needed to examine the effects of promising phytochemical such as OcTe, BaMo, and CaSi on cognitive health in younger, healthy individuals but also on selected target populations. Research is warranted to explore the potential therapeutic benefits of additional herbs traditionally used such as C. longa, C. asiatica, Ganoderma lingzhi, and Salvia rosmarinus. A recent clinical trial demonstrated that a herbal formulation containing turmeric and rosemary improved the speed of cognitive tasks in older adults, suggesting there may also be synergistic effects of such phytochemical.⁶⁶ Such studies should aim to combine pharmacokinetic, neuroimaging, and neuropsychological assessment techniques to determine the neurobiological nature of any effects on behaviors and cognitions observed. Neuropsychopharmacological mechanism of action studies should be performed alongside any future trials reviewing efficacy in cognitive impairment of phytochemical. This work will facilitate onward drug development and innovation. Studies should restrict cognitive assessment to widely used measures to enhance replicability. Only in exceptional cases, with documented justification, should they adopt measures or use novel assessments.

CONCLUSIONS

A limited number of studies performed demonstrated a lack of benefit of BaMo or OcTe on cognitive functioning in healthy adults of working age. Additionally, limited evidence indicated no benefit of CaSi within a Down syndrome population. BaMo, OcTe, and CaSi were well tolerated and caused no serious adverse effects. Despite the limitations of the current clinical evidence base, further explicitly designed, double-blind RCTs are warranted, alongside neuropsychopharmacological mechanism of action studies, to explore the use of phytochemical compounds as cognitive enhancers.

Author Contributions

A.M.: Searches, analysis, quality appraisal, primary manuscript preparation, and review handling; D.Q.: Searches and selection, quality appraisal, manuscript editing, and review handling; M.M., M.H., V.R.: Inception, supervision, search criteria appraisal, manuscript review, and editing.

Alexander Marsh (Conceptualization [equal], Data curation [lead], Formal analysis [lead], Investigation [lead], Methodology [lead], Project administration [lead], Validation [lead], Visualization [lead], Writing—original draft [lead], Writing—review and editing [lead]), Darren Quelch (Investigation [supporting], Methodology [equal], Project administration [equal], Validation [supporting], Writing—original draft [supporting], Writing—review and editing [supporting]), Marion Mackonchie (Conceptualization [equal], Methodology [equal], Project administration [equal], Resources [equal], Supervision [equal], Writing—review and editing [equal]), Momna Hejmadi (Conceptualization [equal], Supervision [equal], Writing—review and editing [equal]), Vivien Rolfe (Conceptualization [equal], Methodology [equal], Resources [equal], Supervision [lead], Writing—review and editing [equal]).

Funding

D.Q. is supported by a Sandwell and West Birmingham NHS Trust research fellowship; this work was supported by Pukka Herbs Limited.

Conflicts of Interest

Authors A.M., D.Q., and M.H. have no conflicts to declare. Author M.M. works for Pukka Herbs, a herbal tea manufacturer. M.M. is employed by a commercial herbal manufacturer; however, literature searches, study selection, quality appraisal, data analysis, and final interpretation were led by the academic authors who have no conflicts of interest. All contributions transparently reported via CRediT authorship.

Data Availability

https://osf.io/6tse9/overview?view_only=034c96998bb646f89ec856bc1362215f

Ethics

Not applicable.