Cognitive effects of two nutraceuticals Ginseng and Bacopa benchmarked against modafinil: a review and comparison of effect sizes

Authors


Correspondence

Professor Andrew Scholey PhD, Centre for Human Psychopharmacology, Swinburne University of Technology, Melbourne, Vic. 3122, Australia.

Tel.: +61(3) 9214 8932

Fax: +61(3) 9214 5525

E-mail: andrew@scholeylab.com

Abstract

Over recent years there has been increasing research into both pharmaceutical and nutraceutical cognition enhancers. Here we aimed to calculate the effect sizes of positive cognitive effect of the pharmaceutical modafinil in order to benchmark the effect of two widely used nutraceuticals Ginseng and Bacopa (which have consistent acute and chronic cognitive effects, respectively). A search strategy was implemented to capture clinical studies into the neurocognitive effects of modafinil, Ginseng and Bacopa. Studies undertaken on healthy human subjects using a double-blind, placebo-controlled design were included. For each study where appropriate data were included, effect sizes (Cohen's d) were calculated for measures showing significant positive and negative effects of treatment over placebo. The highest effect sizes for cognitive outcomes were 0.77 for modafinil (visuospatial memory accuracy), 0.86 for Ginseng (simple reaction time) and 0.95 for Bacopa (delayed word recall). These data confirm that neurocognitive enhancement from well characterized nutraceuticals can produce cognition enhancing effects of similar magnitude to those from pharmaceutical interventions. Future research should compare these effects directly in clinical trials.

Introduction

Over recent years there has been increasing interest in the area of cognitive enhancement. This field has traditionally focused on the effects of so-called nootropics in those with fragile cognition (e.g. those with age-related cognitive decline, mild cognitive impairment or dementia). However in recent years the field of cognition enhancement has also embraced studies into cognition enhancement in young, cognitively intact individuals [1]. In the context of pharmaceutical cognition enhancers, methylphenidate and modafinil (Provigil) have taken centre stage [2]. Modafinil had been taken as a cognitive enhancer by around 10% of respondents to an online survey of the readership of the journal Nature [3], though it is worth noting that this study may have a high response bias from individuals using cognitive enhancers.

As well as pharmaceutical approaches to cognitive enhancement, there is growing interest in the possibility that certain nutraceuticals may enhance cognitive performance. Herbal extracts may contain multiple active components which, in concert, may influence numerous neuronal, metabolic and hormonal systems involved in behavioural processes [4]. Additionally the interactions between these actives may be synergistic, resulting in complex dose and time dependent effects. These factors are challenging for psychopharmacology, making certain positive effects fragile even where there is strict batch to batch consistency across studies, and rendering negative findings sometimes difficult to interpret. Nevertheless there is growing evidence that certain standardized natural products have reproducible neurocognitive effects in humans, possibly because of their inherent polyphamacological properties. Examples of the most promising nutraceuticals in this context include species of Salvia (sage) [5-8], Panax ginseng [9-12] and Bacopa monniera [13-17]. For the purposes of this short review we chose to concentrate on Ginseng and Bacopa. The former has been shown to enhance robustly cognitive performance after acute dosing [18], whilst Bacopa has been shown to have nootropic effects with chronic dosing [16].

We are not aware of any study to date which has directly compared the cognition enhancing effects of a nutraceuticals and a pharmaceutical in a head to head trial in healthy volunteers. The purpose of this review is to examine the literature describing clinical trials of the neurocognitive effects of modafinil, Panax ginseng and Bacopa monniera in order to compare their effect sizes across different cognitive domains.

Methods

Search strategy

The electronic databases SCOPUS, PubMed and PsychInfo were accessed in early 2012 (April). Key word searches were conducted by combining ‘Ginseng’, ‘Bacopa’ or ‘modafinil’ with ‘cognition’, ‘memory’, ‘neuropsychological’, ‘neurocognitive’ and ‘executive function’. Articles that did not relate to human cognitive trials were excluded as were articles that were not in English. Accepted articles were those that had completed double-blind, randomized, placebo-controlled empirical investigations on healthy human subjects using cognitive function as a primary outcome. Additionally studies were included only if they described an intervention with at least one arm assessing one of Ginseng, Bacopa or modafinil. In the case of Ginseng and Bacopa, only studies using ‘pure’ extracts were accepted, i.e. there were no other supplements present within the target nutraceutical arm. Furthermore, all extracts must have been used in isolation and not contaminated by co-use of other supplements as adjunct interventions.

Papers identified as meeting eligibility criteria were analyzed for emerging common cognitive domains between the intervention types. Applying these criteria resulted in eight relevant studies for modafinil, seven for Bacopa, and nine for Ginseng.

Effect size analysis

Effect sizes were calculated on statistically significant data in order to assess the magnitude of effects. Effect sizes were calculated using Cohen's d [19]. The strength of clinical effects were defined as small: d = 0.2, medium: d = 0.5 and large: d = 0.8 effect sizes as defined by Cohen [19]. Chronic trials utilizing a repeated measures design were analyzed at endpoint and not mid-points. Effect sizes are presented in Tables  1-3. It should be noted that all treatment-associated benefits (e.g. increased accuracy, shorter reaction times) are presented as positive effect sizes and impairments (e.g. more errors, slower reaction times) as negative effect sizes respectively.

Table 1. Summary of human trials into the cognitive effects of modafinil
Author/YearnDoseSampleTrial designOutcome measuresResultsCohen's d effect size
  1. CO, crossover; COWAT, controlled oral word association test; DB, double-blind; DMTS L, delayed matching to sample latency; DMTS, delayed matching to sample; DRN, detection of repeated numbers; DS BAS, digit span backward actual span; DS BS, digit span backward score; DS FAS, digit span forward actual span; DS FS, digit span forward score; DT, deliberation time (gamble); IED, intra/extra dimensional set shift; MD, methamphetamine-dependent; PC, placebo-controlled; PG, parallel groups; PRM, pattern recognition memory; RVIP, rapid visual information processing; S RT, Stroop reaction time; SOC, Stockings of Cambridge; VAS, visual analogue scale; VSM E, visual spatial memory error rates; VSM RT, visual spatial memory reaction time; VSM, visual spatial memory.
Turner et al. (2002) [23]60100 mg (n = 20) or 200 mg (n = 20)Placebo mean age 25.30 years, 100 mg mean age 24.35 years, 200 mg mean age 25.10 years

Acute

DB, PC, PG

Decision making (gambling), RVIP, digit span, visual spatial working memory, pattern recognition, stop signal, stroop colour naming

Self rated mood and wakefulness.

Significant improvements from drug on digit span, Stroop colour naming, spatial planning, gambling deliberation and pattern recognition.

Suggestion of reduction in impulsive responsiveness.

DS FAS0.12
DS BAS0.09
DS FS0.12
DS BS0.09
PRM0.09
DMTS L0.09
NTOL0.11
DT0.08
N.B. Effect sizes calculated as average mean and SD of low and high dose combined in keeping with the analysis used by the authors
Randall et al. (2003) [30]30100 mg or 200 mgPlacebo mean age 20.7 years, 100 mg mean age 20.3 years, 200 mg mean age 20.7 years

Acute

DB, PC, PG

Motor screening, RVIP, DMTS, IED, SOC, trail making, Stroop, COWAT, clock drawing.

VAS mood, sleep and wakefulness

No significant cognitive effects

Significant improvements in self-reported mood and wakefulness

No cognitive effects
Müller et al. (2004) [28]16200 mg20–29-year-olds, mean age 24.1 years

Acute

DB, PC, CO

Visual spatial working memory, trail-making, state anxiety, numeric working memory

Significant decreases in error rates at longest delay (8000 ms) in visual spatial task with modafinil, as well as faster reaction times in 1000 and 4000 ms delays.

No speed–accuracy interaction

VSM E 8000 ms0.77
VSM RT 1000 ms0.21
VSM RT 4000 ms0.15
Baranski et al. (2004) [26]184 mg kg−1 – average 300 mg18–40 years, mean age 24.2 years

Acute

DB, PC, CO

Choice reaction time, numeracy, DRN, visual comparison, logical reasoning, self rated mood and fatigueImprovements in reaction time and vigilance as well as fatigue and motivation.Effect sizes unavailable – raw data not present in research paper
Randall et al. (2004) [27]45100 mg or 200 mg50–67 years, placebo mean age 55.7 years, 100 mg mean age 58.8 years, 200 mg mean age 58.2 years

Acute

DB, PC, PG

Motor screening, RVIP, DMTS, IED, SOC, trail making, Stroop, COWAT, clock drawing, VAS mood, sleep and wakefulness

No improvements in executive functioning (trail making, SOC)

Improvements in Stroop response time, more pronounced with 200 mg dose

No improvements in self-reported mood and wakefulness

100 mg 
Stroop RT0.16
200 mg 
Stroop RT0.45
Stoops et al. (2005) [51]6100, 200, 400 mg

22–28 years, mean age

24 years

Acute

DB, PC, CO

Wakefulness and stress measures, mental arithmetic (×3 50 min blocks)

Improved self-reported stress and wakefulness measures.

Improved mental arithmetic performance

Effect sizes unavailable – raw data not present in research paper
Winder-Rhodes et al. (2010) [25]12300 mg

18–39 years

Mean age

26.3 years

Acute

DB, PC, CO

Digit span, RVIP, SOC, Pattern recognition, stop-signalPlacebo and prazosin (used as another data point) decreased accuracy on SOC, whilst modafinil did not exert this negative effect. However, the effects of modafinil were not statistically significantNo significant effects
Ghaharemi et al. (2011) [29]19200 mgMean age 33.7 years

Acute

DB, PC, CO

Part of a larger study using MD participants and fMRI

Abstract picture associative learning

No effect on performance of the task in the healthy control groupEffect sizes unavailable – raw data not present in research paper
Table 2. Summaryof human trials into the cognitive effects of Panax ginseng
Author/YearnDoseSampleTrial designOutcome measuresResultsCohen's d effect size
  1. 3BRT, 3-back reaction time; 3BSI, 3-back sensitivity index; CO, crossover; CRT, choice reaction time; CRTA, choice reaction time accuracy; DB, double-blind; DVA, digit vigilance accuracy; DVFA, digit vigilance false alarms; DVRT, digit vigilance reaction time; DWRA, delayed word recall percentage accuracy; IWR, immediate word recall (% accuracy); MF, mental fatigue; NMC, numeric working memory percentage change; NWMRT, numeric working memory reaction time; PC, placebo-controlled; PR, picture recognition; PRRT, picture recognition reaction time; RT, reaction time; RVIP FA, rapid visual information processing false alarms; RVIP RT, rapid visual information processing reaction Time; RVIP, rapid visual information processing; S3 E, serial threes errors; S3, serial threes correct; S7 E, serial sevens errors; S7, serial sevens correct; SMRT, spatial memory reaction time; SMRT, spatial memory reaction time; SRT, simple reaction time; WR, word recognition; WRRT, word recognition reaction time.
Kennedy et al. (2001) [10]20200, 400, 600 mg (G115)20–27 years, mean age 21.3 years

Acute

DB, PC, CO

CDR testing battery, self-reported mood measures

400 mg lead to improvements on quality of memory and secondary memory at all time point post dose.

Significant negative effect on attention in the 200 and 600 mg doses

200 mg 
IWR 4 h−0.17
DVFA 6 h0.60
DVRT 4 h0.50
DVRT 6 h0.40
CRT 1 h0.36
CRT 4 h0.15
CRT 6 h0.22
SMRT 1 h0.17
SMRT 2.5 h0.05
PR 1 h−0.01
PRRT 6H0.19
400 mg 
IWR 1 h0.02
IWR 2.5 h−0.15
IWR 4 h−0.12
IWR 6 h0.12
SRT 6 h0.52
DVA 1 h0.37
DVA 4 h0.44
DVFA 4 h0.20
DWRT 2.5 h0.47
WR 2.5 h0.08
WR 6 h−0.10
PR 1 h−0.01
600 mg 
IWR 1 h0.13
IWR 2.5 h−0.01
IWR 4 h0.03
SRT 4 h0.86
SRT 6 h0.64
DVA 2.5 h−0.34
DVFA 4 h0.17
DVRT 4 h0.31
DVRT 6 h0.62
WR 2.5 h−0.06
WR 6 h−0.28
Kennedy et al. (2002) [37]20400 mg (G115)Mean age 21.1 years

Acute

DB, PC, CO

CDR testing battery, S3, S7, Bond–Lader

Improvements in secondary memory (CDR score), speed of memory task performance and attentional accuracy.

No effects of G115 on other measures in the CDR battery

IWR 4 h0.54
IWR 6 h0.77
DVFA 2.5 h0.37
DVFA 4 h0.41
CRTA 1 h0.16
CRTA 2.5 h0.23
SMRT 2.5 h0.39
NMC 2.5 h−0.24
NMC 6 h−0.30
DWRA 2.5 h−0.06
DWRA 6 h0.63
WR 2.5 h0.03
WR 4 h−0.08
Scholey et al. (2002) [18]20200, 400, 600 mg (G115)Mean age 19.9 years

Acute

DB, PC, CO

S3, S7

No effects on S3

Fewer subtractions at all time points following 200 mg dose on S7

400 mg showed significant improvement in accuracy with reduced amount of errors

Effect sizes unavailable – raw data not present in research paper
Kennedy et al. (2004) [36]28200 mg (G115)Mean age 21.4 years

Acute

DB, PC, CO

CDR testing battery, S3, S7, sentence verification, logical reasoning, self-reported mood measuresImprovement on speed of attention and speed of memory (CDR scores)DVRT 6 h0.41
CRT 3 h0.15
CRT 4 h0.02
NWMRT 2.5 h0.32
NWMRT 4 h0.29
NWMRT 6 h0.39
WRRT 1 h0.01
WRRT 4 h0.11
PRRT 4 h0.02
Reay et al. (2005) [11]30200 and 400 mg separate dosesMean age 22.6 years

Acute

DB, PC, CO

Mental fatigue (self-rated), RVIP, S3, S7.

RVIP, S3, S7 are administered as a 10 min ‘cognitive demand battery’ and is repeated at hourly intervals at baseline and up to 6 h post-dose

More significant effects using the 200 mg dose. Significant reduction of mental fatigue scores as well as increased performance in S7 in all except one time point (5 h). Significant positive effects shown in RVIP and S3.

400 mg dose showed significant increase in mental fatigue at 3 h post dose as well as modest improvements on RVIP and S3 error rates

200 mg 
MF 2 h0.20
MF 3 h0.59
MF 4 h0.31
MF 5 h0.90
MF 6 h1.03
RVIP RT 5 h0.53
RVIP 6 h−0.11
S3 4 h0.27
S3 E 4 h0.17
S3 E 5 h0.39
S7 1 h0.34
S7 2 h0.36
S7 3 h0.45
S7 4 h0.35
S7 6 h0.39
400 mg 
MF 3 h0.61
RVIP RT 6 h0.29
RVIP 5 h−0.07
S3 6 h−0.09
S3 E 2 h0.17
Sünram-Lea et al. (2005) [39]30200 mg18–25 years, mean age 20 years

Acute

DB, PC, CO

CDR testing battery, self-report mood measures

Improvements on ‘speed of attention’ component of CDR battery from the choice reaction time task.

No other measure was significant

CRT0.33
Reay et al. (2006) [12]27200 mg

Mean age

21.89 years

Acute

DB, PC, CO

Mental fatigue (self-rated), RVIP, S3, and S7.

RVIP, S3, S7 are administered as a 10 min ‘cognitive demand battery’ and is repeated at hourly intervals at baseline and up to 6 h post dose.

Mixed effects on mental fatigue

Significant improvements on RVIP performance and reduced false alarms.

Significant improvements on serial 3 s at 4 and 6 h, but significant reduction in performance at 3 h

MF 5 h1.08
MF 6 h1.40
RVIP 4 h0.24
RVIP FA 6 h0.21
S3 3 h0.28
S3 4 h0.22
S3 6 h0.28
S3E 6 h0.03
Reay et al. (2009) [41]30200 and 400 mgMean age 22.87 yearsDB,PC,CO

Acute and sub-chronic (8 days) testing sessions.

Multiple tests including corsi-blocks, 1/2/3/4 back and random number generation

Reduced reaction time on 3-back reaction time at 400 mg, but increases in reaction time with 200 mg

Increased 3-back sensitivity index scores with 400 mg, but decreased with 200 mg

Improved self-reported mood levels

Day 1 
200 mg 
3BRT 4 h−0.25
400 mg 
3BRT 2.5 h0.81
3BSI 1 h0.40
3BSI 4 h0.30
Day 8 
200 mg 
3BRT 1 h−0.10
3BRT 2.5 h−0.48
3BSI 4 h−0.43
400 mg 
3BSI 1 h0.04
3BSI 4 h0.16
Table 3. Summary of human trials into the cognitive effects of Bacopa monniera
Author/YearnDoseSampleTrial designOutcome measuresResultsCohen's d effect size
  1. Note: 0 in trial design column refers to baseline testing, i.e. prior to ingestion of first tablet. AAMI, age associated memory impairment; AE, adverse events; AVLT, auditory verbal learning test; AVLT1AVLT learning rate; AVLT2, AVLT forgetting rate; AVLT3, AVLT pro-active interference; AVLT4, AVLT delayed recall; AVLT5, AVLT total learning; AVLT6, AVLT retroactive interference; BC, bacoside content; BM, BacoMind; CDR, cognitive drug research; CDRI, Central Drug Research Institute, Lucknow, India; DAT, divided attention task; DB, double-blind; DS, digit span; DSST, memory task and digit symbol substitution test; DWP, delayed word pairs; IT, inspection time; KM, KeenMind; MH, MediHerb; MMSE, Mini Mental State Examination; PC, placebo-controlled; RT, reaction time; RVIP, rapid visual information processing; SCT, speed of comprehension test; SDMT, symbol digit modalities test; TMT, Reitan trail making test; VS, visual span; WAIS, Wechsler adult intelligence scale; WM, working memory (CDR component); WMS, Wechsler memory scale.
Stough et al. (2001) [17]

46

23(B)

23(P)

KM

300 mg daily

BC: min 55%

18–60 years

12 week

DB, PC

IT, AVLT, RT, DSImproved IT & three AVLT subtests (learning rate, proactive interference and forgetting rate) at 12 weeks compared with PL. RT & DS did not improveIT:0.25
AVLT1,:0.27
AVLT2,:0.52
AVLT3,:1.01
Nathan et al. (2001) [51]

38

18 (B)

20 (P)

KM

300 mg single dose

BC: min 55%

18–60 years0 and 2 h DB, PCAVLT; DS; DSST: SDMT: SCT; TMT; RT; ITNo improvements made on all measures.NS
Roodenrys et al. (2002) [15]

76

37 (B)

39 (P)

KM

300 mg (subjects <90 kg)

400 mg (subjects >90 kg)

BC:55%

40–65 years0 and 3 months DB, PCMemory questionnaire, story recall, general knowledge, WP, DS, VS, codingImproved recall of delayed WP task. No improvements in DS, VS, coding, general knowledge, memory questionnaire or story recallDWP :0.23
Raghav et al. (2006) [52]

40

20 (B)

20 (P)

CDRI

250 mg daily

BC: min 55%

55+ years

AAMI ppts

0 and 12 weeks DB, PCWMSDifferences between treatment groups were not statistically detailedN/S
Stough et al. (2008) [16]

62

33 (B)

29 (P)

KM

300 mg daily

BC: min 55%

18–60 years0 and 90 days DB, PCCDR testing batteryImproved two CDR subtests (RVIP and working memory); Remaining nine subtests showed no improvementsWM:0.47
RVIP:0.30
Calabrese et al. (2008) [13]

44

24 (B)

24 (P)

MH

300 mg daily

BC: min 50%

65+ years0 and 12 weeks DB, PC

Primary outcome: AVLT

Secondary outcomes: Stroop, DAT, WAIS letter digit test

Significant AVLT subtest (delayed word recall) and Stroop improvements. No improvements on DAT, WAIS letter digit testAVLT4,:0.36
Stroop:0.32
Barbhaiya et al. (2008) [53]

44

23 (B)

21 (P)

BM

450 mg daily

BC: bacoside A3 >5% w/w

50–75 years

MMSE 24+

0, 12 and 24 weeks DB, PC

Treatment between 0–12 weeks

AVLT, WMS, WAIS, cognitive testing battery administered after 12 weeks treatment and 24 weeks follow-upComparisons not between treatment groups but change from baseline in each groupNo effect sizes available
Morgan & Stevens (2010) [14]

81

36 (B)

45 (P)

BM

300 mg daily

BC: 40–50%

55+ years

0 and 12 weeks

DB, PC

Multiple cognitive measures: AVLT, CFT, TMT, MAC-Q

Significant improvements on verbal learning, memory acquisition and delayed recall AVLT measures

No significant differences on CFT, TMT or MAC-Q

AVLT4:0.95
AVLT5:0.53
AVLT6:0.57

Results

Modafinil

Modafinil (C15H15NO2S [20]) is a pharmaceutical drug used as a licensed treatment for excessive daytime sleepiness associated with narcolepsy or shift-work [2]. The mechanisms responsible for its effects remain largely unknown. It appears to exert a wide range of effects including via modulating catecholamine activity [21]. The human pharmacokinetic profile is known, with peak effects between 2–4 h after oral ingestion and a half-life of 12–15 h [22].

As noted in a review by Repantis et al. [2], as well as its use in the context of sleep disorders, there has been an increase in academics and office workers using modafinil as a cognitive enhancer. Research into the cognitive enhancing effects of modafinil generally falls into one of two types: studies in sleep deprived and non-sleep deprived human subjects. These have typically compared doses of 100 mg and/or 200 mg doses against placebo. As modafinil is used as a treatment for excessive daytime sleepiness, there is a large body of research assessing cognitive effects in sleep-deprived participant groups. However, the results from these studies are not comparable with the participant groups used to assess the cognitive effects of Bacopa or Ginseng so are not included here.

In non-sleep deprived adults, modafinil is associated with improvements in accuracy of pattern recognition and the stop signal task following 100 mg and 200 mg [23], with several studies showing improvements in digit span with the 100 mg dose alone [23, 24]. Furthermore, modafinil improved accuracy of an executive planning task (Stocking of Cambridge) [25]. Faster reaction times have also been shown across a range of tasks, notably the Stroop colour naming task of selective attention [23, 24, 26, 27]. There are also numerous tasks that are unaffected by modafinil, regardless of dose, including trailmaking [28], mathematical processing [26], spatial working memory [24], logical memory [27], associative learning [29] and verbal fluency [30].

Effect sizes restricted to those domains significantly affected by modafinil are presented in Table 1 (note that all benefits are presented as positive effect sizes).

Panax ginseng

Ginseng refers to extracts from the Araliaceae family of plants. It is estimated that, in the US, Ginseng is the second most used psychoactive herbal product [31]. The active components are believed to be the ginsenosides, of which over 30 have been isolated, though many exist in trace amounts [32]. The chemical structure of these aglycone saponins can be used to classify the ginsenosides into three groups: the protopanaxadiol group (e.g. Rb1, Rb2, Rb3, Rc, Rd, Rg3, Rh2, Rs1), the protopanaxatriol group (e.g. Re, Rf, Rg1, Rg2, Rh1) and the oleanolic acid group (e.g. Ro) [33].

Over the last decade or so a number of studies have revealed that single doses of Panax ginseng (also known as Asian ginseng) can modulate aspects of cognitive function [9] such as brain activity as measured by electroencephalography [34] and peripheral blood glucose concentrations [11, 35], in healthy young volunteers. Most of these studies have used a standardized extract (G115) which contains an invariant 4% ginsenosides.

With regards to cognitive function, a number of controlled studies have identified both positive and negative behavioural effects. The most consistent finding, however, is one of improved secondary memory (i.e. declarative memory involving recollection) performance following extract G115 alone [10, 36, 37] and in combination with both Ginkgo biloba [37] and Paullinia cupana (guaraná) [36]. In addition, Panax ginseng (G115) has been shown to enhance aspects of working memory [38], to improve mental arithmetic performance (in a task that loads heavily on working memory resources) [11, 12] and to speed attentional processes [39] in healthy volunteers. These benefits to reaction time occurred without a concomitant cost to accuracy, precluding the possibility of a treatment related shift in the speed/accuracy trade-off. One recent study has shown that acute administration of a standard extract of a different Ginseng species, Panax quinquefolius (American ginseng), which has a ginsenoside profile distinct from that of Panax ginseng, can also improve working memory performance [38].

Despite growing evidence supporting the efficacy of Panax ginseng (G115) in modulating cognitive processes following a single dose, only three empirical studies have directly investigated the cognitive and mood effects following more extended Ginseng ingestion periods (with only two of these studies using the standardized G115 extract). Two early studies revealed improved speed of performing a mental arithmetic task following 12 weeks administration of Panax ginseng (200 mg G115 day–1) in young volunteers [40]. The most recent, Reay et al. [38, 41] found both positive and negative effects of 7 days dosing with G115. Whilst there were beneficial effects of the 400 mg dose on various measures of the 3-back task there were also negative effects on reaction time limited to the 200 mg dose.

Effect sizes restricted to those domains significantly affected by Ginseng from selected publications are presented in Table 2.

Bacopa monniera

Bacopa monniera (BM), known as Bacopa or Brahmi, is a herb from the Scrophulariaceae family of plants which has been used for centuries in Ayurvedic medicine. BM has been shown to contain a complex mixture of constituents including alkaloids, saponins and flavonoids. The key constituents of Bacopa are thought to be the triterpenoid saponins, bacosides A and B [42]. These bacosides usually co-occur, differing only their optical rotation, with the presence of bacoside B believed to be an artefact generated during the isolation of bacoside A [43]. Animal studies have shown BM to be an antioxidant [44], memory enhancer [45], antidepressant [46] and to reduce the concentrations of beta-amyloid in a mouse model of Alzheimer's disease [47].

Human studies reveal consistent cognitive enhancement as a result of BM administration across young, old and impaired adult populations [48]. Unlike modafinil and Ginseng, Bacopa may not acutely improve cognitive functioning (although there are as yet unpublished reports of acute effects during more effortful cognitive processing). To date publications are restricted to effects which are evident only after chronic interventions (typically 12 weeks of a 300 mg daily dose), with no significant improvements occurring after a 5 week intervention [17] or acutely after 2 h [49]. The most robust effects of BM are on memory performance, including positive effects on learning and consolidation of target stimuli [17], delayed recall [13], total learning [14], visual retention of information [15] and working memory [16]. There is also evidence that BM can improve speed of information processing in both the inspection time task and rapid visual information processing [16, 17]. Using BM in an older population group (55 years and over) has shown improvements in executive functioning on the Stroop task and the mental control subtest of the Wechsler memory scale [13, 14]. There are also tasks which appear to be unaffected by BM administration including working memory speed [17], reaction time [13, 16, 17] and divided attention [13]. Effect sizes for significant findings restricted to those domains significantly affected by Bacopa are presented in Table 3.

Effect sizes

With regards to modafinil, our analysis revealed small effect sizes on reaction time and small to large effect sizes on response accuracy during visual-spatial working memory, as well as a dose-dependent increase in effect size for stroop reaction times (Table 1). There appeared to be no cognitive costs (i.e. cognitive impairments in other domains) associated with these effects, with effect sizes ranging from d = 0.083 to d = 0.774 (for accuracy of visual spatial memory).

There was large variation in the effect sizes for Ginseng (Table 2). The largest effect size, d = 1.396, was for amelioration of self-rated mental fatigue during heavily loaded cognitive processing. Regarding cognitive processing the largest effect size was for simple reaction time (d = 0.860). Interestingly another measure, reaction time on the 3-back working memory task, showed a large positive effect size (d = 0.806) following acute Ginseng administration and also had the largest impairment d = −0.481 [41] following dosing for 8 days. This raises the possibility of neuroadaptations to the neural substrates influenced by acute Ginseng administration with longer term dosing resulting in opposite effects to acute dosing.

Chronic BM interventions generally produced the most consistent and largest effect sizes. These ranged from small to medium effect sizes for measures of attention and information processing tasks such as RVIP, Stroop and inspection time. Larger effect sizes were evident for auditory verbal learning tasks where the effect sizes ranged from d = 0.230 for delayed word pair memory to d = 0.950 for delayed word recall (AVLT4) and d = 1.01 for protection from pro-active interference during delayed memory (AVLT3).

Conclusions

All three substances reviewed here exerted overwhelmingly positive effects on neurocognitive function across different cognitive domains. However, it must be made clear that this review has only looked at the statistically significant results and did not include the non-significant results from studies into the neurocognitive effects of these substances. Modafinil had the strongest effects on speed of information processing and executive functioning. Ginseng exerts acute positive effects on secondary memory and more heavily cognitively loaded working memory tasks. Bacopa administration appears to predominantly enhance learning and memory, with effects restricted to chronic administration. The differential effects on cognitive domains presumably reflects different mechanisms of action of each substance. Modafinil has multiple effects on neurotransmitter systems including region specific increases in adrenergic, histaminergic and glutaminergic activity and decreased GABAergic activity [50]. The mechanisms for BM are unknown, but there is evidence that it has pro-cholinergic effects as suggested by the effects on inspection time and rapid visual information processing [17, 51] as well as anti-oxidant and anti-inflammatory properties. As might be expected from an extract with multiple components Ginseng has been reported to have multiple properties relevant to neurocognitive function. These include glucoregulation, modulation of cholinergic and dopaminergic activity as well as increasing nitric oxide synthesis [19]. Whilst the purpose of this review was to compare nutraceuticals and pharmaceuticals, it is worth noting that the majority of the BM research presented here relates to chronic administration whilst the modafinil and Ginseng research is predominantly in the context of acute administration. However, the purpose of this review is to assess substances that enhance cognition.

With regards to modafinil, the results here show the cognitive enhancing effects of the substance on non-sleep deprived subjects, rather than during the treatment of excessive tiredness in narcolepsy or after sleep deprivation. It is also worth noting that many research papers using non-sleep deprived individuals have not adequately reported the cognitive results so we were unable to compute effect sizes for these studies (although this factor is by no means limited to modafinil research). Regarding the effects of Ginseng and Bacopa, research into herbal medicines brings its own difficulties. For example human studies into BM have used different products across trials. Whilst the manufacturers of all BM products included in our review claim standardization, compositions of individual treatments have not been compared. However, all extracts used in studies to date are reported to have standardized bacoside content to levels between 50–55%. All Ginseng studies included in Table 2 used the standardized extract G115.

This review is limited by the number of studies currently available. As the number of double-blind, placebo-controlled, randomized trials is rising, it is likely that future reviews will be able to compare particular groups of individuals such as those with specific cognitive dysfunction or particular age groups. At present, the number of available studies is too low to make such direct comparisons. Future research studies may wish to compare directly the differences of these substances in the same cohort. We are also aware that some of the research into nutraceuticals is not widely available through popular search engines such as those used for this review.

In conclusion the nutraceuticals Ginseng and Bacopa produced effect sizes for cognitive enhancement which were comparable with those seen for modafinil, albeit in different cognitive domains. Future studies should directly compare the cognitive effects of these agents in direct, head to head clinical trials. Furthermore, presentation and statistical analysis of results in certain research papers have made calculating effect sizes difficult in some instances. This is an issue that will need to be addressed in future studies into both pharmaceutical or nutraceutical research aiming to establish any cognitive enhancing effects of said interventions.

Competing Interests

All authors have completed the Unified Competing Interest form at http://www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declare no support from any organization for the submitted work, AS and CS have received funding from Soho Flordis International (SFI) in the previous 3 years and there are no other relationships or activities that could appear to have influenced the submitted work.

Ancillary