Effects of Nutrition on Cognitive Function in Adults with or without Cognitive Impairment: A Systematic Review of Randomized Controlled Clinical Trials

New dietary approaches for the prevention of cognitive impairment are being investigated. However, evidence from dietary interventions is mainly from food and nutrient supplement interventions, with inconsistent results and high heterogeneity between trials. We conducted a comprehensive systematic search of randomized controlled trials (RCTs) published in MEDLINE-PubMed, from January 2018 to July 2021, investigating the impact of dietary counseling, as well as food-based and dietary supplement interventions on cognitive function in adults with or without cognitive impairment. Based on the search strategy, 197 eligible publications were used for data abstraction. Finally, 61 articles were included in the analysis. There was reasonable evidence that dietary patterns, as well as food and dietary supplements improved cognitive domains or measures of brain integrity. The Mediterranean diet showed promising results, whereas the role of the DASH diet was not clear. Healthy food consumption improved cognitive function, although the quality of these studies was relatively low. The role of dietary supplements was mixed, with strong evidence of the benefits of polyphenols and combinations of nutrients, but with low evidence for PUFAs, vitamin D, specific protein, amino acids, and other types of supplements. Further well-designed RCTs are needed to guide the development of dietary approaches for the prevention of cognitive impairment.


Introduction
Dementia represents a serious public health challenge for elderly people [1], while Alzheimer's disease (AD), the most common type of dementia, has become one of the biggest mental burdens [2]. Except for the recent and controversial approval of Aduhelm to treat patients with Alzheimer's disease, there are no effective treatments to prevent or delay its progression, only for treating the symptoms of mild to moderate dementia [3]. Therefore, there is an urgent need to identify effective strategies to prevent dementia. It

Materials and Methods
This systematic review was conducted following the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guidelines (http://www.prisma-statement.org/ (accessed on 21 October 2021). Refs. [23,24] and the review protocol was registered with PROSPERO (PROSPERO 2021: CRD42021245941).

Eligibility Criteria
The eligible studies had to fulfill the following criteria: (1) they were RCTs, including crossover and parallel designs; (2) they assessed the impact of a dietary intervention (i.e., diet counseling intervention, food-based intervention, or dietary supplement trial) in adults with or without cognitive impairment but without previous diagnoses of AD; (3) they were published in English and available in PubMed between 1 January 2018 and 31 July 2021. The reason for the selection of this period was that previous systematic reviews had been performed on this topic before [13,21,22]. Studies were excluded if they involved multifactorial lifestyle interventions, in which effects of dietary factors could not be distinguished from other lifestyle factors (such as multidomain or physical training intervention); they used probiotics in the form of pills (not in a food matrix); or they were observational studies, case reports, comments, letters, editorials, in vitro studies, animal studies, duplicate studies, narrative or systematic reviews and study protocols. Additionally, assessment of cognitive functioning could include the application of different neuropsychological batteries and/or specific questionnaires, as well as the assessment of brain integrity, as revealed by structural, functional and metabolic changes. Studies containing combined interventions (i.e., nutritional plus physical activity or behavioral interventions) were excluded unless the nutritional effect on the outcomes was analyzed separately from other interventions. When there was more than one study of the same trial, the publication with the largest sample size and longest follow-up was selected.

Literature Search
A literature search was carried out using the MEDLINE-PubMed database. The search was limited to the period of 1 January 2018 to 31 July 2021. For this we employed the algorithm described in Supplementary Files. The search strategy used keywords and selected terms related to diet, specific food or food groups, nutrients and nutritional supplements, combined with terms related to cognitive impairment, Alzheimer's disease, and dementia.

Screening and Data Extraction
Screening of eligible studies and data extraction were performed independently by three independent reviewers (L.G., A.F., and M.R.), and disagreement between the three authors was resolved by consulting senior researchers (C.P. and M.B.).
Searching for duplicate publications was performed using an electronic database and then the full text of each potentially relevant study was reviewed. A specific grid was adopted to facilitate data extraction and selection, including the following information: the author's surname, year of publication, country where the study was conducted, characteristics of the trial participants (cognitive health, sample size based on participants randomized, and age), study design (including the design of a parallel or crossover intervention and blinding), duration, type of the intervention and the comparator (including dose of exposure), study outcomes and major findings. Cognitive function was assessed by using different neuropsychological batteries and/or specific questionnaires, as well as measures of brain integrity, as revealed by structural, functional and metabolic changes. the Cambridge Neuropsychological Test Automated Battery (CANTAB), the Consortium to Establish a Registry for Alzheimer's Disease neuropsychological battery (CERAD.NB), etc. Regarding brain integrity, fMRI methods assessing blood oxygen level-dependent (BOLD) and (18F) FDG PET scans were considered to assess brain glucose metabolism.

Quality Assessment
The quality assessment was based on extracted data using the second version of the Cochrane Risk-of-Bias instrument (RoB 2) [25] and discrepancies were resolved by consensus between two additional authors. Risk-of-Bias was assessed in terms of the following domains: (1) randomization process; (2) deviation from intended interventions; (3) missing outcomes data; (4) measurement of the outcome; (5) selective outcome reporting; and (6) overall bias. Each domain was considered to be of "low risk", "some concern", or "high risk" of bias. According to RoB 2, a result of "high risk" for any individual domain leads to an overall high risk of bias, while a result of "some concern" in any individual domain is judged as an overall risk of "some concern". If multiple domains are problematic, confidence in the outcome decreases and the overall risk should be considered "high" [25,26].

Data Synthesis
Studies were grouped according to their design (i.e., food-based intervention, dietary counselling intervention or dietary supplement intervention) and secondly according to outcome. The review emphasized the impact of the identified nutritional interventions on participants' cognitive health, which was assessed through changes in mean values of measures of cognitive function, standard deviation or 95% confidence interval, odds ratio and/or relative risk.

Search and Selection of Studies
Through a database search, 197 articles were identified (Figure 1). After examining titles, abstracts and full-texts, 83 articles were considered to be out of the scope of this review and were excluded. Further screening was performed by reading full text and 53 articles were non-eligible due to: participants having a diagnosis of AD (n = 2), studies without nutritional interventions (n = 12), multi-domain interventions (n = 14), not assessing cognition (n = 16), interventions administrating drugs or probiotics in pill form (n = 6), and manuscripts with commentaries or protocols (n = 3). Finally, a total of 61 articles met the inclusion criteria and were finally considered in the present systematic review.

Diet Counselling Interventions (n = 5)
The effect of the Mediterranean diet enriched with a single food was evaluated in 3 different studies. A 8-week RCT with a crossover design revealed improvements in processing speed in 33 older cognitively healthy subjects who followed a Mediterranean diet enriched with fresh lean pork (2-3 servings/week) compared to a low-fat diet [27]. Similarly, the same researchers evaluated the effect of a Mediterranean diet enriched with 3-4 servings/day of dairy products in 43 older cognitively healthy subjects, showing a significant enhancement in processing speed in the Mediterranean diet group compared to a low-fat diet group [28]. The effect of a Mediterranean diet enriched with either high phenolic extra virgin olive oil (50 mL/d) or moderate phenolic extra virgin olive oil (50 mL/d), or a Mediterranean diet alone for 12 months was also evaluated in 50 Greek older adults during 12 months. Participants allocated in both olive oil intervention groups showed significant improvements in cognitive function compared to the control group, with a greater effect in the high phenolic extra virgin olive oil consumption group [29]. Additionally, a 6-month RCT was conducted in 79 adults with cognitive impairment but no dementia and those who followed the DASH diet did not reveal significant improvements on general cognition [30]. The influence of adhering to Finnish Nutrition Recommendations for 4 years was also tested in men and women aged 57-78 in a parallel RCT with no beneficial effects on cognition compared to the control group [31].

Food-Based Interventions (n = 7)
Six food-based intervention studies were conducted on cognitively healthy participants. In a parallel intervention study carried out in 61 subjects over 60 years, the con-sumption of 3 g/d of matcha green tea powder for 12 weeks improved cognitive and memory function [32]. Similarly, daily avocado consumption (140-175 g/d) for 3 months ameliorated attentional inhibition in 47 young adults [33]. In a longer trial conducted over 52 older adults, consumption of 100 g/d of ultra-high hydrostatic pressurized brown rice for 2 years reduced overall cognitive decline compared to white rice intake [34]. Similarly, a crossover study of 31 older adults showed a reduction in overall cognitive decline after consuming 3 servings/d of dewaxed brown rice for 6 months [35]. In contrast, an acute crossover study comparing the consumption of rye-based bread with white wheat flour bread for 3 days did not find any effect neither in working memory nor in attention [36]. In a recent long-term (2 years) trial with 708 older participants, the consumption of 60 g/d walnuts had no any beneficial effect in global cognition [37]. Only one study has evaluated the effect on animal-based food consumption on cognition. In this regard, the consumption of 33.4 g/d of camembert cheese or the same amount of processed cheese made from mozzarella and cream cheese for 3 months did not produce improvements in general cognitive functions in 65 subjects with MCI [38].

Dietary Supplement Interventions (n = 49)
Forty-nine trials were categorized according to the type of supplementation into proteins and amino acids, vitamins, polyphenols, PUFAs, combinations of nutrients and other supplements.

Protein and Amino Acid Supplements (n = 7)
A randomized, double-blind, placebo-controlled 12-week trial evaluated the effect of combined anserine and carnosine supplementation (750:250 mg/d) in 25 MCI subjects. Those subjects assigned to the active group showed a better response to the global Clinical Dementia Rating (gloCDR), without showing any beneficial effect on the other psychometric tests including the Mini-Mental State Examination (MMSE), the Wechsler Memory Scale, and the Alzheimer's Disease Assessment Scale (ADAS). However, when APOE4+ or APOE4-subjects were separately analyzed, improvements in MMSE and gloCDR were observed in the APOE4+ subjects [39]. A study conducted over 268 community-dwelling adults without dementia showed that administration of 200 µg/d of a casein-derived peptide Met-Lys-Pro (MKP) for 24 weeks significantly improved the ADAS-Cog subscale, but not the ADAS-Cog total score compared to the placebo group [40]. Similarly, the intake of lactotripeptide tablets containing 1.4 mg/d valine-proline-proline and 2.0 mg/d isoleucine-proline-proline for 8 weeks also did not produce significant improvements in cognitive assessments in 35 middle-aged and older adults [41]. A parallel trial investigating the effect of taurine supplementation (150 mL/d water containing 1.5 g taurine) for 14 weeks also failed to show significant benefits on cognition in older women [42]. The only crossover trial with amino acid supplementation conducted in 20 young volunteers receiving either 125 mg/d of L-homoarginine for 4 weeks or placebo, did not find changes neither in vascular nor in neuronal function in young volunteers [43]. The effect of 1 g/d of whey peptide supplementation for 12 weeks resulted in higher verbal fluency test (VFT) scores in the intervention group compared with the placebo. A further subgroup analyses according to the degree of subjective fatigue showed that changes in the VFT as well as the Stroop and subjective memory function tests were significantly better in subjects with high-level fatigue from the whey peptide group [44]. Finally, another RCT was conducted on 19 male subjects consuming either β-alanine (12 g/d) or placebo for 14 days prior to a 24-h simulated military training (SUSOP) during which participants received cognitive assessment. Individuals in the placebo group showed worse performance in the visuomotor training task compared to the active group [45].

Vitamin Supplements (n = 4)
Three RCTs evaluated the effect of vitamin D on cognition and one study evaluated B group vitamins. A 3-year trial conducted in 260 older postmenopausal African-American women showed no significant difference in cognition over time, based on the MMSE score, after the intake of D3 vitamin (daily dose of 2400 IU, 3600 IU, or 4800 IU to achieve and maintain serum levels of 30 ng/mL) [46]. A study in 181 subjects with MCI evaluated the supplementation of a daily dose of 400 IU vitamin D3 for 12 months and demonstrated an overall significant improvement of cognitive functioning in the vitamin D3 group compared to placebo [47]. Another 6-month double-blind, four-arm placebo, controlled trial evaluated the effects of vitamin D supplementation delivered either as enhanced vitamin D2 (UV-exposed mushroom, D2M) or D3 (synthetic) for 24 weeks compared to standard mushroom (SM) and placebo (PL) controls on cognition and mood in 436 healthy subjects aged over 60 years. There were no significant effects of treatment on any of the measures of cognitive function or mood, thus not supporting any benefit of vitamin D supplementation for mood or cognition [48].
Regarding B-group vitamins, vitamin B9 (400 µg/d) and vitamin B12 (25 µg/d) were administered alone or in combination (400 µg/d + 25 µg/d B9 and B12, respectively) to 180 subjects with MCI for 6 months. The folic acid plus vitamin B12 supplementation significantly improved the Full-Scale IQ, Information and Digit Span scores. Although the supplementation with vitamin B9 and vitamin B12 significantly improved cognition, the combination of folic acid and vitamin B12 was superior to either folic acid or vitamin B12 alone [49].

Polyphenols Supplements (n = 10)
The acute consumption of cocoa flavanols was tested in 11 young adults with type-1 diabetes (T1D) and their matched healthy controls (15 mg in healthy and 900 mg in T1D subjects) 2 h before a flanker test. The authors found faster reaction times and greater BOLD signal amplitude in the supramarginal gyrus and inferior frontal gyrus compared to the placebo in both groups [50]. In another larger trial including 101 healthy adults aged 40-60 years with overweight and obesity, the long-term (24 weeks) supplementation with anthocyanin-rich Aronia melanocarpa extract (90 mg and 150 mg) shown a significant improvement in psychomotor speed but not in the rest of cognitive tests [51].
One of the target populations for polyphenol supplementation was healthy older adults. In this case, curcumin supplementation (80 mg/d curcumin) in 38 subjects showed a beneficial effect on working memory after 12 weeks, as well as improved mood and lower fatigue at both 4 and 12 weeks of the trial [52]. Similarly, the consumption of 258 mg/d of flavonoids from a polyphenol-rich extract of grape and blueberry for 6 months in 92 subjects improved age-related episodic memory impairment in those subjects with greater cognitive impairment [53]. The effect of consuming two cups of Montmorency tart cherry juice improved memory and performance on learning tasks after a 12-week randomized controlled trial in 37 older adults [54].
One double-blind controlled trial was performed in 90 young subjects with SCD. The effect of the spearmint extract intake was evaluated at two different doses (600 or 900 mg/d) for 90 days. Results indicated that subjects consuming 900 mg/d dose significantly improved working memory compared to the placebo group [55].
There were also four studies with 12-week interventions, one with Lactobacillus plantarum C29-fermented soybean consumption (800 mg/d) showing enhanced cognitive performance and attention in a parallel trial in 100 MCI subjects [56], another with Cosmos caudaus supplement (500 mg/d) that potentially improved global cognition function in 23 MCI patients [57]. A crossover trial evaluated the effect of continuous intake of chlorogenic acid (1.107 g/d) in 28 MCI patients, showing an enhanced attention and executive function [58]. Additionally, a 6-month trial conducted in 36 MCI patients investigated the effect of Biokesum, a registered Persicaria minor extract supplement rich in polyphenols (500 mg/d), supplementation and observed improved visual memory, negative mood and bilateral dorsolateral prefrontal cortex activation [59].

PUFAs Supplements (n = 5)
A double-blinded placebo-controlled parallel trial enrolled 555 officers who daily consumed 8 capsules of krill oil (2.3 g/d of omega-3) or macadamia nut oil containing between 53% and 67% oleic acid and 16% and 24% palmitoleic acid (control group) for 20 weeks. The results did not support a significant beneficial effect of supplemented krill-based omega-3 fatty acids on cognitive function [60]. In the MAPT study, 1680 French community-dwellers aged 70 or over reporting subjective memory complaints received either w-3 PUFAs supplementation (800 mg/d of DHA and ≤225 mg/d of EPA) alone, a multidomain intervention (nutrition, physical activity and cognitive training), w-3 PUFAs supplementation plus the multidomain intervention or placebo for 3 years. Participants who consumed PUFAs had no significant improvement in cognition compared to placebo [61][62][63]. Moreover, 33 of the participants from the MAPT study underwent 2-(18F)fluoro-2-deoxy-D-glucose (FDG) PET and received either V0137 (800 mg/d of an active supplement containing a minimum of 65% DHA and a maximum of 15% EPA) or placebo for 12 months, although non-significant effects on cognition were observed [64].

Combination of PUFAs, Polyphenols and Vitamins (n = 8)
The effect of combining PUFAs, polyphenols and vitamins has also been evaluated. The consumption of omega-3 PUFAs, catechins and ginsenosides (960 mg/d EPA, 624 mg/d DHA, 16 mg/d ginsenosides and 26 mg/d green tea catechins) for 1 month induced enhanced cognitive functioning and increased fMRI functional connectivity during task execution in 10 healthy older adults [65]. In contrast, combining omega-3 PUFAs with vitamin E (1.72 g/d DHA and 0.6 g/d EPA, mixed with tocopherols at 11.2-18 mg/g) exerted a small negative effect on psychomotor speed but a significant increased perception of self-reported cognitive failures compared to participants in the control group who were consuming 4 capsules/d, each containing 990 mg of low polyphenol olive oil and 10 mg of fish oil, equivalent to 1.8 mg EPA and 1.2 mg DHA [66]. Eighteen older adults with MCI supplemented with a combination of omega-3 and omega-6 fatty acids, and vitamins (4.14 g DHA/0.810 mg EPA, 1.8 g GLA/3.15 g LA, 0.6 mg vitamin A, 22 mg alfa-tocopherol, 760 mg gamma-tocopherol per day) for 24 weeks, significantly reduced cognitive function and functional capacity [67]. Another trial with a three-arm design assigned 152 cognitively healthy participants to take fish oil (2 g/d DHA + 0.4 g/d EPA), or curcumin (160 mg/d), or a combination of both, or placebo for 16 weeks. The research showed that only in males was fish oil supplementation able to improve processing speed, while curcumin intake showed benefits in working memory. However, the combination of fish oil with curcumin did not produce additional benefits [68]. A similar 24-week RCT, with an additional 24-week post-intervention follow-up was carried out in 94 older adults with SCD who were assigned to consume fish oil (1.6 g/d EPA, 0.8 g/d DHA) + placebo powder (24 g/d), blueberry powder (25 g/d) + placebo oil, fish oil + blueberry powder or placebo (placebo oil + placebo powder). The study demonstrated that both treatments were associated with an enhancement of perceived functional capability that was maintained beyond the 24-week discontinuation phase as well as reduced interference in memory for those receiving blueberry powder supplementation. However, there were no positive outcomes in cognition after following the combined treatment [69]. A fourarm double-blind, placebo-controlled trial conducted in 240 MCI adults evaluated the effect of folic acid (FA) + DHA (FA 800 µg/d + DHA 800 mg/d), FA (800 µg/d), DHA (800 mg/d) and placebo for 6 months, showing that FA, DHA and their combination significantly improved cognitive function, although FA + DHA was more beneficial than each individual nutrient on their own [70]. In contrast, there were no significant differences in cognition measures among 33 cognitively healthy older adults who were provided with a vitamin B complex (1 mg of vitamin B12, 100 mg of vitamin B6 and 800 µg of folic acid per day) and randomized to either 2152 mg/d of DHA or placebo over 6 months [71]. Finally, cerebral perfusion was assessed in a subset of 13 subjects from 49 MCI patients, aged 50-80 years, who were given a combination of PUFAs and vitamin E (daily intake of 1320 mg EPA, 880 mg DHA and 15 mg vitamin E) or placebo (sunflower oil) for 26 weeks, showing an improvement of cerebral perfusion in the treatment group [72].

Other Types of Dietary Supplements (n = 15)
Two double-blind RCTs used terpenoids as a dietary supplement with recognized antioxidant properties. Consumption of astaxanthin and sesamin (6 mg/d and 10 mg/d, respectively) for 3 months significantly improved psychomotor and processing speed in a study of 21 MCI patients [73]. In contrast, auraptene-enriched juice (6 mg/d of auraptene) for 24 weeks did not provide any extra benefit for global cognitive functioning as assessed by MMSE in 41 cognitively healthy adult subjects [74]. Other single dietary supplements or their combinations have been tested either in cognitively healthy adult subjects or in subjects with some degree of cognitive decline with contradictory results. In cognitively healthy subjects, 5 intervention studies have evaluated the effect of fungi and algae extracts, other plant extracts, or combination of minerals and vitamins. A 12-week parallel intervention study in 31 cognitively healthy adults allocated to consume either a combination of an extract of Eleutherococcussenticosus (ES) and rhizomes of Drynariafortunei (DR) (203.01 mg/d of ES leaf extract, and 20.01 mg/d of DR), showed significant improvements in language, semantic fluency and figure recall scores [75]. Consumption of Hericiumerinaceus supplement (3.2 g/d) for 12 weeks also improved cognitive function and prevented cognitive decline in a study of 34 cognitively healthy subjects over the age of 50 years [76]. The effect of fermented Laminaria japonica was also evaluated in 60 moderately active senior subjects randomized to consume either 1.5 g/d of this supplement or a sucrose pill as placebo. After 6 weeks of intervention, several neuropsychological tests were performed and a significant improvement in most of them was observed for the intervention group [77]. In contrast, another study in 44 cognitively healthy subjects (minimum age was 40 years old) who consumed 10 mL/d of an ethanol extract of traditional herbal medicines sage, rosemary and melissa) or 10 mL/d of placebo for 2 weeks, and did not find significant differences between treatment and placebo, only significant improvements to delay word recall in a subgroup analysis in subjects under 63 years [78]. Finally, an ancillary study nested within the Personalized Prevention of Colorectal Cancer Trial was conducted in 250 subjects with a calcium intake ≥700 mg/d and lower than 2000 mg/d and a Ca:Mg ratio ≥ 2.6. Participants were randomly allocated either to consume a personalized dose of Mg supplementation to reduce the Ca:Mg intake ratio around 2.3 or to the placebo group for 12 weeks. Subjects who reduced the Ca:Mg ratio by the personalized treatment significantly improved global cognitive function assessed by the MoCa test [79]. Additionally, two other RCTs were conducted in cognitively normal participants. A 6-week parallel RCT in 54 high-perfectionist men showed the consumption of milk protein concentrate drink enriched with phospholipids, sphingomyelin, phosphatidylcholine and phosphatidylethanolamine (250 mL/d with 2.7 g of phospholipids) enhanced performance on switch attention task [80]; while 33 elderly nursing home residents allocated to receive medium-chain triglycerides (6 g/d) either alone or in combination with L-Leucine (1.2 g/d) and cholecalciferol (20 µg/d) for 3 months significantly improved their cognition compared to the control group (6 tg/d long-chain triglycerides) [81]. Other studies with different dietary supplements or their combinations have also evaluated their potential beneficial role on cognition in subjects with subjective cognitive decline. In this case, two 12-week studies conducted by the same investigators on subjects with subjective cognitive decline revealed improvements in verbal fluency, executive functions, memory retrieval and enhanced attention after matured hop bitter acid consumption. One of the trials was conducted in 60 adults (aged 45-64 years) who received either orally administered matured hop bitter acids (35 mg/d) derived from Humulus lupulus L, a traditional herb used in beer brewing, or placebo for 12 weeks [82]. Treatment improved verbal memory and executive function compared to participants in the placebo group [82]. The same supplement was tested in 49 adults (45-69 years), increasing the Symbol Digit Modalities Test score and also memory retrieval in the subgroup with SCD [83]. The effect of low (600 mg/d, n = 30) and high-doses (1200 mg/d, n = 30) of Tremella fuciformis (TF) consumption on cognition was evaluated in 60 young adults after 4 and 8 weeks of intervention, and subjects in the intervention group showed fewer self-reported cognitive complaints, better short-memory and executive functions compared to placebo in a non-dependent dose manner. Along with this, TF supplementation was related to increases in gray matter volumes of specific brain regions [84]. Another study conducted in 30 older subjects (60-80 y) evaluated the administration of a daily dose of 750 mg of spermidine-rich plant extract containing 1.2 mg of the natural polyamine plus 510 mg cellulose compared to the placebo containing 750 mg potato starch and 510 mg cellulose and revealed a moderate enhancement of memory performance and mnemonic discrimination ability after 3 months of intervention [85]. The combination of different dietary supplements on cognition in subjects with cognitive impairment has also been evaluated in 2 different studies with different intervention durations (6 and 12 months). The first study was a multicenter trial conducted in 99 participants randomized to receive multinutrient capsules (3 capsules/d, each capsule containing 250 mg DHA, 40 mg EPA, 5 mg vitamin E, 15 mg phosphatidylserine, 90 mg tryptophan, 125 ug vitamin B, 250 ug folate and 60 mg ginkgo biloba) or placebo for 12 months. Subjects in the active group had no significant improvement in global cognitive function when compared to the placebo group [86]. The second study was a multicenter, randomized, double-blind and placebo-controlled trial evaluating the effect of Twendee X, a patented supplement with 8 antioxidants, on MMSE and Hasegawa Dementia Scale-revised test (HDS-R) in 78 MCI patients who were randomized to the treatment or placebo group for 6 months. Cognitive assessment exhibited a significant improvement in the treatment group compared to placebo [87].

Risk of Bias Assessment
In terms of quality assessment, 48% of the studies were judged at low risk of bias, 13% with some concerns and 39% at high risk of bias (see Figures 2-4). Many of the lowquality studies did not inform on the randomization process or on the blinding of outcome assessors, especially regarding diet counselling and food-based interventions. Some others did not evaluate adherence to the intervention or presented deviations from the intended interventions. Notably, only 39% of the studies followed an intention-to-treat assessment ( Figure S2), whereas 61% followed a per-protocol assessment ( Figure S3), which is generally considered a high-risk factor for bias. Moreover, studies with a small sample size were usually assessed at high risk of bias, whereas large sample sizes were often judged to be at low risk of bias ( Figure 5).

Discussion
The aim of the present systematic review of RCTs was to update and provide the most comprehensive analysis to date of the effectiveness of dietary patterns, specific foods and nutritional supplements on cognitive health outcomes in subjects without symptoms of dementia. Overall, significant improvement in cognitive functions was observed in most (n = 44) but not all (n = 17) studies, irrespective of treatment type (diet, food, supplement), duration or cognitive status (cognitively healthy subjects or cognitively impaired older adults).
Diet has been suggested to have an important role in cognitive health and the development of dementia [88]. However, most of the intervention studies conducted to date mainly assessed the effect of dietary supplements on cognitive function, while few trials were designed to include specific foods or dietary patterns. Within the dietary patterns, the Mediterranean diet is probably the most investigated and with the most promising results [89]. A sub-analysis previously conducted in the framework of the PREDIMED trial showed that following a Mediterranean diet enriched with olive oil or nuts improved global cognition, memory and attention after 4 years of follow-up [90]. Since 2018, 3 RCTs evaluating the effect of following a Mediterranean diet enriched with either fresh lean pork (MedPork study) or dairy products (MedDairy study) or high phenolic early harvest extra virgin olive oil (HP-EH-EVOO) in cognitively healthy subjects have been conducted, showing increased processing speed as compared to a low-fat diet [27,28] and better scores on global cognition, attention and fluency tasks compared to a Mediterranean diet not supplemented with high-phenolic EVOO [29]. However, because in the MedPork [27] and MedDairy [28] studies the control groups followed a low-fat diet instead of a Mediterranean diet, it cannot be elucidated whether the observed benefits were due to the synergistic role of the healthy components of the Mediterranean diet, rather than the addition of fresh pork or dairy products. Similarly, in the DR's EXTRA study, no difference in the 4-year change in global cognition was observed between the diet group and the control group, and this could be due to the fact that diet quality also improved in the control group, which could confound the effect of a healthy diet [31]. In addition, the DASH diet, with similar dietary composition to the Mediterranean diet, but with a higher intake of low-fat dairy products, potassium and less sodium, was shown in previous observational studies to be associated with improved cognition and lower risk of cognitive impairment [91]. This is probably due to the fact that the DASH diet shares similar characteristics with the Mediterranean diet, in addition to reduced sodium intake, which is associated with improved blood-brain barrier integrity [92,93]. The lack of significant effects in Blumenthal et al., might be due to the low number of participants in the intervention group and/or the relatively modest adherence to the diet regime [30].
Some key foods of the Mediterranean diet, such as EVOO, but also other plant-based foods including brown rice, matcha green tea and avocado, have also shown significant improvements in different cognitive domains such as speed of execution [34], orientation and language [34,35], and attention and global cognition [33][34][35]. Surprisingly, the Walnuts and Healthy Aging study, conducted at two centers, found no consistent effect of walnut consumption on global cognition in all participants. However, when data was analyzed separately by recruitment centers, an improvement in global cognition and perception scores was observed in participants recruited in Barcelona but not in the Loma Linda subjects. These differences between centers were potentially attributed to lower levels of education and lower plasma α-linolenic acid concentrations in the Barcelona participants at baseline [37]. Animal-based foods, such as camembert cheese, worsened cognition, even when levels of the neuroprotective BDNF were high [38], but, in contrast, intake of bovine milk-derived phospholipid was shown to protect cognitive performance under stress in terms of better sustained attention. Taken together, the effects of phospholipid supplementation may be limited to subjects with high levels of stress, as they attenuate stress response mediated by the HPA axis [80].
It is important to note that all but one study [37] included a relatively small number of participants, between 20 to 45 subjects, with a medium follow-up (<12 weeks), showing a high risk of bias. The studies on dietary counselling and food-based interventions included in this systematic review were conducted in cognitively healthy subjects and in patients with MCI, showing similar improvements in both groups, which cannot be extrapolated to adults with subjective cognitive decline.
Most of the RCTs included in this systematic review (48 out of 61) have evaluated the effect of dietary supplements (mostly plant-based supplements) on cognitive health using parallel study designs and different intervention durations (maximum 3 years). Half of the interventions evaluating dietary supplements (24 out of 48) were carried out for more than 12 weeks, 9 studies lasted less than this period and 16 studies were conducted for exactly 12 weeks.

Protein and Amino Acid Supplements
Met-Lys-Pro peptide, whey peptide and β-alanine significantly improved cognitive function [40,44,45]. Interestingly, the effects of whey peptide at week 12 were reduced compared to week 6 [44]. Some studies have suggested that this reduction is due to the learning effects on the tests, especially the less complex ones, resulting in smaller differences between the placebo and intervention groups. Still, the consumption of L-homoarginine and lactotripeptide did not exert any beneficial effect, which could be attributed to the short supplementation period of 4 and 8 weeks, respectively, and/or the small sample size (n = 20 and n = 35) [41,43]. Taurine is a sulfur compound with many functions in the nervous system. Recent data from animal studies suggest that taurine can improve cognitive function [94]. Considering that its plasma concentration decreases in older people, the lack of effect of taurine might be due to the small sample size (n = 24) or that a sufficient plasma taurine concentration was not reached to exert any beneficial effect [42]. Finally, anserine and carnosine only improved cognition in APOE4+ subjects, but not in the general population. Whether this over-stratification provokes false positive or there is an interaction between APOE4 and nutritional interventions deserves further attention [39].

Vitamin Supplements
The association of low vitamin D with global cognitive deficits is evident in several cross-sectional and longitudinal studies, but causality remains unclear [95]. The 3 RCTs included in this systematic review did not contribute to clarifying the role of vitamin D in cognitive decline. Thus, while vitamin D supplementation did not show beneficial effects in studies in cognitively healthy participants [46,48], it did in patients with MCI [47]. It is possible that the effects of vitamin D only become apparent when cognitive impairment is present. As for vitamin B12, it may exert a neuroprotective effect on cognitive function [49], alone or in combination with folic acid. The lack of effect was observed in one study included in our systematic review and this may be due to the use of diabetic medication in the study participants, including ACE inhibitor, angiotensin receptor blocker or metformin, which were also found to be neuroprotective [96]. Whether this over-stratification provokes false positive or there is an interaction between APOE4 and nutritional interventions deserves further attention [39].

Polyphenol Supplements
In general, polyphenol supplementation using extracts of fruits rich in anthocyanins such as blueberries, Montmorency tart cherries and grapes, as well as cocoa flavanols, Aronia melanocarpa, curcumin, spearmint, Lactobacillus plantarum C29-fermented soybean, Cosmos caudatus, chlorogenic acid and the nutraceutical Biokesum, have shown significant reduction in age-related decline in memory and cognition that could be explained by their antioxidant and anti-inflammatory properties [50][51][52][53][54][55][56][57][58][59]. However, a crossover trial published in 2016 used a single 60 mL dose of Montmorency tart cherry concentrate for 14 days and failed to show any beneficial effect on cognition, including attention, working memory, inhibition and cognitive flexibility in middle-aged participants, despite improvements in certain vascular functions [97]. These contrasting results could be partially explained by the lower doses used in this study compared to another study included in this review in which elderly participants consumed 2 cups of Montmorency tart cherry juice over a longer period (12 weeks) [54].

PUFAs Supplements
Despite the well-known role of omega-3 fatty acids in improving vascular endothelial function [98], the effect of omega-3 fatty acid supplementation on cognitive function has provided mixed results [99,100] and showed no improvement in the 5 RCT included in this review. Even if the lack of beneficial effect in the trial by Giudici et al. [61] might be due to decreasing adherence over time, similar results were obtained in the trial by Chhetri et al. [62] leading to the suggestion that a multidomain intervention might be more useful than PUFAs alone. However, in The Multidomain Alzheimer Preventive Trial (MAPT), no effects on brain glucose metabolism [64] or cognitive function were observed in subjects with SCD [63]. No beneficial effect was also observed in young men [60]. This is consistent with a previous RCT, in which an increase in oxyhemoglobin concentration was observed during a mental arithmetic task in participants assignedfiveto krill oil supplementation, but no changes in behavioral performance were reported [101].
However, contrary to these results, Cook et al. [102] reported that cognitively healthy young women with lower omega-3s had lower attention scores. Therefore, discrepancies between these studies might be due to gender differences in absorption and/or response to omega-3 fatty acid supplementation. Additionally, the poor retention rate of only 44.1% and the questionable compliance of the participants who completed the study make the results underpowered to test the alternative hypotheses.

Combination of PUFAs, Polyphenols and Vitamins
The synergistic action of antioxidants and anti-inflammatory bioactive compounds may also exert beneficial effects on some cognitive functions, such as executive [65], attention [65], working memory [65], intelligence [70], verbal memory and processing speed [68]. However, this synergistic action did not exert a beneficial effect in all studies. In this context, a combination of fish oil and blueberry extracts did not improve cognition [69]. One explanation is that long-term, daily supplementation with the combination treatments may lead to an over-regulation of the transcription factor Nrf2, whose excessive activation could contribute to adverse effects [103]. Additionally, for n-3 FA supplementation to exert an effect on cognitive outcomes, it is likely that adequate brain levels of n-3 FA must be reached. In the study by Arellanes et al. [71], vitamin B complex and n-3 FA were combined, but their levels could not be properly measured, so the lack of effects observed in this study may be due to inadequate brain delivery. Nevertheless, omega-3 FA may improve perfusion in cerebral regions typically affected in AD, which would help maintain brain structure and function and potentially delay conversion to dementia [72]. It is also possible that fish oil supports cognitive function in the short term, but that the positive effects may not apply over longer periods [66].

Other Types of Dietary Supplements
Other dietary supplements that have shown significant improvements on different cognitive domains are fungi and algae extracts (Hericiumerinaceus, Tremella fuciformis and Laminaria japonica), other plant extracts (Eleutherococcussenticosus leaf, Drynariafortuneirhizome, combined sage, rosemary and melissa, auraptene, astaxanthin and matured hop bitter acids), minerals such as magnesium, medium-chain triglycerides and the complex Twendee X (supplement containing a mix of eight antioxidants). The consumption of the above antioxidant and anti-inflammatory supplements not only significantly improved subjective memory complaints [84] and scores in different cognitive domains such as psychomotor and processing speed [73], working memory [76,77,79,81], short-term memory [74,76,77,84], spatial working memory [75,79,81,87], verbal episodic memory [78], verbal fluency [75,82,83], language [75,79,81,87], executive functions [79,82,84] and attention [79,83,87], but also cortical gray matter integrity in AD-related brain regions [84]. The spermidine-rich plant extract also had an impact on memory performance in subjects with SCD, although this was not significant, probably due to the small sample size (n = 30) and/or the short follow-up period (3 months) [85]. Differences in cognitive outcomes were also dependent on the patients' nutritional status. Well-nourished subjects showed a significant improvement in memory following multi-nutrient supplementation, the main component of which was omega-3 PUFA, but this improvement was not apparent in those with poorer nutritional status [86]. This might be explained by a combination of the potential synergies of the nutrients in the supplement and the good nutritional status [104]. Similarly, whereas antioxidant vitamins C and E were reported to have little effect on cognition [105,106], the combination of antioxidants in Twendee X showed stronger antioxidant and anti-inflammatory effects than single antioxidant vitamins in MCI subjects [87].

Potential Mechanisms
Oxidative stress and inflammation are common features in age-related cognitive decline [107]. In this vein, recent studies performed in cognitively normal elderly subjects have shown that levels of total antioxidant capacity are associated with decreased cortical integrity, poorer cognitive function, and increased cerebral amyloid beta burden [108,109], the earliest neuropathological event in AD. Therefore, antioxidants and anti-inflammatory agents are of particular interest in the modulation of reactive oxygen species (ROS) and inflammatory biomarkers [110] (Figure 6).
Plant-based food interventions, particularly those using colorful fruit and vegetablebearing plant foods, are rich in polyphenols, anthocyanins, carotenoids, flavonoids, folic acid, vitamins and fiber. On the other hand, omega-3 fatty acids contained in plants (α-linolenic acid (ALA)), but also in fish oil and supplements, such as DHA and EPA, constitute blood cell membranes and are essential for normal brain function [111,112]. These compounds may act through multiple pathways. They can reduce lipid peroxidation, protein and DNA oxidation, modulate the immune system and regulate synaptic activity and axonal density [113]. Several studies also suggest that they mediate microglia cells through the gut-brain axis, thus regulating the release of pro-inflammatory cytokines, which have a detrimental role in cognitive impairment [114,115]. DHA also stimulates neuronal plasticity through the Akt pathway and the neurotrophin BDNF, a signaling molecule that modulates synaptic plasticity and protects against neuronal apoptosis [116], but also influences energy metabolism, insulin sensitivity and glucose and lipid metabolism [117]. BDNF has been shown to increase in brain areas related to AD after the consumption of plant foods and supplements, as well as fermented foods [118]. For example, DHA supplementation has been found to elevate levels of hippocampal BDNF and enhance cognitive function in mice [119]. Besides facilitating synaptic plasticity and membrane fluidity, omega-3 fatty acids might also exert an effect on metabolism, stimulating glucose utilization [120], mitochondrial function [121] and reducing ROS [119]. Oxidative stress leads to vascular dysfunction and cell death. Adherence to the Mediterranean diet reduces vascular risk factors, such as LDL cholesterol and increases favorable HDL cholesterol [122], reduces systolic blood perfusion, exerts vasodilation and might promote angiogenesis. Moreover, components of this diet, such as legumes and whole-grain foods, may lower the glycemic index and improve cognitive function, which is of particular interest in the case of diabetic patients to cease the infiltration of inflammatory mediators into the brain [123].
Overall, the components of the Mediterranean diet may have a beneficial effect in ameliorating the risk factors for metabolic syndrome, thus contributing to better cognitive performance. In addition, high-fiber foods may have indirect effects on cognition through regulating intestinal activity [124] and facilitating nutrient absorption, for example that of GABA. Furthermore, the combination of various micronutrients in foods and supplements, such as antioxidants and flavonoids, might have an additive or synergistic effect on cognition. Of special interest are also the changes on brain integrity, as revealed by structural, function and metabolic changes. Brain image could be a surrogate marker of cognitive changes associated with nutrition, but few in this review have used them [27,28,37,41,50,64,65,72]. Since vascular dysfunction and synaptic loss are common features of cognitive impairment, increase in blood perfusion, glucose metabolism and neuronal activation might be good biomarkers of cognitive impairment.
Lastly, there are sex-specific differences in the effect of some nutrient components, such as DHA [125]. These differences could be related to sex hormones, as they influence PUFA metabolism differently.

Limitations and Strengths
There are several limitations that need to be taken into consideration when assessing the findings. Firstly, the dietary interventions were highly heterogeneous, with a wide age range of participants, various geographic regions, different durations of interventions of which many with short follow-up periods, and different measures of cognitive functioning, making it difficult to compare the effectiveness of the interventions across trials. Secondly, the characterization of subjects into cognitively healthy, SCD and cognitively impaired is not always well established in the trials, and even between articles there may be some varia-tion when classifying individuals. It is possible that a fraction of MCI and SCD individuals might by AD patients, but in almost all the studies they used these terms in an unspecific way, without biomarkers of AD to support it and/or measured using a MMSE/MoCa or similar test. Third, it is noteworthy that most of the studies dealt with supplement interventions, many of them of plant origin, as well as that more than half of the studies were conducted in cognitively healthy participants. Moreover, in this systematic review multimodal interventions were excluded because of the complexity of disentangling observed effects. However, several multimodal interventions (i.e., MAPT [126], FINGER [127]) have been relatively successful in combating complex chronic disorders. Overall, as reflected in the RoB and Figure 6, many studies judged at high risk of bias were conducted with a small sample size. Therefore, larger and high-quality studies are needed to avoid potential false positives.
The present systematic review also has strengths. Firstly, it presents the most comprehensive and updated assessment of this topic; secondly, it examines dietary patterns in people who are not diagnosed with dementia. This might be a useful strategy in relation to other nutrients, since it prevents dementia from a manageable approach easy to implement in dietary habits.

Further Study Directions
Future RCTs could test other types of dietary patterns, such as vegetarianism or pescovegetarianism, and foods such as dairy products or seafood/fish, among subjects with SCD and cognitive impairment. Additionally, it would be interesting to study gender responses to the interventions, and the potential modulatory role of gut microbiota composition to better establish the gut-brain-axis system [128].

Conclusions
Oxidative stress and inflammation may play an important role in cognitive function in healthy subjects, those subjects with cognitive impairment and self-perceived cognitive decline. Dietary counselling interventions, food-based interventions and dietary supplements with antioxidant and anti-inflammatory properties may help to ameliorate inflammation and oxidative damage.
The studies included in this review support that healthy dietary patterns, specific foods and dietary supplements can (a) improve memory, language, attention and concentration, executive functions, psychomotor speed and further cognitive domains, and (b) increase blood perfusion in areas typically related to AD.
Taken together, the existing evidence from RCTs over the last four years is in line with most of the previous findings already reported in the preceding systematic reviews. It confirms the beneficial role of the Mediterranean diet, plant foods and protein-rich supplements, some amino acids and minerals, polyphenols and the combination of them, as well as other types of dietary supplements. In addition, it supports the need for further research, especially in terms of vitamins and PUFAs, since different effects on cognitive function have been reported, but also other types of nutritional supplements which have been little studied. Moreover, most of the studies evaluated cognition, but it would also be necessary to determine brain changes, which could be useful to identify the mechanisms behind these benefits.
These data highlight the need for nutritional strategies that are most appropriate in maintaining cognitive function and protecting against age-related and AD cognitive impairment, thus reducing the risk, delaying the onset or slowing the progression of cognitive impairment and dementia in healthy, SCD and cognitively impaired subjects.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/nu13113728/s1, Figure S1: PubMed search strategy, Figure S2: Quality assessment of the RCTs that followed an intention-to-treat analysis, Figure S3: Quality assessment of the RCTs that followed a per-protocol analysis. Funding: This work was supported by the Spanish Ministry of Health (Institute of Health Carlos III) and the Spanish Ministry of Economy and Competitiveness-Fondo Europeo de Desarrollo Regional (Project PI19/00854, PID2020-119978RB-I00 and PID2020-118825GB-I00), the Regional Government of Andalusia (PY20_00858 8), the CSIC Fundation-Research Program for a Long-Life Society (0551_PSL_6_E) and by the CIBERNED (CB06/05/1111). L.G. received a pre-doctoral fellowship from the Generalitat de Catalunya's Department of Health (SLT01720000047) and M.R. from the Generalitat de Catalunya's Department of Universities (2021 FI_B 00657). C.P. is recipient of the Instituto de Salud Carlos III Miguel Servet fellowship (grant CP 19/00189).

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.

Data Availability Statement:
No new data were created or analyzed in this study. Data sharing is not applicable to this article.