This systematic review was designed to comprehensively investigate the protective role of flavonoids in amnesia. The review adhered strictly to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines to ensure transparency, accuracy, and reproducibility throughout the process.

A comprehensive literature search of published articles investigating the protective role of flavonoids in amnesia was conducted across three major electronic databases: Web of Science (WoS), Scopus, and PubMed. The search was performed on the 10 ^th of December 2024. Search terms included the keywords “flavonoid” and “amnesia,” with the search fields restricted to “Title,” “Abstract,” and “Keywords” in WoS and Scopus databases to ensure a robust and targeted retrieval of relevant studies. Boolean operators (AND, OR, NOT), quotation marks, parentheses, wildcards, and asterisk (*) symbols were strategically employed to optimize the search results, following established search optimization techniques. The search strategy combined multiple related flavonoid terms and was structured as follows for WoS and Scopus: (“flavonoid” OR “flavonol” OR “flavanone” OR “flavone” OR “flavan-3-ol” OR “isoflavone” OR “daidzein” OR “genistein” OR “kaempferol” OR “apigenin” OR “catechin” OR “epicatechin” OR “epigallocatechin” OR “gallocatechin” OR “luteolin” OR “hesperetin” OR “quercetin” OR “biochanin” OR “theaflavin” OR “formononetin” OR “baicalein” OR “myricetin” OR “chrysin” OR “naringenin” OR “glycitein” OR “eriodictyol” OR “isorhamnetin” OR “thearubigin” OR “anthocyanin” OR “delphinidin” OR “peonidin” OR “malvidin” OR “anthocyanidin” OR “petunidin” OR “cyanidin” OR “pelargonidin”) AND (“amnesia”). For PubMed, the search was adapted to include MeSH terms and field tags, resulting in the following syntax: (“flavonoid” [Title/Abstract] OR “flavonol” [Title/Abstract] OR …) AND (“amnesia” [Title/Abstract]) NOT (review [Publication Type]).

All retrieved articles from the database searches were downloaded in CSV format and subsequently uploaded onto the Rayyan platform, a web-based tool designed to streamline systematic review processes. Rayyan was used to facilitate the efficient and blinded screening of titles and abstracts based on predefined eligibility criteria. Duplicate entries were automatically identified and removed by the Rayyan system before the screening commenced. The selection process followed three sequential stages: title screening, abstract screening, and full-text review. Each article was assessed for relevance against the established inclusion and exclusion criteria and any disagreements during screening were resolved through discussion and consensus among the reviewers. Screening and selection were conducted in strict adherence to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines. Only studies that met all eligibility criteria at each screening level were included for final qualitative synthesis. Articles that failed to meet the criteria were excluded at the appropriate stage, with exclusion reasons documented systematically. Two independent reviewers screened and extracted data from each included article using a predefined data extraction form. Disagreements were resolved through discussion, and no automation tools were used in the data collection process.

Studies that investigated the role of flavonoids in amnesia, including those that evaluated any flavonoid compound for the treatment or management of amnesia, were included. Eligible studies specifically targeted flavonoid-based interventions and reported on the protective functions or therapeutic roles of flavonoids in amnesia. Only studies that presented quantitative outcomes related to the severity of amnesia, were published in peer-reviewed journals in English, and demonstrated clear methodology with detailed reporting regarding the administration of flavonoids or formulations, compounds, or extracts with flavonoid activity were considered for inclusion. Studies that did not focus on flavonoids, lacked experimental models of amnesia (with clinical trials also excluded), or investigated therapies unrelated to flavonoid interventions were excluded. Additionally, studies published in languages other than English, studies without clear reporting of outcomes or methodology, and secondary sources such as review articles, editorials, commentaries, and conference abstracts were excluded from the review.

Extracted data were tabulated and categorized under distinct headings for clarity. Descriptive synthesis was employed to summarize the outcomes, focusing on experimental models, types of flavonoids, behavioral assessments, biochemical markers, gene expression, and affected brain regions. Frequencies and percentages were calculated for key variables, and where possible, trends and correlations were discussed. The Rayyan platform facilitated structured data management throughout the analysis process.

A total of 593 articles were retrieved from three databases: 98 from Web of Science, 440 from Scopus, and 55 from PubMed. After the removal of 394 duplicates, 199 articles proceeded to the screening stage. Screening of titles and abstracts resulted in the exclusion of 139 articles. The full texts of the remaining 60 articles were then assessed for eligibility, leading to the exclusion of 35 articles that did not meet the inclusion criteria. In the end, 25 articles were included in the final review after completing the eligibility and quality assessment process as depicted in Figure 1.

These studies were conducted across various countries, demonstrating considerable diversity in animal models, species, sample sizes, sexes, and age groups used in the research. The studies spanned multiple countries, including Brazil, Pakistan, Cameroon, India, China, Saudi Arabia, South Korea, Tunisia, Romania, Egypt, the United Arab Emirates, and Mongolia, along with contributions from unspecified locations. A variety of animal models were employed, with rats and mice being the most commonly used. Zebrafish and gerbils were also utilized in several studies. Among the mouse models, specific strains included Swiss albino mice, Laca mice, and transgenic APP/PS1 mice. For rat models, Wistar and Sprague-Dawley strains were frequently used. The majority of studies predominantly used male animals, though some included both sexes. Sample sizes varied widely, ranging from as few as six animals per group to larger cohorts of over 100 animals. Animal ages were also diverse, with studies reporting precise age ranges or weights. For instance, some studies focused on young adult animals, while others used aged animals to model disease progression. Several studies also detailed treatment initiation and assessment at specific age milestones. Full details of species, strains, sex distribution, sample sizes, and ages are summarized in Table 1.

The analysis of behavioral models employed across the reviewed studies demonstrates a clear preference for specific memory and learning assessment paradigms as depicted by Figure 2. The Morris Water Maze (MWM) was the most frequently utilized test, reflecting its robust sensitivity for evaluating spatial learning and memory functions, particularly hippocampus-dependent processes. ²⁵ The Novel Object Recognition Test (NORT) and Y-Maze Test (YMT) also featured prominently, emphasizing the importance of assessing recognition memory and spontaneous alternation behavior as measures of cognitive flexibility. ³⁰ Other commonly used models included the Elevated Plus Maze (EPM) and Passive Avoidance Test (PAT), suggesting a broader interest in anxiety-related memory retention and avoidance learning. Less frequently employed tests, such as the Radial Arm Maze and Object-Location Memory Test, provide complementary assessments of working memory and object-place associations. Overall, the distribution of behavioral models indicates that researchers prioritized multi-dimensional evaluation of memory, integrating spatial, recognition, and emotional memory paradigms to capture the complexity of amnesia-related cognitive deficits.

Oxidative stress markers assessed

The evaluation of oxidative stress markers across the included studies highlights the critical role of oxidative damage in amnesia and the potential antioxidative benefits of flavonoid interventions as depicted by Figure 3. Malondialdehyde (MDA) was the most frequently measured marker, reflecting its well-established role as a sensitive indicator of lipid peroxidation and cellular oxidative injury. ^{31, 32} Glutathione (GSH) and superoxide dismutase (SOD) were also commonly assessed, underscoring the importance of endogenous antioxidant defenses in maintaining neuronal integrity. ³³ Other markers such as lipid peroxidase (LPO), catalase (CAT), and nitric oxide (NO) were frequently utilized, further emphasizing the multifactorial nature of oxidative damage pathways in cognitive dysfunction. ³⁴ Less frequently assessed indicators, including FRAP, DPPH, and reactive oxygen species (ROS), provided complementary measures of overall antioxidant capacity and free radical scavenging activity. ³⁵ Overall, the pattern of biomarker evaluation supports that oxidative stress modulation is a central mechanism through which flavonoids exert neuroprotective effects against amnesia.

Plant extracts, phytochemicals, dosages, and routes of administration

The reviewed studies demonstrate the diverse use of plant extracts and isolated flavonoid compounds in the experimental management of amnesia, as summarised in Table 3. Oral administration emerged as the predominant route, reflecting its practicality and relevance for translational research into human therapies. Most studies utilized ethanolic or methanolic extracts (e.g., Averrhoa carambola, Grewia asiatica, Cyamopsis tetragonoloba), highlighting the solubility of active flavonoids in polar solvents and the richness of polyphenolic compounds in these preparations. ^{10, 9, 12} Doses ranged widely from low concentrations (e.g., 5 mg/kg for Biochanin A) to relatively high doses (e.g., 800 mg/kg for Lavandula stoechas extract), suggesting differences in extract potency, bioavailability, and experimental design across studies. ¹⁸ Standardized compounds such as quercetin, kaempferol, silibinin, and pycnogenol® were also assessed, providing insights into the neuroprotective potential of purified flavonoids. ^{7, 17, 19} Less common administration methods, such as intraperitoneal injection ( Pinus pinaster, Glycine max) and immersion ( Tribulus terrestris), highlight methodological adaptations for certain experimental models. Notably, several studies did not report dosage or administration details, which may affect reproducibility and interpretation. ¹⁷ Overall, the data suggest that both crude extracts and purified flavonoid compounds confer neuroprotective effects against amnesia through varied dosing strategies and predominantly oral delivery, supporting their therapeutic potential.

Animal models utilized

The distribution of animal models in the reviewed studies highlights a predominant reliance on rats (40.74%) and Swiss albino mice (37.04%) for evaluating the protective effects of flavonoids against amnesia, as depicted in Figure 4. This preference reflects the extensive validation of these models in behavioral neuroscience, particularly in memory and learning paradigms. Among rat models, Wistar rats (14.81%) and Sprague Dawley rats (7.41%) were commonly employed due to their well-characterized cognitive profiles and reproducibility in experimental settings. ^{29, 30} The use of Danio rerio (zebrafish) (7.41%) illustrates emerging trends favoring alternative, cost-effective models for high-throughput behavioral screening. ^{21, 22} Less frequently used models, including Mongolian gerbils, APP/PS1 transgenic mice, ICR mice, Laca mice, and aged wild-type mice (each at 3.70%), reflect specific study designs targeting particular aspects of neurodegeneration, aging, or genetic predisposition to memory deficits. Overall, the data indicate a strong preference for traditional rodent models, with growing exploration of alternative species to diversify cognitive research in amnesia.

This systematic review is limited by its exclusive focus on preclinical experimental studies conducted in animal models, which restricts the generalizability of the findings to human populations. The lack of clinical trials highlights a translational gap that must be addressed to confirm these results in clinical settings. There was considerable heterogeneity across the included studies in terms of animal species, flavonoid types, dosages, routes of administration, and behavioral assessment tools. This variability hindered direct comparisons across studies and made meta-analysis infeasible. Many studies did not provide detailed phytochemical characterization or standardization of the flavonoid content in their extracts. This absence of consistency affects the reproducibility and pharmacological interpretation of the findings. In several cases, the duration of treatment and follow-up periods was either unclear or not reported, limiting conclusions about the sustainability of effects over time. Risk of bias assessments and certainty grading of evidence, such as the use of GRADE, were not conducted, which weakens the ability to critically appraise the overall strength of the evidence. The review was also limited to studies published in English, which may have introduced language bias and excluded potentially relevant findings from non-English sources. These limitations emphasize the importance of methodological standardization, clinical translation, and broader inclusion criteria in future systematic reviews on this topic.

Future research should prioritize the standardization of flavonoid preparations and dosing protocols to enhance comparability. Rigorous pharmacokinetic and pharmacodynamic studies are necessary to optimize bioavailability and therapeutic windows. To enhance clinical relevance, translational studies, including well-designed human trials, should be conducted to validate the neuroprotective efficacy of flavonoids in amnesia and related cognitive disorders. Combining flavonoids with current pharmacotherapies should also be explored for potential synergistic effects. Additionally, the development of flavonoid-based nutraceuticals or supplements with well-defined compositions and safety profiles may offer accessible interventions, especially in aging or at-risk populations. Finally, future reviews should consider broader inclusion criteria, such as non-English studies and grey literature, to capture a more holistic evidence base.