A century of coffee and tea research in cognitive health and Alzheimer's disease: Structural,thematic,and translational insights (1911

Abstract

Background

Although studies have explored tea and coffee in relation to Alzheimer's disease, no century-scale analysis has jointly examined both within a unified primary-evidence framework.

Objective

This study maps the structural, thematic, and temporal evolution of coffee and tea research in cognitive aging.

Methods

Scopus-indexed original English articles (1911–2025) were retrieved using a structured Boolean strategy under PRISMA guidance. Analyses were conducted using Bibliometrix and VOSviewer. Performance metrics, collaboration networks, co-citation mapping, Bradford's Law, co-word clustering, thematic evolution, and overlay visualization were applied to full-period and recent (2021–2025) datasets.

Results

A total of 2873 articles across 1285 sources demonstrated steady annual growth (4.96%) and substantial citation impact (mean 40.17 citations per document). Research productivity was concentrated in high-income countries, led by the United States, United Kingdom, and China, reflecting core–periphery stratification and citation asymmetry. Collaboration networks showed hub dominance with dense transatlantic–Eurasian linkages. Bradford analysis identified a limited core of highly productive journals. Thematic evolution revealed persistent anchoring in Alzheimer's disease, oxidative stress, and neuroprotection, with sustained prominence of coffee, caffeine, tea, and polyphenols. Recent years indicate translational expansion integrating microbiome science and computational methods. Overlay visualization demonstrated temporal stratification, highlighting emerging themes such as gut microbiota and deep learning alongside stable beverage-related cognitive frameworks.

Conclusions

Coffee and tea research in cognitive aging has evolved into a mature, mechanistically grounded, and globally stratified field, increasingly integrating translational, microbiome, and computational approaches in dementia-related investigations.

Keywords

Alzheimer's disease caffeine coffee cognitive function dementia neuroprotection oxidative stress tea

Introduction

Coffee and tea intake has been robustly investigated for its association with dementia risk and cognitive function, with multiple large-scale cohort studies and meta-analyses reporting inverse relationships between caffeinated beverage consumption and cognitive decline.^1–3 In a recent prospective analysis of 131,821 participants, higher intake of caffeinated coffee was significantly associated with a lower risk of dementia, with those in the highest consumption quartile exhibiting approximately an 18% reduced risk compared with low consumers (hazard ratio ∼0.82) and a modest reduction in subjective cognitive decline, while caffeinated tea showed similar protective associations; decaffeinated coffee did not confer these benefits.^1,4,5 Meta-analytic evidence also supports non-linear inverse associations between coffee intake and dementia risk, suggesting that moderate consumption (around 2–3 cups per day) is linked with the greatest reduction in incident cognitive disorders and slower cognitive decline, with diminishing returns observed at very high consumption levels.⁶ A systematic review of cohort studies documented that increased tea consumption was associated with a decreased risk of dementia and Alzheimer's disease, with a linear inverse relationship between tea intake and dementia risk, and that each additional cup of tea per day was associated with lower risk of cognitive disorders.⁶ Furthermore, meta-analyses have identified statistically significant reductions in risk of cognitive disorders for moderate daily coffee and tea intake, reinforcing public health recommendations for dementia prevention through diet and lifestyle.^3,6

Mechanistic and observational studies help contextualize these epidemiological associations, suggesting that bioactive compounds in coffee and tea—including caffeine, polyphenols, catechins, and flavonoids—may exert neuroprotective effects by reducing neuroinflammation, oxidative stress, and amyloid pathology, all implicated in the pathogenesis of Alzheimer's disease and dementia.^7,8 Evidence from observational cohorts indicates that habitual coffee consumption may be associated with slower cognitive decline and a reduced likelihood of transition from normal cognition to mild cognitive impairment or dementia, potentially mediated by caffeine's antagonism of adenosine receptors and improvement of vascular and metabolic health.^1,4,6,8,9 Similarly, tea intake—especially green tea—has been linked to lower risk of cognitive disorders in dose-response analyses, with higher intake associated with progressively reduced risk, although some heterogeneity exists by tea type and study design.^1–3^,6,10 While not all studies are concordant and some report null associations, the preponderance of evidence supports a protective association between moderate coffee and tea consumption and reduced dementia risk and better cognitive outcomes, recognizing that causality cannot be definitively established from current observational data and that confounding factors such as lifestyle, diet, and genetic predisposition may influence observed effects.¹

Although bibliometric investigations have explored coffee- and tea-related research in Alzheimer's disease, important conceptual and methodological gaps persist. Prior studies^11,12 have either focused broadly on dietary strategies in AD without isolating stimulant-based beverages (coffee and tea) as independent analytical domains, or have examined tea within a restricted temporal window (e.g., 2014–2023) without integrating coffee and caffeine research into a unified framework. Moreover, existing analyses have primarily relied on the Web of Science Core Collection and have frequently included multiple document types, potentially inflating citation-driven visibility of reviews rather than capturing the authentic trajectory of primary evidence generation. To date, no study has conducted a century-scale bibliometric mapping (1911–2025) of original research articles exclusively indexed in Scopus, integrating coffee, caffeine, and tea within a single comparative scientometric architecture (Figure 1). The present study addresses these gaps by restricting inclusion to original English-language articles, applying a dual-dataset strategy to assess both long-term intellectual evolution and recent research frontiers, and employing a multi-software analytical framework—VOSviewer and Bibliometrix—to provide a comprehensive structural, thematic, and temporal mapping of stimulant-related dementia research. This integrated approach offers the first century-spanning, primary-evidence-focused scientometric synthesis of coffee and tea research within the cognitive aging domain.

Methods

Justification for the use of Scopus

Scopus (Elsevier) was selected as the primary data source due to its extensive multidisciplinary coverage, robust citation indexing, and compatibility with advanced bibliometric tools such as Bibliometrix (R), VOSviewer, and CiteSpace. Scopus indexes over 27,000 active peer-reviewed sources and provides standardized metadata, including author affiliations, funding data, citation counts, abstracts, and keywords, thereby enabling reliable performance analysis, co-citation mapping, thematic evolution, and collaboration network assessment. Compared with single-discipline databases, Scopus offers broader representation of neuroscience, nutrition, epidemiology, pharmacology, and translational research—domains central to coffee, tea, and cognitive aging research. Its refined filtering system, including document type, language, and publication year, further enhances methodological transparency, dataset consistency, and reproducibility.

Search strategy

This bibliometric study was conducted and reported in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) framework to ensure transparency, reproducibility, and structured documentation of the identification, screening, eligibility, and inclusion processes. The Scopus database (Elsevier) was selected as the sole data source due to its extensive multidisciplinary coverage, standardized citation indexing, and compatibility with bibliometric software tools such as Bibliometrix, VOSviewer, and CiteSpace. The search was performed on 10 January 2026 and included all records indexed up to 31 December 2025.

A comprehensive Boolean search strategy (Figure 1) was developed using truncation () and phrase searching to capture variations of beverage-related and cognitive-related terminology. The initial search was conducted in the TITLE-ABS-KEY fields using the following structure: (coffee OR caffeine OR “Coffea” OR tea OR “Camellia sinensis” OR “green tea” OR “black tea”) AND (dementia OR “Alzheimer*” OR “cognitive function” OR “cognitive decline” OR “cognitive impairment” OR “neuroprotective” OR memory). This broad query yielded 6569 records. The search strategy was intentionally designed to capture exposure-based research examining the relationship between coffee and tea consumption and cognitive outcomes, including Alzheimer's disease. Therefore, highly specific pathological terms such as “amyloid-beta” and “neurodegeneration” were not included in the Boolean query. Inclusion of these terms would bias retrieval toward mechanistic and molecular studies and potentially exclude a substantial body of epidemiological, clinical, and cognitive research central to the study objectives. The term “neurodegenerative outcomes” is used descriptively to reflect the broader clinical context rather than as a primary search construct.

To enhance precision and ensure dataset homogeneity, filters were applied to include only original research articles (DOCTYPE = “ar”) published in English, reducing the dataset to 3919 records. To further improve specificity and minimize incidental keyword co-occurrence, a refined search was implemented targeting Title and Abstract fields separately while maintaining Boolean consistency. Publications from 2026 were excluded to avoid incomplete citation windows and indexing bias. After applying these refinements, 2873 articles met the eligibility criteria and constituted the final dataset (Dataset 1). In addition, a temporal subset comprising publications from 2021 to 2025 was extracted to assess recent research dynamics and thematic evolution, yielding 1015 articles (Dataset 2). This five-year window was selected to represent the most recent phase of research development and is commonly used in bibliometric analyses to identify emerging trends, thematic shifts, and evolving research frontiers, enabling comparison between long-term knowledge structures and contemporary research dynamics.

Studies were included if they were Scopus-indexed original research articles written in English and explicitly examined coffee, caffeine, tea, green tea, black tea, or Camellia sinensis in relation to dementia, Alzheimer's disease, cognitive function, cognitive decline, memory, or neuroprotective outcomes. Records were excluded if they were reviews, conference papers, editorials, letters, book chapters, non-English publications, studies published in 2026, or articles lacking substantive relevance to both beverage exposure and cognitive or neurodegenerative outcomes.

Restricting the analysis to original research articles was methodologically justified to ensure that the bibliometric mapping reflects the authentic trajectory of primary evidence generation rather than secondary synthesis. Reviews, editorials, and conference papers often accumulate disproportionately high citations due to their summarizing function, which can artificially inflate visibility and distort thematic centrality in co-citation and co-word networks. By focusing exclusively on original articles, the study captures empirical innovation, methodological progression, and the evolution of experimental and epidemiological inquiry in coffee and tea research related to cognitive aging. This approach enhances analytical precision, reduces citation bias, and allows clearer identification of genuine research frontiers, intellectual foundations, and translational developments shaping the field over time

Analyses

The analyses integrated performance, relational, and longitudinal science-mapping techniques. Descriptive performance indicators (annual growth rate, citation per document, authorship patterns, SCP/MCP ratios, and institutional productivity) were computed using Bibliometrix (version 4.2.3, R package; Italy) within the R statistical environment (R version 4.3.2; Italy).¹³ Journal dispersion was evaluated using Bradford's Law, and leading scholars were assessed through Scopus-derived H-index metrics. Network-based analyses—including country collaboration, international co-authorship, source co-citation, and keyword co-occurrence—were conducted using VOSviewer (version 1.6.19; Leiden University, Netherlands), applying total link strength (TLS) and modularity-based clustering.¹⁴ Thematic structure and evolution were examined via co-word analysis using the Louvain community detection algorithm¹³ and positioned on Callon's centrality–density strategic diagrams. Longitudinal thematic flows were visualized through Sankey mapping, and emerging topics were identified using overlay visualization based on average publication year. Collectively, these approaches enabled integrated performance analysis, structural mapping, and temporal stratification of the knowledge domain.

Results

Data overview

The dataset covers the period 1911–2025 and includes 2873 documents published in 1285 sources, demonstrating a steady annual growth rate of 4.96%. The average document age is 10.6 years, with a mean citation rate of 40.17 citations per document, indicating strong scholarly impact. The corpus contains 15,557 Keywords Plus and 6367 author keywords, reflecting substantial thematic diversity. A total of 11,713 authors contributed to the dataset, with only 145 single-authored documents, highlighting a predominantly collaborative research structure. The average number of co-authors per document is 5.98, and 21.98% of publications involve international collaboration, suggesting moderate global research integration within the field.

Annual publication output and citation dynamics

Annual publication output and citation dynamics across the observation period (1911–2025) are presented in Figure 2. The corpus comprised 2873 publications, with output exhibiting a pronounced exponential growth trajectory, culminating in peak annual productivity of 249 articles in 2025. Mean total citations per article reached a maximum of 136.48 in 2006, thereafter declining progressively—a pattern attributable to the well-established bibliometric recency bias, whereby recently published articles have had insufficient time to accumulate citations. Mean total citations per year similarly peaked in 2006 (6.50), corroborating the citation concentration within mid-period literature. The inverse relationship between publication volume and mean citation intensity underscores the temporal lag inherent to scholarly citation accumulation within this research domain.

Figure 1.

Scopus database search strategy and sequential dataset construction for the bibliometric analysis of caffeine, tea, and cognitive function/dementia literature (1911–2025). The flowchart delineates four successive filtering stages applied to an initial retrieval of 6569 records, encompassing document type restriction (articles only), language limitation (English), field-specific term refinement (title and abstract fields), and exclusion of incomplete 2026 publication records, yielding a full analytical corpus of 2873 documents (Dataset 1) and a recent quinquennium subset of 1015 documents (Dataset 2; 2021–2025).

Figure 2.

Annual publication output and citation metrics in coffee and tea and cognitive research literature (1911–2025). Vertical bars represent annual publication frequency (left axis); solid line denotes mean total citations per article (right axis); dashed line indicates mean total citations per year (offset right axis). Shading reflects relative citation intensity across the observation period.

Global scientific production and collaboration structure

Global productivity exhibits pronounced research concentration, with output clustered in a limited set of high-income countries, reflecting a classic core–periphery scientometric structure. Citation distribution demonstrates marked impact asymmetry, as leading producers (e.g., United States, United Kingdom, China) accumulate disproportionately higher citation shares relative to peripheral nations. However, citation impact does not perfectly mirror productivity; several mid-volume countries display elevated citation intensity, indicating stratified influence rather than linear output–impact correspondence. This pattern underscores geopolitical research dominance, where established funding ecosystems, advanced infrastructure, and strong international collaboration networks amplify visibility and citation accrual. Persistent North–South disparities suggest structural inequalities in research capacity, access to high-impact journals, and participation in global consortia. Overall, productivity (Figure 3A) and citation (Figure 3D) hierarchies align with national investment in biomedical research, historical regulatory engagement, and differential disease burden, reinforcing the central role of resource-rich systems in shaping global knowledge production and agenda setting.

Figure 3.

Global scientific production and collaboration structure. (A) World map of country scientific production based on total publications. (B) Overlay visualization of country productivity by average publication year, indicating temporal dynamics. (C) Corresponding author's countries showing single-country (SCP) and multiple-country publications (MCP). (D) Density visualization of country co-authorship intensity. (E) Global bilateral collaboration map illustrating inter-country linkages. (F) International co-authorship network displaying clusters, node size proportional to productivity, and link thickness representing collaboration strength (https://tinyurl.com/27s5vx93).

The country production overlay indicates a historically dominant core led by the United States, United Kingdom, Germany, China, and Japan, reflecting sustained output over earlier periods. Blue and green nodes denote established contributors with continuous productivity (Figure 3B). In contrast, yellow nodes represent recent entrants or rapidly expanding participants, including countries such as the Philippines, Ukraine, Iraq, and several African and Middle Eastern nations. This pattern suggests geographic diversification, with emerging economies increasingly integrating into the global research system and contributing to contemporary expansion.

Figure 3C reveals a pronounced SCP dominance across leading countries, consistent with a core–periphery configuration in which core nations sustain high levels of domestic production. In mature research economies such as China and the United States, elevated SCP volumes reflect strong internal scientific capacity and dense national funding ecosystems, indicating structural scientific independence. However, high SCP ratios may also signal partial insulation from cross-border integration, particularly when MCP proportions remain comparatively lower. In contrast, emerging systems often display moderate SCP levels but rely more heavily on international collaboration to enhance visibility and credibility. Regulatory-driven national agendas—especially in nutrition, aging, and neurodegeneration—can further reinforce domestic publication concentration. SCP dominance in core countries centralizes knowledge generation, potentially constraining diffusion pathways, whereas balanced SCP–MCP profiles facilitate broader global integration and transnational knowledge circulation within the field.

The bilateral collaboration network demonstrates a highly centralized architecture dominated by the United States, which anchors the strongest dyads, particularly USA–China (59), USA–United Kingdom (36), and USA–Canada (34). These high-frequency links indicate a hub-and-spoke configuration in which the U.S. functions as the principal global connector (Figure 3E). China and the United Kingdom act as secondary hubs, forming dense triadic linkages with other high-capacity systems such as Germany, Australia, and Japan. European collaborations (e.g., France–Germany, Spain–Italy, Germany–Netherlands) reveal strong intra-regional cohesion, suggesting a sub-core cluster. Emerging South–South collaborations are present but comparatively weaker and fragmented. Overall, the distribution of collaboration frequencies reflects asymmetric integration, with knowledge exchange concentrated among established research economies while peripheral linkages remain sparse and low-intensity.

The VOSviewer international co-authorship network (60 countries; 483 links; total link strength [TLS] = 1462; 8 clusters) exhibits a clear core–periphery structure characterized by high centralization and hub dominance (Figure 3F). The United States occupies the primary hub position, demonstrating the highest degree centrality and TLS (431), indicating both extensive collaboration breadth and strong tie intensity (Figure 3F). The United Kingdom (TLS = 209) and China (TLS = 156) function as secondary hubs, forming a dense transatlantic–Eurasian collaboration triangle. This configuration reflects a highly centralized network topology, where a limited number of dominant actors mediate global knowledge flows.

Cluster density is moderate, as evidenced by 483 links among 60 nodes, suggesting substantial but uneven connectivity. European countries (Germany, France, Switzerland, Italy) form a tightly interconnected meso-core, reinforcing regional clustering within the global core. Peripheral nodes—primarily emerging economies—display lower degree centrality and weaker TLS, indicating dependency on core collaborations rather than autonomous network integration.

The skewed TLS distribution implies knowledge concentration within scientifically advanced nations, reflecting disparities in research infrastructure, funding capacity, and publication visibility. Hub dominance signals geopolitical research leadership concentrated in North America and Western Europe, with China emerging as a major bridging power. Overall, the collaboration structure indicates a mature yet hierarchically organized scientific system, where transdisciplinary integration is facilitated by core hubs, while peripheral participation remains structurally constrained by asymmetrical connectivity.

Most productive affiliations

The most productive affiliations in this research domain are led by major academic and governmental institutions with strong biomedical and neuroscience infrastructures. The Ministry of Education of the People's Republic of China and Harvard Medical School each contributed 45 publications, followed closely by Inserm (44). Prominent contributors also include Universidade Federal do Rio Grande do Sul (33), University of South Florida (30), and Chinese Academy of Sciences (29). Leading clinical and translational centers such as Massachusetts General Hospital (25) and Brigham and Women's Hospital (24) demonstrate strong involvement, while global representation is evident through institutions including University of Oxford (26), King Saud University (19), Karolinska Institutet (17), and National University of Singapore (16). Overall, productivity is concentrated within internationally recognized research-intensive universities and medical centers spanning North America, Europe, Asia, and Latin America.

Bradford's Law analysis of journal productivity

Bradford's Law states that scholarly output is distributed across journals in concentric zones, with roughly one-third of articles concentrated in a small core (Zone 1), one-third in moderately productive journals (Zone 2), and one-third dispersed across numerous peripheral titles (Zone 3). In this dataset (2873 articles; 1285 sources), Zone 1 comprised 21 journals (1.6%) producing 33.4% of articles; Zone 2 included 95 journals (7.4%) accounting for 66.7% cumulatively; and Zone 3 contained 1169 journals (91.0%) generating the remaining 33.3% (Figure 4). Core sources included Nutrients, Journal of Alzheimer's Disease, PLoS One, Psychopharmacology, and Scientific Reports. The Bradford multiplier (k ≈ 4.5) indicates geometric dispersion consistent with theoretical expectations.

Figure 4.

Core sources identified by Bradford's Law. Bradford distribution curve illustrating journal productivity ranked by descending article output. The shaded region represents Zone 1 (core journals), which accounts for approximately one-third of total publications. The x-axis displays source log(rank), and the y-axis indicates article frequency per journal, demonstrating geometric dispersion across successive zones.

Figure 4 also presents temporal trajectories for 20 consolidated author keywords (1985–2025). Caffeine (n = 6346) and Alzheimer's disease (n = 2578) showed the highest frequencies, with all keywords peaking in 2024–2025, demonstrating sustained and accelerating thematic growth.

Leading scholars

Table 1 presents the first 15 leading scholars in the field, ranked by publication output (N), institutional affiliation, and bibliometric impact indicators. Cunha, R.A. leads in productivity (25 publications), while Youdim, M.B.H. exhibits the highest H-index (114), reflecting substantial citation impact. Several authors demonstrate strong influence, including Schwarzschild, MA (H-index 88), Müller, C.E. (89), and Chen, J.F. (83). The cohort represents geographically diverse institutions across Europe, North America, Asia, and South America, indicating broad international engagement. Brazilian and Taiwanese institutions show notable representation. Overall, Table 1 highlights a concentration of high-impact scholars driving research productivity and citation performance within the domain.

Table 1.

Leading scholars.

Author	Number of articles	Affiliation	H-index	Scopus ID:
Cunha, R.A.	25	University of Coimbra, Center for Neuroscience and Cell Biology, Coimbra, Portugal	84	7102879525
Schwarzschild, MA	15	Massachusetts General Hospital, Boston, United States	88	7003651412
Chen, J.F.	14	Wenzhou Medical University, Wenzhou, China	83	56008825700
Porciúncula, L.O.	14	Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil	35	6603042120
Souza, D.O.	14	Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil	73	7102611542
Tan, J.	13	Hangzhou City University, Hangzhou, China	68	7402302301
Youdim, M.B.H.	13	Technion—Israel Institute of Technology, Haifa, Israel	114	36051309100
Soung, H.S.	11	National Defense Medical University Taiwan, Taipei, Taiwan	14	55148303700
Wang, M.H.	11	En Chu Kon Hospital, New Taipei City, Taiwan	15	7406686610
Zeman, A.	11	Edinburgh Medical School, Edinburgh, United Kingdom	42	7005604768
Mello-Carpes, P.B.	10	Universidade Federal do Pampa, Bage, Brazil	22	55552024400
Müller, C.E.	10	Universität Bonn, Bonn, Germany	89	7404109739
Tseng, H.C.	10	Shin Kong Wu Ho-Su Memorial Hospital, Taipei, Taiwan	13	56522356400
Unno, K.	10	University of Shizuoka, Shizuoka, Japan	30	7004931096
Arendash, G.W.	9	RF Longevity, Phoenix, United States	59	7006472652

Temporal trajectory panels

Figure 5 presents individual temporal trajectory panels for 20 consolidated author keywords identified across the caffeine and cognitive research corpus (1985–2025). Each panel depicts annual publication frequency as an area-filled curve, enabling visual assessment of growth trajectories, inflection points, and acceleration patterns. Caffeine (n = 6346) and Alzheimer's Disease (n = 2578) exhibited the highest cumulative frequencies, reflecting sustained and escalating scholarly attention across the observation period. Annotation insets within each panel document three empirical metrics: year of first appearance, peak frequency year and corresponding value, and the percentage of cumulative occurrences concentrated within the most recent quinquennium (2021–2025). The navy-shaded band consistently demarcates this interval, facilitating cross-panel comparison of contemporary research intensity. All 20 keywords demonstrate monotonically increasing trajectories, with peak values uniformly recorded in 2024–2025, indicating unabated growth across thematic domains and confirming the progressive acceleration of scholarly output throughout the observation period.

Figure 5.

Temporal trajectories of the top 20 consolidated author keywords in caffeine and cognitive research literature (1985–2025). Area-filled curves represent annual publication frequency per keyword; annotated insets denote year of first appearance, peak frequency year and value, and the proportion of cumulative occurrences recorded within the final quinquennium (2021–2025). The shaded band demarcates the 2021–2025 interval.

Thematic evolution across four periods (1911–2025)

The longitudinal thematic mapping (Figure 6) identified distinct thematic configurations across four consecutive periods (1911–2011, 2012–2018, 2019–2022, and 2023–2025). Across the full timespan, recurrent high-frequency themes included Alzheimer's disease, oxidative stress, neuroprotection, caffeine, polyphenols, aging, and cognitive function. During the period 1911–2011, dominant themes included oxidative stress, neuroprotection, aging, and Alzheimer's disease. Additional recurring topics were caffeine, adenosine, adenosine A2A receptor, cognition, and transient epileptic amnesia. Thematic connections indicate linkages between oxidative stress and aging, caffeine and cognition, and Alzheimer's disease with mechanistic receptor-based investigations.

Figure 6.

Thematic evolution of research on coffee, tea, and cognitive health (1911–2025). Sankey diagram illustrating the longitudinal development and interconnection of dominant research themes across four periods (1911–2011, 2012–2018, 2019–2022, and 2023–2025). Node size reflects thematic frequency, while link thickness represents the strength of conceptual continuity between periods. Early research emphasized oxidative stress, neuroprotection, aging, and adenosine-related mechanisms. Subsequent phases show consolidation around Alzheimer's disease, caffeine, and polyphenols, followed by an intensified focus on cognition and memory. The most recent period highlights the emergence of gut microbiota and deep learning, indicating interdisciplinary expansion toward systems biology and computational approaches in cognitive and neurodegenerative research.

In the 2012–2018 period, Alzheimer's disease and oxidative stress remained central themes. Polyphenols, caffeine, green tea, and cognition increased in prominence. Thematic flows show continued linkage of oxidative stress with Alzheimer's disease and aging, alongside emerging connections between polyphenols and neuroprotection. Coffee-related terminology became more explicitly connected to dementia and cognitive-related descriptors. Between 2019 and 2022, caffeine, cognition, and memory appeared as highly developed and centrally positioned themes. Alzheimer's disease, oxidative stress, polyphenols, green tea, neuroprotection, and aging remained present. Additional keywords included antioxidant, apoptosis, working memory, and neurodegeneration. Thematic continuity is observed between caffeine and cognition, polyphenols and oxidative stress, and Alzheimer's disease across consecutive periods.

In the most recent period (2023–2025), Alzheimer's disease, oxidative stress, caffeine, polyphenols, aging, and cognitive function persisted as high-frequency themes. Newly appearing themes included gut microbiota, deep learning, machine learning, sleep, diet, exercise, metabolomics, and neuroinflammation. Thematic links show continuity from earlier periods for caffeine, oxidative stress, and Alzheimer's disease, alongside new associations connecting cognitive function with computational and microbiome-related terms. Across all four periods, Alzheimer's disease and oxidative stress maintained consistent presence. Caffeine appeared continuously and demonstrated increasing central positioning in later periods. Polyphenols and green tea remained recurrent themes with shifting co-occurrence patterns. The Sankey visualization demonstrates sustained thematic continuity with the addition of emerging interdisciplinary keywords in the final period.

Thematic map

Thematic map: whole period (1911–2025)

The thematic structure for the entire study period was generated using Bibliometrix via co-word analysis of author keywords, applying Louvain clustering and Callon's centrality and density metrics (Figure 7A). The motor-theme quadrant is dominated by the cluster “Alzheimer's disease—oxidative stress—neuroprotection,” demonstrating both high internal development and strong integration within the broader network. The “caffeine—memory—cognition” cluster appears as a basic theme, indicating high centrality but comparatively lower density, thus functioning as a transversal and foundational domain. The “coffee—dementia—cognitive function” cluster occupies a transitional position near the central axis, reflecting ongoing consolidation. Niche themes include “transient epileptic amnesia—epilepsy” and “deep learning,” which show high internal cohesion but limited systemic connectivity. Overall, the long-term structure reflects a field anchored in neurodegenerative mechanisms with stable cognitive-pharmacological foundations.

Figure 7.

Strategic thematic maps of research clusters based on Callon's centrality (relevance) and density (development). (A) Whole-period thematic structure (1911–2025). The cluster “Alzheimer's disease–oxidative stress–neuroprotection” occupies the motor-theme quadrant, indicating high development and strong network integration. “Caffeine–memory–cognition” functions as a basic theme, while “deep learning” and “transient epileptic amnesia” appear as niche themes. (B) Recent-period thematic structure (2021–2025). The “caffeine–cognition–memory” cluster shifts toward the motor quadrant, reflecting increased consolidation and strategic relevance. “Coffee–Alzheimer's disease–cognitive function” strengthens its centrality, whereas “oxidative stress–neuroprotection–inflammation” stabilizes as a foundational basic theme. Emerging clusters (e.g., aging–gut microbiota–polyphenols) indicate interdisciplinary diversification.

Thematic map: last five years (2021–2025)

The recent-period thematic map (Figure 7B) reveals structural reconfiguration. The cluster “caffeine—cognition—memory” shifts toward the motor quadrant, indicating increased thematic consolidation and strategic relevance. “Coffee—Alzheimer's disease—cognitive function” strengthens its central position, reflecting translational expansion. In contrast, “oxidative stress—neuroprotection—inflammation” remains a basic theme, suggesting continued foundational importance but relative thematic stabilization. Emerging themes include “aging—gut microbiota—polyphenols,” signaling interdisciplinary diversification, while computational approaches such as “deep learning—machine learning” persist as niche domains. Collectively, the recent map indicates thematic drift toward integrative, translational, and data-driven paradigms, reflecting knowledge maturation and methodological modernization within the field.

Co-citation analysis

Source co-citation analysis was conducted using VOSviewer with a minimum threshold of 40 citations per source. Of 2626 indexed sources, 39 met this criterion and were included in the network. The resulting map comprised 39 nodes interconnected by 476 co-citation links, generating a cumulative total link strength (TLS) of 1739 and resolving into four distinct clusters, reflecting thematic consolidation within the intellectual structure (Figure 8).

Figure 8.

Source Co-Citation Network of Core Journals. VOSviewer-generated source co-citation map illustrating the intellectual structure of the field. Nodes represent highly cited sources, with node size proportional to citation frequency and link thickness indicating co-citation strength. Colors denote distinct thematic clusters reflecting multidisciplinary neuroscience (red), clinical neurology and epidemiology (green), and psychopharmacology and experimental research (blue). The dense interconnections among clusters demonstrate strong bibliometric proximity and transdisciplinary integration within the domain (https://tinyurl.com/2avvlzoc).

The most prominent sources, demonstrating high nodal centrality and co-citation strength, include Nature, Science, Journal of Neuroscience, New England Journal of Medicine, and The Lancet. These journals exhibit both high citation frequency and elevated TLS, positioning them as structural hubs that anchor the domain's epistemic architecture.

Cluster 1 (red) is dominated by high-impact multidisciplinary and neuroscience journals (e.g., Nature, Science, Neuron, PNAS), representing foundational experimental and translational neuroscience. Cluster 2 (green) comprises clinical neurology and epidemiology outlets (e.g., Neurology, Annals of Neurology, Archives of Neurology, The Lancet, NEJM), reflecting clinical and population-based research. Cluster 3 (blue) centers on psychopharmacology and toxicology (e.g., Psychopharmacology, Pharmacological Reviews, Food and Chemical Toxicology), indicating mechanistic and therapeutic inquiry. Cluster 4 (yellow) integrates specialized neuroscience and disease-focused journals (e.g., Journal of Alzheimer's Disease, Nature Reviews Neuroscience), representing thematic specialization.

Dense inter-cluster bibliometric proximity, particularly between multidisciplinary science journals and clinical neurology outlets, demonstrates strong transdisciplinary knowledge integration. High co-citation strength across clusters indicates an intellectually cohesive yet thematically differentiated field, where foundational neuroscience, clinical medicine, pharmacology, and epidemiology converge within a structurally interconnected citation network.

Emerging themes: overlay visualization

Figure 9 presents a VOSviewer overlay visualization of keyword co-occurrence, where node size reflects frequency, link thickness indicates co-occurrence strength, and color denotes average publication year (blue = earlier; yellow = recent). Structurally central terms such as caffeine, Alzheimer's disease, neuroprotection, and hippocampus demonstrate longitudinal stability, anchoring the field's intellectual core.

Figure 9.

Overlay Visualization of Keyword Co-Occurrence Network. VOSviewer-generated overlay map illustrating the temporal evolution of research themes. Node size represents keyword frequency, link thickness indicates co-occurrence strength, and colors reflect the average publication year (blue = earlier; yellow = recent). The visualization demonstrates temporal stratification within the intellectual structure, highlighting stable core constructs alongside emerging themes that signal methodological modernization and translational expansion.

Earlier themes emphasize neurochemical and experimental paradigms (e.g., adenosine, dopamine, rats, working memory), reflecting a mechanistic turn. More recent keywords—including mild cognitive impairment, older adults, molecular docking, and deep learning—indicate thematic drift toward translational expansion and methodological modernization. This temporal stratification suggests knowledge maturation, shifting from reductionist neuroscience to integrative, clinically oriented, and computationally enhanced research trajectories.

Thematic classification of highly cited articles

As shown in Table 2, the highly cited articles (1992–2023) cluster around caffeine neuropharmacology, green tea polyphenols (EGCG)–mediated neuroprotection, dietary flavonoids and dementia risk, epidemiological determinants of Alzheimer's disease, and translational models. The most cited study, published in American Journal of Epidemiology (2002; 1101 citations), examined Alzheimer's risk factors, while mechanistic EGCG and flavonoid studies in Free Radical Biology and Medicine (2001; 800 citations) and Journal of Neuroscience (2005; 663 citations) shaped the molecular framework. Additional influential works in Neuroscience (2006; 430 citations) and Current Neuropharmacology (2015; 439 citations) highlight caffeine's role in cognitive modulation and amyloid regulation.

Table 2.

Highly cited articles.

Rank	Journal	Year	Volume	DOI	Citations
1	American Journal of Epidemiology	2002	156	10.1093/aje/kwf074	1101
2	Free Radical Biology and Medicine	2001	30	10.1016/S0891-5849(00)00498-6	800
3	Journal of Neuroscience	2005	25	10.1523/JNEUROSCI.1521-05.2005	663
4	European Journal of Epidemiology	2000	16	10.1023/A:1007614613771	616
5	Pediatrics	1992	90	https://pubmed.ncbi.nlm.nih.gov/1437425	604
6	Journal of Neurochemistry	2001	78	10.1046/j.1471-4159.2001.00490.x	591
7	Pharmacology & Toxicology	1995	76	10.1111/j.1600-0773.1995.tb00111.x	527
8	Psychopharmacology	2002	164	10.1007/s00213-002-1217-9	513
9	Biotechnology and Bioengineering	2017	114	10.1002/bit.26045	495
10	Neonatology	2023	120	10.1159/000528914	492
11	American Journal of Medicine	2006	119	10.1016/j.amjmed.2006.03.045	485
12	Brain Research	2008	1214	10.1016/j.brainres.2008.02.107	451
13	Current Neuropharmacology	2015	13	10.2174/1570159X13666141210215655	439
14	Neuroscience	2006	142	10.1016/j.neuroscience.2006.07.021	430
15	Journal of Biological Chemistry	2002	277	10.1074/jbc.M202832200	425

Discussion

This century-scale bibliometric analysis demonstrates that research on coffee and tea in relation to cognitive aging has evolved into a structurally mature and mechanistically anchored domain, with Alzheimer's disease and oxidative stress emerging as central organizing themes. While beverage-related cognitive research remains foundational, recent years reveal a discernible shift toward disease-specific neuroprotection and translational integration. These findings directly address the study's objective by clarifying the intellectual architecture, temporal evolution, and emerging frontiers shaping stimulant-related dementia research.

The observed productivity and citation hierarchies indicate that coffee and tea research in cognitive aging is embedded within a structurally stratified global knowledge system. The dominance of high-income countries reflects sustained biomedical investment, advanced research infrastructure, and long-standing regulatory engagement with neurodegenerative disorders.^15,16 The divergence between productivity and citation intensity suggests that scientific influence is not purely volume-driven; rather, epistemic authority is concentrated within systems capable of generating high-visibility, internationally collaborative outputs.¹⁷ Mid-volume countries with elevated citation impact likely derive influence through participation in high-impact consortia rather than independent leadership. Persistent North–South disparities further point to constrained inclusion of low- and middle-income regions in agenda-setting processes.¹⁸ Comparable bibliometric investigations in Alzheimer's disease and nutritional neuroscience report similar core–periphery configurations and impact asymmetries,^19,20 reinforcing that stimulant-related cognitive research operates within a hierarchically organized yet globally interconnected scientific ecosystem shaped by geopolitical research capacity and collaborative centrality. This global interpretation should, however, be considered alongside the English-language restriction applied during data extraction, which may have comparatively underrepresented contributions from non-English-speaking regions and partially amplified the apparent dominance of English-speaking and high-income research systems in the observed collaboration topology, country rankings, and leading-scholar distribution.

R.A. Cunha, based at the University of Coimbra's Center for Neuroscience and Cell Biology, is the most prolific scholar in this domain, with 25 publications and an H-index of 84. His research has decisively advanced understanding of how caffeine and related methylxanthines regulate cognition through adenosine receptor modulation. Experimental studies demonstrated that early-life caffeine exposure induces sex-dependent memory alterations,²¹ while chronic caffeine intake counteracts diabetes-associated hippocampal dysfunction²² and restores memory deficits in depression-prone models.²³ Cunha's work further established that hippocampal A₂A receptor activation is sufficient to impair memory via cyclic adenosine monophosphate response element-binding protein (CREB) signaling²⁴ and that amygdala A₂A receptors regulate fear memory.²⁵ Mechanistic and translational insights extend to ADHD models,²⁶ chlorogenic acids,²⁷ theobromine,²⁸ and chronic caffeine–synaptic metabolic coupling.²⁹ Comprehensive reviews on adenosine receptors in Alzheimer's disease further consolidate his theoretical framework.³⁰ Collectively, these studies position adenosinergic neuromodulation as a central axis linking coffee bioactives to neuroprotection and cognitive resilience.

The thematic evolution analysis indicates a progressive transition from mechanistic neurobiology toward integrative and translational research frameworks. During the early period (1911–2011), the field was dominated by oxidative stress, neuroprotection, and adenosine receptor–mediated mechanisms, reflecting foundational investigations into caffeine's neuropharmacological effects.^31,32 Between 2012 and 2018, Alzheimer's disease and oxidative stress consolidated as central themes, while polyphenols and green tea gained prominence, aligning with growing interest in dietary antioxidants and dementia prevention.^{3,6,8,10,15,33} From 2019 to 2022, cognition and memory emerged as dominant drivers, paralleling epidemiological evidence linking moderate coffee and tea intake with reduced cognitive decline.^1,3,15 In the most recent phase (2023–2025), the emergence of gut microbiota and deep learning reflects interdisciplinary expansion toward gut–brain axis research and computational modeling approaches in neurodegeneration.^34,35

The strategic thematic map highlights a structurally mature yet heterogeneous intellectual landscape within the field. The cluster centered on Alzheimer's disease, oxidative stress, and neuroprotection occupies the motor-theme quadrant, indicating both high centrality and strong internal cohesion. This positioning confirms its role as the primary driving axis of the field, integrating mechanistic neurobiology with translational neurodegeneration research. The strong PageRank values of its core publications further reinforce its dominance and sustained scholarly influence.¹⁷ In contrast, the caffeine–memory–cognition cluster is positioned within the basic-theme quadrant, reflecting high relevance but comparatively lower internal density. This suggests that while caffeine-related cognitive research is foundational and widely interconnected across subdomains, it remains conceptually broad rather than tightly specialized. The coffee–dementia–cognitive function cluster demonstrates moderate centrality with lower density, indicating an actively evolving domain where epidemiological and clinical associations are still consolidating into a more cohesive research stream.

The transient epileptic amnesia–epilepsy cluster represents a classic niche theme: highly developed internally yet weakly connected to the broader intellectual network. Its specialized neurological focus limits cross-theme integration despite internal maturity. The most structurally isolated cluster, deep learning, also falls within the niche quadrant but exhibits extremely low centrality. Its recent publication dates and low PageRank values suggest an emerging methodological application rather than an established thematic core. The limited integration of computational approaches into the dominant neurodegeneration and cognitive frameworks indicates that artificial intelligence remains peripheral within this dataset. Overall, the map demonstrates a field anchored by neurodegenerative mechanisms, supported by foundational cognitive pharmacology, and progressively incorporating lifestyle exposures such as coffee and tea. While specialized neurological subfields remain distinct, computational methodologies have yet to achieve structural integration.^34,36 This configuration reflects a mature but selectively interconnected research ecosystem, with clear potential for greater interdisciplinary convergence in future investigations.

Comparing the cumulative structure (1911–2025; Figure 7B) with the recent configuration (2021–2025; Figure 7A) enables assessment of thematic drift and knowledge maturation.¹³ In the whole-period map (Figure 7B), “caffeine–cognition–memory” and “coffee–Alzheimer's disease–cognitive function” function as motor themes, reflecting historically central and cohesive translational domains. “Oxidative stress–neuroprotection–inflammation” appears as a basic theme, consistent with its enduring mechanistic role in Alzheimer's disease. Interdisciplinary expansion is evident in microbiome- and polyphenol-related clusters, aligning with growing attention to the gut–brain axis and nutraceutical modulation. In contrast, the recent map (Figure 7A) shows consolidation of “Alzheimer's disease–oxidative stress–neuroprotection” as the dominant motor theme, indicating a mechanistic turn toward pathway-centered neurodegeneration research. Meanwhile, caffeine-related cognition shifts to the basic quadrant, suggesting integration into broader disease frameworks rather than thematic dominance. “Deep learning” remains a niche theme despite its broader scientific rise. Collectively, Figures 7A–B indicate paradigm transition toward mechanistic neuroprotection, with lifestyle exposures acting as connective platforms. Interpretation should acknowledge standard co-word limitations, including dependence on author keywords and heuristic clustering variability.³⁷

Co-citation analysis is a science-mapping technique used to identify the intellectual structure of a research field by examining how frequently two sources are cited together, thereby revealing conceptual proximity and knowledge consolidation.³⁸ The present results indicate that coffee and tea research in cognitive aging is anchored in a relatively small set of highly influential journals, reflecting strong epistemic concentration. The prominence of multidisciplinary outlets such as Nature and Science, alongside leading neuroscience and clinical journals, demonstrates that this domain is embedded within high-impact biomedical science.³⁹ The four-cluster configuration reveals thematic differentiation across experimental neuroscience, clinical neurology and epidemiology, psychopharmacology, and Alzheimer's-focused specialization. Dense inter-cluster connectivity highlights robust transdisciplinary integration, linking mechanistic, translational, and clinical research streams. Collectively, these findings indicate a mature, hierarchically organized, and intellectually cohesive field with strong translational continuity from molecular mechanisms to population-level outcomes.

Trend analysis indicates that “deep learning” has emerged as a prominent and rapidly expanding theme, reflecting the integration of artificial intelligence into tea-related quality assessment, agricultural monitoring, and neurophysiological analytics. Early applications demonstrated the feasibility of deep learning models such as Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) for identifying black tea storage years using hyperspectral imaging⁴⁰ and for soil environment classification in Pu-erh tea plantations.⁴¹ Subsequent studies extended deep learning to electroencephalography (EEG)-based classification of neuromodulatory effects⁴² and advanced neurophysiological signal decoding using CNN architectures.⁴³ In agricultural systems, deep learning enhanced Internet of Things (IoT) anomaly detection in tea plantations⁴⁴ and unmanned aerial vehicle (UAV)-based nitrogen prediction through ResNet and CNN models.⁴⁵ Applications further expanded to intelligent sentiment modeling⁴⁶ and counterfeit tea detection using image-based classifiers.⁴⁷ More recent innovations include deep learning-assisted colorimetric sensor arrays for withering monitoring,⁴⁸ commodity price forecasting,⁴⁹ and portable near-infrared spectroscopy (NIR)-based quality detection.⁵⁰ Collectively, these findings demonstrate a methodological shift toward data-driven, predictive, and real-time analytical frameworks within tea-related research ecosystems.

These highly cited articles are foundational because they collectively established the mechanistic, epidemiological, translational, and methodological pillars of research linking caffeine, polyphenols, cognition, and neurodegeneration. Early mechanistic studies clarified caffeine's action through adenosine receptor antagonism and its downstream neuroprotective effects,^51,52 while flavonoid and EGCG investigations demonstrated antioxidant, anti-amyloidogenic, and tau-modulating mechanisms in cellular and transgenic models.^5,53–56 Parallel epidemiological studies provided population-level validation, linking flavonoid intake, coffee consumption, and lifestyle factors to reduced dementia risk.^2,9,10 Broader cognitive frameworks were strengthened through controlled human performance studies and comprehensive neuropharmacological reviews.^57–59 Additionally, methodological advances, including blood–brain barrier modeling and clinical guideline development, expanded translational and clinical applicability.^60,61 Together, these works shaped the conceptual architecture that continues to guide contemporary research.

Limitations

This study is subject to several limitations. First, reliance on Scopus may omit relevant publications indexed exclusively in other databases, potentially affecting coverage breadth. Second, restricting inclusion to English-language original articles may underrepresent regional scholarship. Third, co-word and co-citation analyses depend on author keyword consistency and citation behavior, which can introduce conceptual bias. Finally, heuristic clustering algorithms may yield non-deterministic results. Future research should incorporate multi-database integration, multilingual datasets, and comparative language-sensitive bibliometric modeling to enhance robustness and provide a more globally representative view of coffee and tea research in Alzheimer's disease and cognitive aging.

Conclusions

This century-spanning bibliometric analysis demonstrates that coffee and tea research in cognitive aging has matured into a structurally consolidated and mechanistically centered domain anchored in Alzheimer's disease and oxidative stress pathways. Across 2873 original research articles, the field exhibits sustained exponential growth, a clearly defined core–periphery global structure, and progressive thematic consolidation. Beverage-related cognitive research has transitioned from general pharmacological exploration toward disease-specific neuroprotective and translational frameworks. Integrated analyses of collaboration networks, co-citation structures, and thematic mapping reveal an intellectually cohesive yet increasingly interdisciplinary field, incorporating microbiome science, multi-omics profiling, and computational methodologies, including emerging artificial intelligence approaches. By restricting inclusion to primary evidence and spanning more than a century, this study provides the first unified scientometric synthesis of coffee and tea research in dementia. Future research should expand multi-database integration, develop biomarker-driven translational models, and implement explainable AI frameworks to bridge epidemiological exposure data with mechanistic validation. Collectively, these findings clarify the knowledge architecture shaping stimulant-related dementia research and delineate strategic pathways for advancing evidence-based cognitive health interventions.

Footnotes

Acknowledgements

This research article is derived from a research grant funded by the Research, Development, and Innovation Authority (RDIA)—Kingdom of Saudi Arabia—with grant number (12899-jazzan-2023-JZU-R-2-1-HW).

ORCID iD

Siddig Ibrahim Abdelwahab

Ethical considerations

Not applicable

Consent to participate

Not applicable

Consent for publication

Not applicable

Author contribution(s)

Manal Mohamed Elhassan Taha: Formal analysis; Funding acquisition; Resources; Writing – original draft.

Siddig Ibrahim Abdelwahab: Conceptualization; Data curation; Formal analysis; Funding acquisition; Investigation; Methodology; Project administration; Resources; Software; Supervision; Validation; Visualization; Writing – original draft; Writing – review & editing.

Khaled A. Sahli: Data curation; Funding acquisition; Resources; Writing – review & editing.

Mohammed Jeraiby: Funding acquisition; Resources; Software; Supervision.

Marwa Qadri: Funding acquisition; Project administration; Resources; Writing – review & editing.

Abdulaziz Alarifi: Methodology; Project administration; Resources; Software; Supervision.

Abdullah Farasani: Conceptualization; Data curation; Formal analysis; Funding acquisition; Investigation; Methodology; Project administration; Resources; Software; Supervision; Validation; Visualization; Writing – original draft; Writing – review & editing.

Jobran M. Moshi: Conceptualization; Data curation; Formal analysis; Funding acquisition; Investigation; Methodology; Project administration; Resources; Software; Supervision; Validation; Visualization; Writing – original draft; Writing – review & editing.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research article is derived from a research grant funded by the Research, Development, and Innovation Authority (RDIA)—Kingdom of Saudi Arabia—with grant number (12821-jazzan-2023-JZU-R-2-1-HW); Jazan University.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data availability statement

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

References

Zhang

Liu

, et al. Coffee and tea intake, dementia risk, and cognitive function. JAMA 2026; 335: 961–974.

Commenges

Scotet

Renaud

, et al. Intake of flavonoids and risk of dementia. Eur J Epidemiol 2000; 16: 357–363.

Shen

Chen

Gong

, et al. Association between plant-based diets and incident dementia: results from prospective cohort studies and a meta-analysis. J Prev Alzheimers Dis 2026; 13: 100457.

Mazidi

Mikhailidis

Dehghan

, et al. The association between coffee and caffeine consumption and renal function: insight from individual-level data, Mendelian randomization, and meta-analysis. Arch Med Sci 2021; 18: 900.

Rezai-Zadeh

Shytle

Sun

, et al. Green tea epigallocatechin-3-gallate (EGCG) modulates amyloid precursor protein cleavage and reduces cerebral amyloidosis in Alzheimer transgenic mice. J Neurosci 2005; 25: 8807–8814.

Rainey-Smith

Sewell

Brown

, et al. Moderate coffee and tea consumption is associated with slower cognitive decline. J Alzheimers Dis 2025; 107: 170–178.

Pergolizzi Jr

Tenenbaum

Pergolizzi

, et al. Neurocognitive and neurological effects of coffee and caffeine: a narrative review. Cureus 2025; 17: e94742.

Socała

Szopa

Serefko

, et al. Neuroprotective effects of coffee bioactive compounds: a review. Int J Mol Sci 2020; 22: 107.

Lindsay

Laurin

Verreault

, et al. Risk factors for Alzheimer's disease: a prospective analysis from the Canadian Study of Health and Aging. Am J Epidemiol 2002; 156: 445–453.

10.

Dai

Borenstein

, et al. Fruit and vegetable juices and Alzheimer's disease: the Kame Project. Am J Med 2006; 119: 751–759.

11.

Meng

Wang

, et al. Bibliometric analysis and description of research trends on nutritional management in Alzheimer's disease patients (1988-2024). Front Nutr 2025; 12: 1545951.

12.

Meng

Cui

Liang

, et al. Trends and hotspots in tea and Alzheimer's disease research from 2014 to 2023: a bibliometric and visual analysis. Heliyon 2024; 10: e30063.

13.

Aria

Cuccurullo

. Bibliometrix: an R-tool for comprehensive science mapping analysis. J Informetr 2017; 11: 959–975.

14.

Van Eck

Waltman

. Software survey: VOSviewer, a computer program for bibliometric mapping. Scientometrics 2010; 84: 523–538.

15.

Vilor-Tejedor

Danso

Albanus

RDO

, et al. Advancing global dementia research through equity and inclusion. Alzheimers Dement 2026; 22: e71069.

16.

Murray

Smith

Menon

, et al. Accelerating biomedical discoveries in brain health through transformative neuropathology of aging and neurodegeneration. Neuron 2025; 113: 3703–3721.

17.

Verma

Kumar

Pabreja

, et al. A decade of data science research: insights from a bibliometric study. Discover Data 2025; 3: 61.

18.

Anghileri

Kandel

Austen

, et al. Rethinking north–south research partnerships amidst global uncertainties: leveraging lessons learned from UK GCRF projects during COVID-19. Land (Basel) 2023; 12: 791.

19.

Sun

Gong

, et al. Bibliometric analysis of neuroinflammation and Alzheimer's disease. Front Aging Neurosci 2024; 16: 1423139.

20.

Ding

Xiong

Ouyang

. A Bibliometric analysis of neuroscience tools use in construction health and safety management. Sensors 2023; 23: 9522.

21.

Ardais

Rocha

Borges

, et al. Caffeine exposure during rat brain development causes memory impairment in a sex selective manner that is offset by caffeine consumption throughout life. Behav Brain Res 2016; 303: 76–84.

22.

Duarte

JMN

Skoug

Silva

, et al. Impact of caffeine consumption on type 2 diabetes-induced spatial memory impairment and neurochemical alterations in the hippocampus. Front Neurosci 2019; 13: 1015.

23.

Machado

Simoes

Silva

, et al. Caffeine reverts memory but not mood impairment in a depression-prone mouse strain with up-regulated adenosine A2A receptor in hippocampal glutamate synapses. Mol Neurobiol 2017; 54: 1552–1563.

24.

Rial

Canas

, et al. Optogenetic activation of intracellular adenosine A2A receptor signaling in the hippocampus is sufficient to trigger CREB phosphorylation and impair memory. Mol Psychiatry 2015; 20: 1339–1349.

25.

Simoes

Machado

Gonçalves

, et al. Adenosine A2A receptors in the amygdala control synaptic plasticity and contextual fear memory. Neuropsychopharmacology 2016; 41: 2862–2871.

26.

França

Schamne

de Souza

, et al. Caffeine consumption plus physical exercise improves behavioral impairments and stimulates neuroplasticity in spontaneously hypertensive rats (SHR): an animal model of attention deficit hyperactivity disorder. Mol Neurobiol 2020; 57: 3902–3919.

27.

Fernandes

MYD

Dobrachinski

Silva

, et al. Neuromodulation and neuroprotective effects of chlorogenic acids in excitatory synapses of mouse hippocampal slices. Sci Rep 2021; 11: 10488.

28.

Valada

Alçada-Morais

Cunha

, et al. Thebromine targets adenosine receptors to control hippocampal neuronal function and damage. Int J Mol Sci 2022; 23: 10510.

29.

Lopes

Oliveira

Gaspar

, et al. Effects of chronic caffeine consumption on synaptic function, metabolism and adenosine modulation in different brain areas. Biomolecules 2023; 13: 106.

30.

Canas

Cunha

Agostinho

. Adenosine receptors in Alzheimer’s disease. Receptors 2018; 34: 259–280.

31.

Zhong

Dong

Zhou

, et al. Caffeine's neuroprotective effect on memory impairment: suppression of adenosine A(2)A receptor and enhancement of tyrosine hydroxylase in dopaminergic neurons under hypobaric hypoxia conditions. CNS Neurosci Ther 2024; 30: e70134.

32.

Gomes

Kaster

Tomé

, et al. Adenosine receptors and brain diseases: neuroprotection and neurodegeneration. Biochim Biophys Acta Biomembr 2011; 1808: 1380–1399.

33.

Darvesh

Carroll

Bishayee

, et al. Oxidative stress and Alzheimer's disease: dietary polyphenols as potential therapeutic agents. Expert Rev Neurother 2010; 10: 729–745.

34.

Zhu

Yang

, et al. Artificial intelligence empowers gut microbiota research in neurodegenerative diseases molecular mechanisms and precision therapy. iScience 2025; 28: 114044.

35.

Guan

Tang

, et al. Herbal medicine and gut microbiota: exploring untapped therapeutic potential in neurodegenerative disease management. Arch Pharm Res 2024; 47: 146–164.

36.

Zhang

, et al. Artificial intelligence in neurodegenerative diseases research: a bibliometric analysis since 2000. Front Neurol 2025; 16: 1607924.

37.

Gkintoni

Vassilopoulos

Nikolaou

. Mindfulness-based cognitive therapy in clinical practice: a systematic review of neurocognitive outcomes and applications for mental health and well-being. J Clin Med 2025; 14: 1703.

38.

Hjørland

. Citation analysis: a social and dynamic approach to knowledge organization. Inf Process Manag 2013; 49: 1313–1325.

39.

Shirai

Kuriki

Otsuka

, et al. Green tea and coffee intake and risk of cognitive decline in older adults: the National Institute for Longevity Sciences, Longitudinal Study of Aging. Public Health Nutr 2020; 23: 1049–1057.

40.

Hong

Zhang

Kong

, et al. Identification of storage years of black tea using near-infrared hyperspectral imaging with deep learning methods. Infrared Phys Technol 2021; 114: 103666.

41.

Cai

Yuan

Liu

, et al. Deep learning model for soil environment quality classification of Pu-erh tea. Forests 2022; 13: 1778.

42.

Uyulan

Mayor

Steffert

, et al. Classification of the central effects of transcutaneous electroacupuncture stimulation (TEAS) at different frequencies: a deep learning approach using wavelet packet decomposition with an entropy estimator. Appl Sci (Basel) 2023; 13: 2703.

43.

Vasei

Gediya

Ravan

, et al. Investigating brain responses to transcutaneous electroacupuncture stimulation: a deep learning approach. Algorithms 2024; 17: 477.

44.

Wang

Feng

Zhang

, et al. Detection and correction of abnormal IoT data from tea plantations based on deep learning. Agriculture 2023; 13: 480.

45.

Wang

Zhao

, et al. Pixel-class prediction for nitrogen content of tea plants based on unmanned aerial vehicle images using machine learning and deep learning. Expert Syst Appl 2023; 227: 120351.

46.

Alohali

Alasmari

Almalki

, et al. Textual emotion analysis using improved metaheuristics with deep learning model for intelligent systems. Trans Emerg Telecommun Technol 2024; 35: e4846.

47.

Besharati

Pourdarbani

Sabzi

, et al. Counterfeit detection of Iranian black tea using image processing and deep learning based on patched and unpatched images. Horticulturae 2024; 10: 665.

48.

Cai

Lin

Ouyang

, et al. Deep learning-assisted self-cleaning cellulose colorimetric sensor array for monitoring black tea withering dynamics. Food Chem 2025; 487: 144727.

49.

Drira

Boutouria

Boujelbéne

. Impact of key risk, green and behavioral factors on agricultural commodity price forecasting amid volatile markets: a deep learning approach. EuroMed J Bus Epub ahead of print 2025. DOI: 10.1108/EMJB-01-2025-0020

50.

Sheng

Adade

SYSS

, et al. Comparison of machine learning and deep learning models for detecting quality components of vine tea using smartphone-based portable near-infrared device. Food Control 2025; 174: 111244.

51.

Arendash

Schleif

Rezai-Zadeh

, et al. Caffeine protects Alzheimer's mice against cognitive impairment and reduces brain β-amyloid production. Neuroscience 2006; 142: 941–952.

52.

Fredholm

. Adenosine, adenosine receptors and the actions of caffeine. Pharmacol Toxicol 1995; 76: 93–101.

53.

Ishige

Schubert

Sagara

. Flavonoids protect neuronal cells from oxidative stress by three distinct mechanisms. Free Radic Biol Med 2001; 30: 433–446.

54.

Levites

Amit

Youdim

MBH

, et al. Involvement of protein kinase C activation and cell survival/cell cycle genes in green tea polyphenol (-)-epigallocatechin 3-gallate neuroprotective action. J Biol Chem 2002; 277: 30574–30580.

55.

Levites

Weinreb

Maor

, et al. Green tea polyphenol (-)-epigallocatechin-3-gallate prevents N-methyl-4-phenyl- 1,2,3,6-tetrahydropyridine-induced dopaminergic neurodegeneration. J Neurochem 2001; 78: 1073–1082.

56.

Rezai-Zadeh

Arendash

Hou

, et al. Green tea epigallocatechin-3-gallate (EGCG) reduces β-amyloid mediated cognitive impairment and modulates tau pathology in Alzheimer transgenic mice. Brain Res 2008; 1214: 177–187.

57.

Bellinger

Stiles

Needleman

. Low-level lead exposure, intelligence and academic achievement: a long- term follow-up study. Pediatrics 1992; 90: 855–861.

58.

Cappelletti

Piacentino

Sani

, et al.

Caffeine: cognitive and physical performance enhancer or psychoactive drug?

Curr Neuropharmacol 2015; 13: 71–88.

59.

Lieberman

Tharion

Shukitt-Hale

, et al. Effects of caffeine, sleep loss, and stress on cognitive performance and mood during U.S. Navy SEAL training. Psychopharmacology 2002; 164: 250–261.

60.

Sweet

Carnielli

Greisen

, et al. European consensus guidelines on the management of respiratory distress syndrome: 2022 update. Neonatology 2023; 120: 3–23.

61.

Wang

Abaci

Shuler

. Microfluidic blood–brain barrier model provides in vivo-like barrier properties for drug permeability screening. Biotechnol Bioeng 2017; 114: 184–194.

A century of coffee and tea research in cognitive health and Alzheimer's disease: Structural,thematic,and translational insights (1911–2025)

Abstract

Background

Objective

Methods

Results

Conclusions

Keywords

Introduction

Methods

Justification for the use of Scopus

Search strategy

Analyses

Results

Data overview

Annual publication output and citation dynamics

Global scientific production and collaboration structure

Most productive affiliations

Bradford's Law analysis of journal productivity

Leading scholars

Temporal trajectory panels

Thematic evolution across four periods (1911–2025)

Thematic map

Thematic map: whole period (1911–2025)

Thematic map: last five years (2021–2025)

Co-citation analysis

Emerging themes: overlay visualization

Thematic classification of highly cited articles

Discussion

Limitations

Conclusions

Footnotes

Acknowledgements

ORCID iD

Ethical considerations

Consent to participate

Consent for publication

Author contribution(s)

Funding

Declaration of conflicting interests

Data availability statement

References