缺血性卒中特异性认知功能障碍：研究演变的文献计量学分析。
全文（开放获取） · 中文译文
摘要 背景：卒中后认知障碍（PSCI）是卒中后的常见并发症，显著影响患者的生活质量和长期预后。尽管PSCI已被广泛研究，但大多数研究将缺血性卒中和出血性卒中合并在一起。从文献计量学角度来看，缺血性卒中特异性认知功能障碍（ISCD）背后独特的炎症机制和血管损伤是否遵循了独立的研究轨迹仍属未知。 方法：使用Web of Science核心合集（2004—2024年），我们的研究确定了1,062篇聚焦于ISCD的文献进行文献计量学分析。在与PubMed交叉验证确认文献完全一致后，采用文献计量学工具（CiteSpace、VOSviewer、基于R的bibliometrix包）分析国家、机构、期刊和作者的发表趋势、贡献及合作关系。亚组分析量化了炎症生物标志物、评估工具和预测模型的趋势。 结果：年发文量在2022年达到峰值（n = 133），随后在2023年和2024年有所下降。中国在生产力方面占主导地位（31.4%），而美国表现出最高的合作中心性（0.30）。关键词突现识别出缺血性特异性热点，包括“模型”（2021年，突现强度=4.66）、“验证”（2022年，突现强度=4.96）和“炎症”（2021年，突现强度=3.83）。炎症生物标志物和评估工具一直是ISCD的研究热点，近期的研究努力正转向开发急性缺血性卒中后认知结局的预测模型。 结论：这项缺血性特异性文献计量分析证实，炎症生物标志物、评估工具和预测模型是ISCD持续的热点，并且明显地从描述性研究转向预测模型的开发和验证。这些发现为指导未来研究方向提供了基于文献的定量依据。 引言 卒中是全球第二大致死原因，其中缺血性卒中（IS）构成其主要亚型（约占80%的病例）（）。这种脑血管病变的特征是脑动脉闭塞，导致缺血、缺氧以及随之而来的神经结构和功能损害，卒中后认知障碍（PSCI）影响近三分之二（67%）的幸存者，显著影响功能恢复和长期生活质量（, ）。 在缺血性卒中（IS）中，组织缺氧和代谢紊乱导致细胞损伤或死亡，释放出损伤相关分子模式（DAMPs），如高迁移率族蛋白B1（HMGB1）、ATP和热休克蛋白（）。这些分子在再灌注后被免疫细胞识别，从而诱导促炎细胞因子（如IL-6、TNF-α）的释放，同时激活补体系统和中性粒细胞浸润，加重炎症损伤（）。而炎症反应又通过免疫激活和血管损伤等机制加剧再灌注损伤（）。这种相互作用形成恶性循环，构成缺血性卒中的核心病理机制，严重影响患者神经功能和认知结局的预后。 鉴于缺血性卒中后认知功能障碍（ISCD）独特的病理生理机制及其在临床上的高患病率，亟待对其研究现状进行系统分析。文献计量学分析已成为一种强有力的工具，可定量评估科学领域的发展、合作和新兴趋势（）。该方法已越来越多地应用于包括卒中研究在内的各医学学科，以绘制知识结构、识别关键贡献者并预测前沿主题（, ）。例如，近期文献计量学研究揭示了卒中治疗中干细胞（）、卒中中炎症与线粒体应激（）等领域的研究演进。然而，尽管神经病学领域已有大量文献计量学综述，但尚无一项研究专门运用文献计量学和可视化分析方法对ISCD的主题演进和研究重点进行全面分析。因此，本文献计量学研究旨在系统描绘ISCD的研究现状，特别关注识别关键研究热点和新兴趋势。最终，我们力求加深对该疾病的理解，并为未来创新提供指导。 材料与方法 数据收集与检索策略 为确保文献数据的全面性并检验使用单一数据库可能带来的选择偏倚，我们的研究通过以PubMed数据库补充Web of Science核心合集（WoSCC）进行了交叉验证分析。我们的检索词源自PubMed中针对“缺血性卒中”和“认知功能障碍”的医学主题词表（MeSH）框架。对于WoSCC，我们使用了等效的关键词组合以确保覆盖范围一致。完整的检索策略和结果见，。文献检索于2025年2月7日进行。两个数据库的检索时间范围均为2004—2024年。对检索结果的比较显示，鉴于我们特定的检索词和时间范围，所有核心出版物均被WoSCC收录，未发现仅在PubMed中收录的独有出版物。这种完全重叠可以解释为：WoSCC作为文献计量学分析的“金标准”，采用了严格的期刊选择和 curation 流程（），而成熟研究领域的高质量文献通常同时被两个数据库收录，尽管PubMed的索引政策更广泛但不够严格。鉴于WoSCC在文献计量工具方面具有更优越的元数据标准化水平，我们选择WoSCC作为后续分析的唯一数据来源。 为了捕捉不断演变的研究趋势，同时保持对主要科学贡献的关注，我们将检索限制为2004年至2024年间发表的文献类型“Article”。语言限定为英文。然后，我们根据预先设定的纳入和排除标准，手动筛选检索到的记录的标题和摘要。纳入标准为：(i) 明确将缺血性卒中后认知障碍（ISCD）作为主要研究对象的研究，包括但不限于对其发病机制、神经影像学特征、危险因素、生物标志物、评估方法、干预措施或康复结局的探讨；以及(ii) 涉及缺血性卒中（IS）患者的研究，或利用公认且广泛接受的IS动物模型开展的基础实验研究。排除研究的标准为：(i) 仅关注一般性认知老化、阿尔茨海默病或其他与IS无关的继发性认知障碍，且未明确将IS确定为核心暴露或病因因素的研究；以及(ii) 与ISCD无直接相关性的研究，例如仅评估急性期溶栓治疗或血管内取栓疗效但未涉及认知结局的研究。最终，共纳入1,062篇独特文章用于后续分析。展示了整个筛选过程。 本研究的文献检索与筛选过程。 数据分析 导出的数据被导入CiteSpace（版本6.4.R1）进行初步去重，以确保数据集无重复（）。后续分析确定了关键指标，包括发文量、作者、机构和国家的贡献、期刊分布、关键词趋势以及引用模式。在“选择标准”中，“g指数”统一设置为K = 5以涵盖所有节点，默认参数为TOP N = 50和TOP N% = 10.0%。 该数据集还使用R语言（版本4.4.1）中的bibliometrix包和VOSviewer（版本1.6.20）进行分析，以生成定量文献计量指标和可视化结果（）。VOSviewer主要用于构建和可视化合作网络及关键词共现图谱，重点关注条目之间的关联强度。R语言的Bibliometrix包用于描述性统计分析和趋势绘图。 为了全面描绘 ISCD 研究领域的全貌，我们首先分析了发表量和引用量，以了解生产力和发展趋势。然后，我们考察了国家贡献和国际合作，以描绘全球合作网络。评估了机构表现，以识别主要研究中心及其组织间合作关系。随后，我们进行了期刊和共被引期刊分析，以勾勒核心发表渠道和学科知识基础。我们评价了作者和共被引作者，以确定关键意见领袖和有影响力的研究团队。此外，进行了关键词共现分析，以揭示研究热点和主题聚类。最后，我们利用参考文献分析来追溯该领域的知识起源和知识演化。这些分析共同构成了我们系统性文献计量分析的基础。 结果 发表与引用趋势分析 展示了2004年至2024年ISCD相关出版物和引用的时间演变。出版物数量经历了不同阶段（）。2004年至2019年期间以稳定但有波动的增长为特征，随后在2019年至2020年间急剧上升。2021年年度出版物超过100篇（n = 111），在2022年达到峰值133篇。然而，之后的几年（2023-2024）见证了产出的显著下降。如 所示，拟合曲线（R2 = 0.96）表明出版年份与产出量之间存在强相关性。虽然该曲线预测2023年和2024年的出版物分别为152篇和180篇，但实际数量仅为108篇和93篇。根据 ，引用数量逐年递增，在2024年达到最高水平（n = 4,482）。 2004年至2024年ISCD领域年度出版物和引用趋势。（A）2004年至2024年年度出版物和引用数量。（B）年度出版物的拟合曲线以及预测与实际出版物的对比。 国家和合作分析 在分析的时间范围内，有65个国家积极参与了ISCD领域。为了更好地解读各国的合作角色，我们采用了介数中心性，该指标衡量一个节点在连接网络中不同节点之间起桥梁作用的程度。较高的中心性值表示在连接不同研究群体方面具有更强中介作用，意味着更大的影响力，而较低的值则反映更边缘的位置。如和所示，中国以333篇出版物（占总数31.4%）居首位，但在全球合作网络中的整合程度有限（中心性 = 0.04）。美国是第二大贡献者（n = 264，占总数24.9%），同时表现出显著更强的合作影响力（中心性 = 0.30）。英国以74篇排名第三（7.0%，0.05）。日本位居第十，贡献了40篇出版物（3.8%，0.04）。 2004年至2024年ISCD研究领域的国家文献计量学分析。（A）产出最多的前10个国家。（B）通讯作者所在国家。（C）基于VOSviewer的国家合作网络。每个节点代表一个国家，节点越大表示该国在该领域的产出越多。节点之间的连线表示国家之间的合作关系，连线越粗表示合作越强。（D）国家的叠加可视化图谱。蓝色和紫色节点标注的国家在2017年之前的时间进程中较早开展了研究活动，而在2018年前后ISCD领域高产出的国家以黄色编码，在2020年或之后变得突出的国家显示为红色。 产出最多的前10个国家。 使用Bibliometrix提供了通讯作者所在国家的详细统计细分，区分了合作类型，以评估独立研究能力与国际合作情况。中国在单国出版物（SCP）和多国出版物（MCP）方面均遥遥领先。然而，在所有高产出国家中，SCP始终超过MCP，这表明相对于国际合作，国内研究产出更强。 此外，我们使用VOSviewer软件创建了,, ，展示了ISCD研究中全球合作的网络可视化和叠加映射。例如，中国与美国和澳大利亚保持着非常密切的合作，近年来科研活动尤为活跃。 机构分析 基于CiteSpace的文献计量分析显示，机构的产出量和中心性呈现明显的集中性，如 . 所示。排名前10的机构共贡献了总发文量的22.98%。首都医科大学以38篇（3.58%，中心性 = 0.17）领先，其次是哈佛大学（n = 30，2.82%，中心性 = 0.08）和美国退伍军人事务部（n = 29，2.73%，中心性 = 0.02）。 ISCD中机构的可视化。(A) ISCD中前10位机构的发文量和中心性。(B) 基于VOSviewer的机构合作网络。每个节点代表一个机构。节点越大表示产出越多。节点之间的连线表示它们之间的合作关系，连线越粗表示合作越强。 为了描绘机构研究活动与影响力，我们使用VOSviewer生成了一个由至少发表10篇论文的机构组成的合著网络。如图所示，首都医科大学、哈佛大学和哥伦比亚大学等知名机构以较大的节点表示，表明这些机构在该领域发表了大量论文。网络可视化揭示了广泛的机构间合作模式。例如，哈佛大学、哥伦比亚大学和约翰斯·霍普金斯大学，如它们之间密集的连线所示，形成了一个紧密合作的集群。 期刊及共被引期刊分析 在本研究中，所有文献来自400种不同期刊。，展示了该领域按文献量和影响因子（IF）排名前10的期刊。Stroke 以47篇位居首位，位于期刊引证报告（JCR）一区（Q1），IF = 7.9，其次是 Journal of Stroke and Cerebrovascular Diseases（n = 42，Q3，IF = 2）和 Frontiers in Neurology（n = 34，Q2，IF = 2.7），表明它们在 ISCD 研究中具有较高的活跃度。高影响因子反映期刊在学术界具有权威地位（）。The Lancet Neurology 的影响因子最高（Q1，46.6），但其累计发文量相对较低（n = 4）。类似地，Nature Reviews Neurology（Q1，28.2）仅发表了一篇相关文章。相比之下，Stroke 兼具高产出量（n = 47）与学术卓越性（Q1，IF = 7.9）。总体而言，虽然发文量和影响因子都是评估期刊学术影响力的关键指标，但它们并不总是呈正相关。 ISCD中期刊及共被引期刊的分析。(A) 产出最多的前10种期刊及其影响因子（IF）。相同颜色的条形图代表同一期刊引证报告（JCR）分区的期刊。(B) 按影响因子排序的前10种期刊及其发文量。(C) 期刊分析的叠加可视化图谱。(D) 期刊密度图。红色区域表示高密度。ISCD研究在期刊上的发文量与密度水平和字号大小均呈正相关。(E) CiteSpace生成的共被引期刊网络。当两种期刊被同一来源共同引用时，二者之间即产生连接，形成共被引关系。中心性>0.1的期刊将以紫色圆圈标出。(F) 共被引期刊的时间线图谱。不同期刊的时间区由其首次被共被引的最早年份决定。彩色环显示引文随时间分布的情况，内环代表早期年份，外环代表近年。 展示了ISCD中至少发表5篇论文的期刊的叠加可视化图谱和密度图。在高产且发文量居前五的期刊中，Frontiers in Aging Neuroscience被单独标为黄色，既表示其近期的发表时间框架，也表明其已成为前沿研究首选期刊。Stroke和Journal of Stroke & Cerebrovascular diseases被指定为高密度区域，说明它们是该领域的基础知识库。 我们使用 CiteSpace 完成了共被引期刊的可视化图谱。如 和 所示，Stroke 是最常被共被引的期刊（共被引次数 = 928），其次是 Neurology（共被引次数 = 692）、The Lancet Neurology（共被引次数 = 528）、Journal of Neurology, Neurosurgery & Psychiatry（共被引次数 = 382）和 PLOS One（共被引次数 = 374）。高共被引次数凸显了期刊在推动学科发展中的基础性作用。在共被引期刊网络中，PLOS One 的中心性最高（0.11）（节点以紫色圆圈表示），其次是 Neurology（0.11）和 The Lancet（0.11）。此外，最高产期刊与最常被共被引期刊之间存在显著的重合，Stroke、Neurology 和 PLOS One 均列入这两类排名的前 10 位。如 所示，期刊按时间顺序（从左到右）分布，以可视化共被引关系随时间的演变。被引频次前 10 的期刊均位于左侧，并具有宽大的红色圆环，为该领域的研究奠定了基础。 按被引频次计，ISCD 的前 10 位共被引期刊。 我们采用文献计量学分析，基于生产力和引文指标评估 ISCD 研究中研究人员的影响力。我们的分析确定了该时期 6,503 位贡献作者。排名前 10 的高产作者列于 。Mok Vincent 以 16 篇出版物居首，紧随其后的是 Lim Jae Sung 和 Yu KH（n = 15）。其中，Mok Vincent 的中心性最高（0.05），而这些值均相对较低。h 指数是衡量研究产出和影响的综合指标。h 指数为 N 表示该作者有 N 篇出版物，且每篇至少被引 N 次。其中，Sacco RL 的 h 指数最高（h = 134），反映出其发表记录在数量和质量之间达到了稳健的平衡。 排名前 10 的高产作者（依据已发表文章）。 合著网络分析提供了一种可靠的方法来突出主要作者并描绘其合作网络的密度（）。我们应用 VOSviewer 分析合著网络（至少发表 5 篇出版物且被引 50 次的作者），以确保聚焦于高影响力研究者（）。在图中，节点之间的连线代表作者之间的合作关系，连线的密度和粗细表示他们合作的强度。作者时区图（）追踪了作者在整个研究时间范围内的动态参与情况。一些高产作者如 Lim Jae Sung、Yu KH 和 Oh Mi Sun 从 2021 年开始发表文章。 ISCD中作者和共被引作者的可视化。（A）作者的合著网络
全文（开放获取） · 英文原文
【Abstract】 BACKGROUND: Post-Stroke Cognitive Impairment (PSCI) is a common complication after stroke, significantly affecting patients' quality of life and long-term prognosis. Although PSCI has been extensively studied, most research combines ischemic and hemorrhagic strokes. Whether the distinct inflammatory mechanisms and vascular injury underlying ischemic stroke-specific cognitive dysfunction (ISCD) have followed an independent research trajectory remains unknown from a bibliometric perspective. METHODS: Using the Web of Science Core Collection (2004-2024), our study identified 1,062 articles focused on ISCD for bibliometric analyses. Following cross-validation with PubMed confirming complete literature concordance, bibliometric tools (CiteSpace, VOSviewer, the R-based bibliometrix package) were employed to analyze publication trends, contributions and collaborations among countries, institutions, journals, and authors. Subgroup analyses quantified trends in inflammatory biomarkers, assessment tools and prediction models. RESULTS: Annual publication peaked in 2022 (n = 133), followed by a decline in 2023 and 2024. China dominated in productivity (31.4%), while the United States exhibited the highest collaborative centrality (0.30). Keyword bursts identified ischemic-specific hotspots, including "model" (2021, burst strength = 4.66), "validation" (2022, burst strength = 4.96), and "inflammation" (2021, burst strength = 3.83). Inflammatory biomarkers and assessment tools have consistently been research hotspots in ISCD, recent efforts are shifting toward the development of prediction models for cognitive outcomes following acute ischemic stroke. CONCLUSION: This ischemic-specific bibliometric analysis confirms that inflammatory biomarkers, assessment tools, and prediction models are persistent hotspots in ISCD, with a clear ongoing shift from descriptive studies toward predictive model development and validation. These findings provide a quantitative literature-based rationale for guiding future research directions. 【Introduction】 Stroke represents the second leading cause of global mortality, with ischemic stroke (IS) constituting its predominant subtype (approximately 80% of cases) (). This cerebrovascular pathology is characterized by cerebral artery occlusion, resulting in ischemia, hypoxia, and consequent structural and functional neurological damage, with post-stroke cognitive impairment (PSCI) affecting nearly two-thirds (67%) of survivors, significantly impacting functional recovery and long-term quality of life (, ). In IS, tissue hypoxia and metabolic disorders lead to cell damage or death, releasing damage-associated molecular patterns (DAMPs) such as high-mobility group box 1 (HMGB1), ATP, and heat shock proteins (). These molecules are recognized by immune cells after reperfusion, thereby inducing the release of proinflammatory cytokines (e.g., IL-6, TNF-α), while activating the complement system and neutrophil infiltration, exacerbating inflammatory damage (). The inflammatory response, in turn, exacerbates reperfusion injury through mechanisms such as immune activation and vascular damage (). This interaction forms a vicious cycle that constitutes a core pathological mechanism of ischemic stroke, seriously affecting the prognosis of patients’ neurological function and cognitive outcomes. Given the distinct pathophysiological mechanisms underlying ischemic stroke-specific cognitive dysfunction (ISCD) and its substantial clinical prevalence, a systematic analysis of the research landscape is urgently needed. Bibliometric analysis has emerged as a powerful tool to quantitatively assess the development, collaboration, and emerging trends within scientific fields (). This methodology has been increasingly applied across various medical disciplines, including stroke research, to map knowledge structures, identify key contributors, and forecast frontier topics (, ). For instance, recent bibliometric studies have illuminated research evolution in areas such as stem cells in stroke treatment (), inflammation and mitochondrial stress in stroke (). However, despite the proliferation of bibliometric reviews in neurology, no comprehensive analysis has specifically focused on the thematic evolution and research priorities of ISCD using both bibliometric and visual analytic approaches. Therefore, this bibliometric study aims to systematically delineate the research landscape of ISCD, with a particular focus on identifying key research hotspots and emerging trends. Ultimately, we seek to deepen the understanding of this disease and provide guidance for future innovation. 【Materials and methods】 【Data collection and search strategy】 To ensure the comprehensiveness of the literature data and examine potential selection bias associated with using a single database, our study performed a cross-validation analysis by supplementing the Web of Science Core Collection (WoSCC) with the PubMed database. Our search terms were derived from the Medical Subject Headings (MeSH) framework for “Ischemic Stroke” and “Cognitive Dysfunction” in PubMed. For WoSCC, we used equivalent keyword combinations to ensure consistent coverage. The complete search strategy and results are provided in , . The literature search was conducted on February 7, 2025. The search period was 2004–2024 for both databases. A comparison of the retrieval results revealed that, given our specific search terms and time frame, all core publications were indexed in WoSCC, and no unique publications were identified exclusively in PubMed. This complete overlap can be explained by the fact that WoSCC, as the “gold standard” for bibliometric analyses, applies a rigorous journal selection and curation process (), whereas high-quality literature in established research fields is typically indexed in both databases, even though PubMed has a broader but less stringent indexing policy. Given the superior metadata standardization of WoSCC for bibliometric tools, we selected WoSCC as the sole data source for subsequent analyses. To capture evolving research trends while maintaining focus on primary scientific contributions, we limited our search to document type “Article,” published between 2004 and 2024. The language was restricted to English. We then manually screened the titles and abstracts of the retrieved records against predefined inclusion and exclusion criteria. The inclusion criteria were: (i) studies explicitly addressing ISCD as a primary focus, including but not limited to investigations into its pathogenesis, neuroimaging features, risk factors, biomarkers, assessment methods, interventions, or rehabilitation outcomes; and (ii) studies involving patients with IS, or basic experimental research utilizing well-established and broadly recognized animal models of IS. Studies were excluded if they: (i) exclusively focused on general cognitive aging, Alzheimer’s disease, or other secondary cognitive impairments unrelated to IS, without clearly identifying IS as the core exposure or etiological factor; and (ii) lacked direct relevance to ISCD, such as those solely evaluating the efficacy of acute-phase thrombolytic therapy or endovascular thrombectomy without addressing cognitive outcomes. Finally, a total of 1,062 unique articles were included for subsequent analysis. illustrates the entire screening process. Literature search and screening process of this study. 【Data analysis】 The exported data were imported into CiteSpace (version 6.4.R1) for an initial removal of duplicate records, thereby ensuring a non-redundant dataset (). Subsequent analyses identified key metrics, including publication volume, authorship, institutional and national contributions, journal distribution, keyword trends, and citation patterns. In the “Selection Criteria,” the “g-index” was uniformly set to K = 5 to encompass all nodes, with default parameters of TOP N = 50 and TOP N% = 10.0%. The dataset was also analyzed using the bibliometrix package in R (version 4.4.1) and VOSviewer (version 1.6.20) to generate quantitative bibliometric indicators and visualizations (). VOSviewer was primarily employed for constructing and visualizing collaboration networks and keyword co-occurrence maps, focusing on the strength of associations between items. The R package Bibliometrix was used for descriptive statistical analyses and trend plotting. To comprehensively characterize the ISCD research landscape, we first analyzed publication and citation volume to gain an overview of productivity and development trends. We then examined country contributions and international collaboration to delineate the global cooperative network. Institutional performance was assessed to recognize major research centers and their inter-organizational partnerships. Subsequently, we conducted journal and co-cited journal analyses to outline the core publication outlets and disciplinary knowledge base. We evaluated authors and co-cited authors to pinpoint key opinion leaders and influential research groups. Furthermore, keyword co-occurrence analysis was performed to uncover research hotspots and thematic clusters. Finally, we employed reference analysis to trace the intellectual origins and knowledge evolution of the field. These analyses together formed the basis for our systematic bibliometric analysis. 【Results】 【Analysis of publication and citation trends】 illustrates the temporal evolution of ISCD-related publications and citations from 2004 to 2024. The publication count evolved through different phases (). The period from 2004 to 2019 was marked by steady though fluctuating growth, followed by a sharp rise between 2019 and 2020. Annual publications exceeded 100 in 2021 (n = 111), culminating in a peak of 133 publications in 2022. However, subsequent years (2023–2024) witnessed a notable decline in output. As shown in , the fitted curve (R2 = 0.96) indicated a strong correlation between publication year and output volume. While the curve predicted 152 and 180 publications for 2023 and 2024, respectively, the actual numbers were only 108 and 93. According to , citation numbers increased progressively each year, attaining the highest level in 2024 (n = 4,482). Trends of annual publications and citations in ISCD field from 2004 to 2024. (A) Annual number of publications and citations from 2004 to 2024. (B) The fitted curve of annual publications and the contrast between predicted and actual publications. 【Analysis of countries and collaboration】 During the analyzed timeframe, 65 countries were actively engaged in the ISCD field. To better interpret the collaborative roles of countries, we employed betweenness centrality, a metric that measures the extent to which a node serves as a bridge connecting disparate nodes within a network. A higher centrality value indicates a stronger intermediary role in linking different research groups, suggesting greater influence, while a lower value reflects a more peripheral position. As visually demonstrated in and , China dominated output with 333 publications (31.4% of the total), yet it exhibited limited integration within the global collaborative network (centrality = 0.04). The United States emerged as the second largest contributor (n = 264, 24.9% of the total), while exhibiting significantly stronger collaborative influence (centrality = 0.30). The United Kingdom ranked third with 74 papers (7.0%, 0.05). Japan occupied the tenth position, contributing 40 publications (3.8%, 0.04). Bibliometric analysis of countries in ISCD research from 2004 to 2024. (A) Top 10 most productive countries. (B) The countries of corresponding authors. (C) Collaboration network of countries based on VOSviewer. Each node corresponds to a country, where larger nodes indicate greater output in this field. The connecting lines between nodes represent collaborative relationships between countries, with thicker lines denoting stronger cooperation. (D) Overlay visualization map of countries. Countries marked by blue and purple nodes engaged in research activities earlier upon time course before 2017, whereas countries that were highly productive in ISCD around 2018 were coded with yellow color and those became prominent around or after 2020 appeared in red. Top 10 most productive countries. provides a detailed statistical breakdown of corresponding authors’ countries using Bibliometrix, distinguishing between collaboration types to evaluate independent research capacity versus international engagement. China commanded a substantial lead in both single country publications (SCP) and multiple country publications (MCP). However, SCP consistently outweighed MCP across all top-producing nations, suggesting stronger domestic research output relative to international collaboration. Additionally, we employed VOSviewer software to create ,, showing both the network visualization and overlay mapping of global cooperation in ISCD studies. For instance, China maintained very close cooperation with the USA and Australia, with notably vibrant scientific research activities in recent years. 【Analysis of institutions】 CiteSpace-based bibliometric analysis revealed a distinct concentration of institutional productivity and centrality, as illustrated in . The top 10 institutions collectively contributed 22.98% of the total publications. Capital Medical University led with 38 publications (3.58%, centrality = 0.17), followed by Harvard University (n = 30, 2.82%, centrality = 0.08), and the US Department of Veterans Affairs (n = 29, 2.73%, centrality = 0.02). Visualization of institutions in ISCD. (A) The publication counts and centrality of the top 10 institutions in ISCD. (B) Collaboration network of institutions based on VOSviewer. Each node represents an institution. Larger nodes indicate greater output. The connecting lines between nodes represent collaborative relationships between them, with thicker lines denoting stronger cooperation. To map institutional research activity and influence, we used VOSviewer to generate a co-authorship network comprising institutions with at least 10 publications. As depicted in , prominent institutions such as Capital Medical University, Harvard University, and Columbia University were represented by larger nodes, indicating their substantial publication output in this field. The network visualization revealed extensive inter-institutional collaboration patterns. For example, Harvard University, Columbia University, and Johns Hopkins University, as indicated by their dense interconnecting lines, formed a closely collaborative cluster. 【Analysis of journals and co-cited journals】 In this study, all publications came from 400 different journals. , display the top 10 journals ranked by publications and impact factor (IF) in this field. Stroke led with 47 articles, in Journal Citation Reports (JCR) Quartile 1 (Q1), IF = 7.9, followed by Journal of Stroke and Cerebrovascular Diseases (n = 42, Q3, IF = 2), and Frontiers in Neurology (n = 34, Q2, IF = 2.7), demonstrating their high activity in ISCD research. High impact factor reflects a journal’s authoritative standing in academia (). The Lancet Neurology held the highest impact factor (Q1, 46.6), though its cumulative publication count was relatively low (n = 4). Similarly, Nature Reviews Neurology (Q1, 28.2) only published one relevant article. In contrast, Stroke combined high productivity (n = 47) with academic excellence (Q1, IF = 7.9). Overall, while both publication volume and impact factor serve as key metrics for evaluating a journal’s scholarly influence, they do not always exhibit a positive correlation. Analysis of journals and co-cited journals in ISCD. (A) Top 10 most productive journals and their impact factor (IF). Bar graphs with the same color represent journals with the same journal citation reports (JCR) partition. (B) Top 10 journals ranked by impact factor and their publications. (C) Overlay visualization map of journals analysis. (D) The density map of journals. Red areas indicate high density. The volume of ISCD studies published in journals correlating positively with both the density level and the word size. (E) The network of co-cited journals from CiteSpace. Where two journals are linked if they are jointly referenced by another source, forming a co-citation relationship. Journals with centrality > 0.1 will be indicated by a purple circle. (F) The timezone map of co-cited journals. The time zones for different journals are determined by the earliest year in which each journal was first co-cited. The color rings show how citations were distributed over time, with inner rings for early years and outer rings for recent years. , illustrate the overlay visualization and density map of journals with at least 5 publications in ISCD. Among the top five high-output prolific journals, Frontiers in Aging Neuroscience was uniquely color-coded yellow, denoting both its recent publication timeframe and its ascendance as a go-to journal for cutting-edge research. Stroke and Journal of Stroke & Cerebrovascular diseases were designated as high-density areas, signifying their roles as foundational knowledge bases in this field. We used CiteSpace to complete the visualization map of co-cited journals. As presented in and , Stroke was the most co-cited journal (co-citation = 928), followed by Neurology (co-citation = 692), The Lancet Neurology (co-citation = 528), Journal of Neurology, Neurosurgery & Psychiatry (co-citation = 382), and PLOS One (co-citation = 374). High co-citation counts highlight a journal’s foundational role in advancing the discipline. In the network of co-cited journals, PLOS One had the highest centrality (0.11) (nodes represented by purple circles), followed by Neurology (0.11), and The Lancet (0.11). Furthermore, there was a notable convergence between the most productive and most co-cited journals, with Stroke, Neurology, and PLOS One featuring in the top 10 of both categories. As shown in , journals are distributed chronologically (left to right) to visualize the evolution of co-citation relationships over time. The top 10 cited journals all positioned on the leftside and had wide red rings, laying the foundation for research in this field. Top 10 co-cited journals of ISCD in terms of citation count. 【Analysis of authors and co-cited authors】 We employed bibliometric analysis to assess researcher influence in ISCD research based on both productivity and citation metrics. Our analysis identified 6,503 contributing authors during this period. Top 10 most productive authors are listed in . Mok Vincent dominated with 16 publications, closely followed by Lim Jae Sung and Yu KH (n = 15). Among them, Mok Vincent had the highest centrality (0.05), while these values were all comparatively low. The h-index is a composite metric measuring both research productivity and impact. An h-index of N means that the author has N publications, each of which has received at least N citations. Among them, Sacco RL had the highest h-index (h = 134), reflecting a robust quantity-quality balance in his publication record. Top 10 productive authors referring to published articles. Co-authorship network analysis offers a robust approach to highlight primary authors and map the density of their collaborative networks (). We applied VOSviewer to analyze the co-authorship network (authors with at least 5 publications and 50 citations), ensuring the focus on high-impact researchers (). In the map, the connecting lines between nodes represent collaborative relationships among authors, with line density and thickness indicating the strength of their collaborations. The author timezone map () tracks the dynamic participation of authors throughout the studied timeframe. Some productive authors such as Lim Jae Sung, Yu KH, and Oh Mi Sun began publishing articles in 2021. Visualization of authors and co-cited authors in ISCD. (A) Author’s co-authorship netwo