This study was searched based on the Web of Science Core Collection (WoSCC). The database features a strict journal selection mechanism and a standardized citation data system, which effectively ensures the homogeneity of bibliometric indicators and makes it the preferred choice for bibliometric research (Zhang et al., 2021; Yang J. et al., 2022; Veiga-del-Baño et al., 2023). Due to some delay in database inclusion, this may lead to systematic bias. To ensure that updates to the database have stabilized, a certain amount of time needs to be allowed for the deadline (December 31, 2024) set by this study for literature inclusion (Lv et al., 2025). Therefore, publications on PI3K/Akt in IS were searched and screened in the WoSCC database on February 1, 2025, for this study. Also excluded due to the small number of publications in this area before January 1, 2010. The search formula used was as follows: TS = (“PI3K/AKT” OR “PI3K/Akt” OR “PI3K-Akt” OR “Phosphoinositide-3 Kinase/Akt” OR “Phosphatidylinositol 3-Kinase/Akt” OR “Phosphoinositide-3 Kinase/protein kinase B” OR “Phosphatidylinositol 3-Kinase/protein kinase B” OR “Phosphoinositide-3 Kinase/serine/threonine kinase” OR “Phosphatidylinositol 3-Kinase/serine/threonine kinase” OR “Phosphatidylinositol 3-Kinase/serine/threonine protein kinase” OR “Phosphatidylinositol-3-hydroxykinase/protein kinase B”) AND TS = (“ischemic stroke” OR “Ischemic Strokes” OR “Stroke, Ischemic” OR “Acute Ischemic Stroke” OR “Acute Ischemic Strokes” OR “cryptogenic ischemic stroke” OR “cryptogenic stroke” OR “cryptogenic embolism stroke” OR “wake up stroke” OR “wake-up stroke” OR “brain ischemia” OR “cerebral ischemia” OR “cerebral stem ischemia”). Based on the given timeframe and search terms, 958 relevant publications on the PI3K/Akt in IS were retrieved from the WOSCC database. Subsequently, we screened the 958 publications based on language type, publication type, and relevance to the search topic. Finally, 635 publications were identified for inclusion in this study’s bibliometric analysis. The complete records of the retrieved publications were exported in “txt” format for subsequent analyses.

In this study, the names of countries/regions and keywords were identified through VOSviewer software. The data were cleaned up using the merge and modify functions of the “thesaurus file.” For the names of countries/regions: (1) Replace the names of countries with internationally recognized names, e.g., replace “Peoples R China” with “China.” (2) The names of regions should be harmonized with the name of the country to which they belong, e.g., “Scotland” and “England” should be unified as “United Kingdom.” For keywords: (1) Combine synonyms according to the MeSH Thesaurus, e.g., “acute ischemic stroke” and “cerebral ischemic stroke” into “ischemic stroke.” (2) Merge of keywords with the same meaning but different forms of expression, e.g., “PI3K/Akt signaling” and “PI3K/Akt pathway” to PI3K/Akt; (3) The full name of the keyword is changed to an academically recognized abbreviation, e.g., “blood-brain barrier” is merged with “blood-brain-barrier” to become “BBB.” (4) Unification of plural forms of nouns into singular forms, e.g., “astrocytes” into “astrocyte.” (Supplementary file S1 lists the records of keyword merging and modification).

The main bibliometric analysis and data visualization tools used were VOSviewer software (version 1.6.19) and CiteSpace software (6.4.R1 advanced version). VOSviewer was used to construct a visual knowledge graph encompassing country, organization, author, co-citation author, journal, co-citation journal, and keyword co-occurrence network. Pajek (version 6.01) and Scimago Graphica (version 1.0.48) were used to refine clustering views for more intuitive visualization. CiteSpace primarily uses time-slicing techniques to construct a time series of network models, which evolve and integrate into a comprehensive overview network to analyze relevant literature systematically (Jiang et al., 2024). We used CiteSpace to generate a visual knowledge graph of journal biplot overlays and co-cited references. Meanwhile, we generated a cluster timeline by clustering the titles of the co-cited references. Additionally, we used CiteSpace to perform “burst detection” on keywords and references. To further characterize the data, we used Microsoft Office 2021 to analyze the exported data from VOSviewer and CiteSpace and construct a table of relevant analyses. A total of 635 articles on the PI3K/Akt in IS were analyzed, far exceeding the minimum recommended sample size of 200 for textual metrics analysis. Accordingly, the feasibility of this study was fully guaranteed. Figure 1 illustrates the flowchart for screening and analyzing the papers.

The search and screening of relevant literature from 2010 to 2024 yielded 635 publications, including 594 original articles (93.54%) and 41 reviews (6.46%). The number of publications was 16 in 2010, peaking at 71 in 2021. Although a slight decline occurred in 2022 (60 publications) and 2023 (62 publications), the number of publications rebounded to 71 in 2024, indicating sustained research interest. Figure 2 presents the annual publication trend of PI3K/Akt studies in IS. Overall, the number of published studies on PI3K/Akt in IS is increasing, with more relevant studies expected to be published in the future.

Between 2010 and 2024, 38 countries published research on the PI3K/Akt in IS. The network of national collaborations was generated using VOSviewer, and Scimago Graphica was used to generate a map of the world’s geographical distribution of the research results. Figure 3a illustrates the distribution of countries contributing to this field on a global scale. Table 1 and Figure 3a present the top 10 most productive countries in this field, with China (533 papers), the United States (54 papers), and South Korea (26 papers) contributing the most. Although Japan published nine papers, none involved collaboration, suggesting that they need to be more actively involved in international collaboration to promote further development in this field. Citation analysis revealed that the United States has the highest average number of citations per paper (43.20), followed by the United Kingdom (39.17) and Australia (39.14) (Table 1). Figure 3c presents the collaboration among the top 10 countries in terms of published literature. Each node represents a country, with node size corresponding to publication volume and the thickness of the connecting lines reflecting the strength of collaboration. Figures 3b,c depict a wide range of cooperation among different countries, with China and the United States exhibiting the closest cooperation, indicating their crucial role in international cooperation.

VOSviewer analysis identified 746 institutions contributing to the 635 papers. Table 2 lists the top 10 institutions in terms of the number of papers, all of which are based in China. Figure 3d depicts the network of institutional collaborations, revealing a dense network of collaborations primarily among Chinese institutions. These findings suggest that China plays a central role in PI3K/Akt research in IS. Strengthening collaborations with Chinese institutions could accelerate progress in this field. Figure 3d indicates that the widest distribution of links occurs in the yellow and green marked years, indicating that inter-institutional collaboration was most frequent between 2018 and 2020 and has continued beyond this period. Citation analysis of institutions (Table 2) reveals that Fudan University has the highest number of citations (557), followed by Capital Medical University (552), and Hebei Medical University (487).

To construct a visual network map of authors and co-cited authors in PI3K/Akt research in IS, we used the VOSviewer visualization software and adjusted the clustering view using Pajek to make it more intuitive and clearer. The field comprises 3,885 authors. Table 3 lists the top 10 most prolific authors, with Wang Lei leading with 8 papers, followed by Wang Na and Wang Ning with 6 papers each. Chen Jing, Ding Yi, Li Li, Li Yunman, Liu Fang, Liu Li, and Liu Nan have each published five papers. Price’s law states that the minimum number of publications by core authors in a given field, m, is proportional to the square root of the number of publications by the most prolific authors, n_max which is given by m = 0.749 × nm⁢a⁢x (Xu et al., 2024). Core authors in the field have at least 3 publications each (rounded upwards). Figure 4a illustrates the collaborative network of the 142 core authors, and it is evident that there are weak collaborative relationships between these core authors. This suggests that more cooperation and communication are required to expedite progress in the field. In terms of authors’ influence, Wang Lei has the highest number of citations at 389, while Liu Li has the highest average number of citations at 58 (Table 3), suggesting that their research findings have a significant international influence. Figure 4b demonstrates the network mapping of the co-cited authors, revealing three clusters. Among the 18,627 co-cited authors, 87 have been cited at least 20 times, with EZ Longa being the most cited (127 citations). Longa developed the middle cerebral artery occlusion model (MCAO) in 1989, a small-animal model of cerebral ischemia (Longa et al., 1989), which remains widely used worldwide.

A total of 231 journals published papers on the PI3K/Akt in IS, with 36 journals publishing at least 5 papers (Figure 5a). Table 4 depicts the top 10 journals in terms of publications. The top 10 journals published 164 articles in the field, accounting for 25.83% of the total number of publications. Brain Research (22 papers), Journal of Ethnopharmacology (22 papers), and Frontiers in Pharmacology (19 papers) were the most prolific. Among the top 10 journals, the impact factor (IF) ranged from 2.70 to 6.90. Six journals, including the Journal of Ethnopharmacology, Frontiers in Pharmacology, Biomedicine & Pharmacotherapy, Molecular Neurobiology, PLoS One, and International Journal of Molecular Sciences, were ranked in the Q1 division based on the 2023 Journal Citation Reports (JCR). As illustrated in Figure 5b, 170 of the 3,243 co-cited journals have at least 40 citations, with different clusters represented by distinct colors. Table 5 details the top 10 most cited journals. Of these, Stroke tops the list with 7,925 citations and is also the journal with the highest IF (7.8 in 2023). Neurorehabilitation and Neural Repair (3,263 citations) and Archives of Physical Medicine and Rehabilitation (3,057 citations) follow closely.

Figure 5c depicts a dual-map overlay of journals. The red dashed box on the left represents citing journals (knowledge frontiers), and the blue dashed box on the right represents cited journals (foundational areas of knowledge). The interactive colored citation links present the strength of the association between the topics of these journals. The most prominent yellow citation path extends from “Molecular, Biology, and Genetics” (knowledge base) to “Molecular, Biology, and Immunology” (knowledge frontiers). This suggests that research on the PI3K/Akt in IS has focused on genetic regulation and immune response and has resulted in a relatively dense knowledge output. However, translational research, including target drug development and drug delivery, remains limited, indicating a gap between basic research findings and clinical applications in this field.

We employed VOSviewer software to conduct a citation analysis of the 635 papers included in the study. Our objective was to identify high-impact papers in the field of PI3K/Akt research in IS. This will facilitate our understanding of the research hotspots and trends in this field (Ravindran et al., 2019). Table 6 presents the top 10 cited papers in the field. Hou et al. (2018) published in Genes & Diseases [Q1 (6.9)], garnered the highest citation count, with 199 citations recorded. Its research focuses on the mechanisms of apoptosis in this field. Fu et al. (2022) published in the Journal of Ethnopharmacology [Q1 (4.8)], despite its late publication (2022), has received 162 citations and is in second place. This suggests that ferroptosis, a regulatory mechanism, is an emerging trend in the field of PI3K/Akt research in IS in recent years. Huang LF (Huang et al., 2018) published in the Journal of Molecular Neuroscience [Q3 (2.8)] was ranked third with 161 citations. Its study is indicative of the elevated level of interest and activity in autophagy research within this field. A significant finding was that six of the 10 highly cited papers concentrated on the active ingredients of traditional Chinese medicine (TCM) and the PI3K/AKT pathway, covering a span of 13 years (2010–2022) (Liu et al., 2010; Yao et al., 2012; Hou et al., 2018; Huang et al., 2018; Xie Y. et al., 2018; Fu et al., 2022). This finding suggests that this line of research not only continues to receive attention but also shows great potential for translation from basic research to clinical applications.

Keywords in scientific research papers reflect the topics of study. Analyzing keyword frequency and co-occurrence helps identify popular topics and emerging trends in a research field (van Eck and Waltman, 2010). We used VOSviewer to visualize keyword co-occurrences in PI3K/Akt research in IS and Pajek to adjust the clustering view. A total of 2,624 keywords were identified over the 15 years from 2010 to 2024, with 198 keywords co-occurring at least 5 times. Table 7 lists the top 10 most frequent keywords, including apoptosis (237), P13K/Akt (145), reperfusion injury (138), activation (135), oxidative stress (132), and neuroprotection (125). Figure 6a presents the co-occurrence network mapping of keywords in this domain. The keywords with strong relevance are grouped into seven colored clusters, all of which contain more than 15 keywords. The red cluster (41 keywords) is the largest and mainly focuses on PI3K/Akt activation to produce neuroprotection after cerebral ischemia/reperfusion injury, including terms brain ischemia, PI3K/Akt, activation and neuroprotection. The green keyword cluster (35 keywords) focuses on the neurorestorative effects of the PI3K/Akt pathway on stroke rats and in-vitro models. The keywords mainly include stroke, in-vitro, neurogenesis, angiogenesis, and rat. The blue cluster (30 keywords) focuses on the impact of oxidative stress and mitochondrial dysfunction on neuronal survival in this field, mainly involving the terms oxidative stress, focal cerebral ischemia, mitochondrial dysfunction, and cell-survival. The yellow cluster (27 keywords) focuses on neuronal apoptosis and expression in this field. The keywords cover apoptosis, expression, and caspase-3. The purple cluster (26 keywords) focuses on reperfusion injury and inflammation in this field and contains mainly the terms cerebral reperfusion injury, inflammation, and microglia. The cyan cluster (23 keywords) focuses on bioinformatics approaches to shed light on the protective mechanisms of the PI3K/Akt signaling pathway. The main terms involved are ischemic stroke, pathway, network pharmacology, and protects. The orange cluster (16 keywords) focuses mainly on autophagy mechanisms in reperfusion injury in this field. Terms such as autophagy, PI3k, Akt, mTOR. These results indicate research hotspots in this field.

The temporal mapping of keyword networks was generated using VOSviewer based on the average year of first keyword occurrence, and the clustering layout was optimally adjusted using Pajek (Figure 6b). The figure depicts that since 2020, ischemic stroke, cerebral reperfusion injury, autophagy, inflammation, microglia, ferroptosis, network pharmacology, and traditional Chinese medicine (TCM) have been the frequently appearing keywords. This suggests that autophagy and neuroinflammation are not only current research hotspots but also future trends in PI3K/Akt studies in IS. Notably, the role of ferroptosis in this field is gradually expanding, indicating the potential for more relevant studies. Additionally, targeting the PI3K/Akt with TCM for IS treatment is expected to become a key direction for translating basic research findings into clinical applications. We generated the top 10 keywords in terms of burst strength using CiteSpace. Figure 6c depicts that keywords such as cyber pharmacology (burst strength of 8.24; 2021–2024) and molecular docking (burst strength of 4.97; 2022–2024) exhibited strong burst trends, which will probably continue in the future. This indicates that using bioinformatics approaches to investigate the targeting of PI3K/Akt by active ingredients of TCM and their application in IS treatment has become an important direction for future research in this field.

Analysis of co-cited references helps to identify the knowledge base and influential articles within a large body of literature, enabling a comprehensive exploration of the field (Zhu et al., 2023). In total, 26,555 references were cited across 635 articles on the PI3K/Akt in IS. Using CiteSpace software, we mapped the co-citation network of these references (Figure 7a). Table 8 details the top 10 co-cited references based on high frequency and centrality, comprising 7 original articles and 3 reviews. It is noteworthy that Hou et al. (2018) attained the top position in terms of co-citation frequency (15) and centrality (0.27). Chen et al. (2019) was the runner-up, with a centrality of 0.25. These studies represent the core nodes in Figure 7a, which establish connections between diverse research directions and provide significant theoretical support and a bridging role for subsequent studies. Feng et al. (2020) exhibited the second-highest co-citation frequency (14), though its centrality (0.01) was minimal. In a manner analogous to Hou et al. (2018), Feng et al. (2020) concentrated on the neuroprotective effects of TCM by targeting the PI3K/Akt pathway. Chen et al. (2020) (centrality 0.11, ranked third) employed a research strategy that integrated proteomics analysis techniques with TCM.

To further investigate research hotspots, we used CiteSpace software to generate 22 clusters. The modularity value and weighted average profile value of the clusters were 0.8779 and 0.9137, respectively, indicating high clustering confidence. Figure 7b presents the distribution of the top 10 clusters. Similarly, in Figure 7c, the horizontal position of the nodes on the time axis represents when the references were cited, while node size represents the frequency of citation. Recent research hotspots in the PI3K/Akt in IS include cluster 0# (autophagy), cluster 2# (microglial response), cluster 3# (neuronal apoptosis), cluster 6# (Chinese herbal medicine), and cluster 9# (molecular docking technology).

A citation burst indicates a significant increase in attention to a specific article in a specific research area within a certain time frame (Xu et al., 2024). References with high values in the strength column are usually milestones in scientific cartographic research (Chen, 2004). Figure 7d presents the top 15 references with the strongest citation bursts. The earliest and longest-spanning reference is “ DOI 10.1161/STROKEAHA.108.53163” (Broughton et al., 2009), with a citation burst concentrated between 2010 and 2014, focusing on apoptosis after stroke. The reference with the highest citation burst strength (10.36) is “ DOI 10.3389/fnana.2015.0004” (Li et al., 2015), which focuses on neuroapoptosis by targeting the PI3K/Akt to provide neuroprotection against stroke. Notably, the reference “ DOI 10.1186/s12974-019-1516-2” maintained a citation peak until 2024. It focuses on post-acute stroke neurological inflammatory mechanisms (Jayaraj et al., 2019). This suggests that inflammation is a current research focus in PI3K/Akt in IS and indicates future research directions.

This study used bibliometrics to analyze 635 original articles and reviews on PI3K/Akt in IS from 2010 to 2024. The number of publications increased steadily each year, peaking in 2021 with 71 publications. Among the 38 countries analyzed, China ranked first with 533 publications, followed by the United States and South Korea. China and the United States exhibited the closest collaboration and made significant contributions to the field. The top 10 research institutions, including Capital Medical University, Chongqing Medical University, and Fudan University, are all based in China, and a relatively close network of cooperation has been established between these institutions. However, further international institutional collaboration is needed to advance the field. Analysis of authors and co-cited authors identified 3,885 authors and 18,627 co-cited authors. Wang Lei was the most prolific author, publishing 8 papers, while Liu Li had the highest average number of citations at 57.80. Among the co-cited authors, EZ Longa had the highest number of citations at 127. This fully reflects their influence and Contribution To The Field.

The 635 articles analyzed were from 231 journals. The top three journals in terms of publication volume were Brain Research, Journal of Ethnopharmacology, and Frontiers in Pharmacology, collectively accounting for 25.83% of all published articles, indicating their significant contributions to the field. Stroke was the most frequently cited journal, with 7,925 citations, highlighting its authority in PI3K/Akt research in IS. However, the dual-map overlay of journals revealed a gap between basic research findings and clinical applications.

We identified several major hotspots for PI3K/Akt research in IS by visualizing and analyzing highly cited papers, keywords, and co-cited literature. These hotspots mainly focus on the pathological mechanisms and targeted therapies of IS.

By analyzing keyword co-occurrence and reference clustering, the trend of PI3K/Akt pathway research in IS was identified. Autophagy, neuroinflammation, and microglia are all current research hotspots and crucial directions for future research. A close and complex relationship has been observed between oxidative stress, apoptosis, and neuroinflammation. Consequently, subsequent investigations on PI3K/Akt in IS may facilitate a more profound examination of the mechanisms through which oxidative stress instigates apoptosis and neuroinflammation. Notably, ferroptosis (iron-dependent programmed apoptosis) is gaining attention in IS research related to the PI3K/Akt. Inhibiting ferroptosis in IS reduces cellular autophagy, promotes microglia polarization toward the M2 phenotype, alleviates neurological deficits, and improves cognition. These mechanisms are closely linked to PI3K/Akt pathway activation (Fu et al., 2022; Zhao et al., 2023; Zhang Y. et al., 2024). As a result, further studies on the role of ferroptosis in IS pathogenesis and its regulation via the PI3K/Akt could provide insights into therapeutic strategies against IS. Additionally, targeting the PI3K/Akt for IS treatment has been extensively supported by bioinformatics analyses, reinforcing the rationale for using TCM. Consequently, this direction is expected to become an important trend in translating basic research findings into clinical applications.

Through an in-depth analysis of the salient topics and emergent trends in this domain, we offer a concise depiction in Figure 8 of the dynamic correlation between the research focal points and the emerging regulatory mechanisms of PI3K/Akt in IS. This contributes to an in-depth understanding of the pathomechanisms of IS and provides important guidance for the development of targeted therapeutic strategies.

It is imperative to acknowledge the limitations inherent in this study. During the literature screening process, the study’s scope was constrained to the WoSCC database. This decision was influenced by the recognition that prevailing software, such as VOSviewer and CiteSpace, encounter difficulties in accommodating data standardization and compatibility issues when conducting cross-database analysis. Nevertheless, this methodological approach may potentially result in the exclusion of significant research findings within this field. Concurrently, the most recent research results from January 1, 2025, to the retrieval date were not incorporated into the analysis, which could have limited the comprehensive grasp of the cutting-edge developments in IS-related PI3K/Akt research. At the technical analysis level, CiteSpace software was limited in authorship identification flaws, including its inability to distinguish between first and corresponding authors and differentiate between authors with the same name from different institutions. Furthermore, its clustering analysis was susceptible to data loss and result bias owing to the lack of parameter standardization.

These limitations may be overcome in the future through the implementation of artificial intelligence technologies. To minimize the probability of overlooking pertinent literature, it is imperative to devise intelligent cross-library standardization instruments that seamlessly integrate disparate data formats from numerous sources. When employed in conjunction with ORCID and other unique identifiers, it establishes an accurate author identification system and effectively addresses the ambiguity problem posed by scholars with identical names. The utilization of machine learning algorithms facilitates the dynamic optimization of data cleaning and parameter standardization processes. This approach is designed to mitigate the impact of data noise and systematic bias on cluster analysis. These measures are expected to enhance the depth and reliability of bibliometric results on a comprehensive scale.