Global trends and hotspots in artificial intelligence for high myopia: a bibliometric analysis

Abstract

Purpose: This study aims to conduct a bibliometric analysis of global publications on the application of artificial intelligence (AI) in high myopia (HM).

Methods: We retrieved publications on AI in HM from the Web of Science Core Collection (WoSCC) database, MEDLINE and Chinese Science Citation Database (CSCD) with data up to 2024. The analysis focused on publication and citation trends, identifying key articles, influential countries, institutions, authors, and journals. Additionally, we explored research domains and emerging keywords.

Results: A total of 167 relevant publications were included. The first AI-related paper on HM was published in 2017, with a significant surge in 2021, followed by a consistent increase in publication and citation counts over the next 3 years. China emerged as the most productive country, with the most extensive international collaboration. East Asian authors dominated the top 10 most influential authors. Yang, Weihua and Investigative Ophthalmology & Visual Science (IOVS) contributed the most publications among authors and institutions, respectively. Keyword analysis revealed that retinal imaging-related terms remained a consistent research focus, while newly emerging keywords included "automated detection" and "childhood."

Conclusion: Recent advancements in AI applications for HM have been significant and are expected to continue. Future research will likely focus on multimodal imaging and improving algorithm accessibility. Our findings offered the first comprehensive overview of global research on AI in HM, thus providing valuable insights for researchers to understand the current status and future trends in this field.

Keywords: artificial intelligence; bibliometric analysis; data visualization; global research; high myopia.

Figure 1

Detailed flowchart of this study.

Figure 2

Trends in the number of AI-related publications and citations in HM.

Figure 3

Publication trends and prediction curve of global and leading countries in AI-related HM research. This figure uses full counting, meaning each country listed in a paper contributes a weight of 1.

Figure 4

Co-authorship network visualization map of countries and regions. Each node represents a country or region, with the circle size reflecting the number of publications. Connecting lines represent collaboration between countries and regions.

Figure 5

Chord diagram of co-authorship among the top institutions. Node data are arranged radially along the circumference, with weighted arcs (indicating collaboration strength) connecting the nodes.

Figure 6

Three-field plot analysis displaying the journal preferences of authors from different countries. The three fields represent (1) AU_CO: Country, (2) SO: Source Journal, and (3) AU: Authors. The width of the nodes indicates the number of publications, while the width of the connecting lines reflects the level of collaboration.

Figure 7

(A) Word cloud of the most frequent keywords from 2017 to 2022. (B) Word cloud of the most frequent keywords from 2023 to 2024.

Figure 8

Burst map of Title-Abstract-Keywords. The numerical columns display the relative burst strength, start, and end times. In the line graph, the red portion represents the burst duration for each time series.