Bibliometric and literature data mining analysis of art therapy research indexed in PubMed using medpulse

Previous bibliometric studies on art therapy have mainly relied on Web of Science, Scopus, ScienceDirect, or topic-specific subsets. In contrast, PubMed-based data-driven mapping and literature mining of general art therapy research remain limited. Using the in silico tool Medpulse, the present study performs a PubMed-based bibliometric and literature data-mining analysis of art therapy research, while further evaluating how retrieval-field selection (topic vs title) affects result relevance and query precision. Topic-based retrieval of “art therapy” in Medpulse showed limited relevance precision: among the top 10 most-cited records and top 10 highest-impact-factor records, only 4 and 8 articles, respectively, were directly related to art therapy, whereas title-based retrieval achieved 100% and 90% relevance in the above sets. Then, we used Medpulse to conduct a title-based bibliometric analysis of art therapy research. We retrieved 479 publications from 204 journals published between 2010 and 2025, involving 1689 authors. The United States was the leading contributor in publication output. It also occupied a central position in the international collaboration network, with visible cross-country collaboration links involving major contributing countries such as China. The key contributors to this research area included Drexel University (in United States), University of Haifa (in Israel), and the University of Sheffield (in United Kingdom). The word cloud suggest that art therapy research centers on depression, anxiety, and mental health. Beyond conventional bibliometric description, this study adopted a literature data-mining perspective using Medpulse to structurally extract and analyze PubMed bibliographic metadata. This approach enabled the mapping of publication patterns, the evaluation of retrieval precision differences, and the identification of structured knowledge patterns in art therapy research.