NIH Launches AI Tool to Link Biomedical Research
The National Institutes of Health (NIH) has introduced an experimental research tool designed to help scientists better understand how individual biomedical studies connect with the broader body of scientific evidence.
Called Linked Discoveries, the new platform was launched on September 24 by the NIH’s National Library of Medicine (NLM). It allows researchers to start with a publication indexed in PubMed and explore other studies that are closely connected to it through citations and other relationships.
NIH Director Jay Bhattacharya said the rapid expansion of biomedical research has created a growing challenge for scientists trying to understand how individual findings fit into the wider scientific landscape. According to Bhattacharya, researchers need more than keyword-based searches to identify meaningful relationships between studies and follow how scientific ideas develop over time.
More than 29 million PubMed publications are available through Linked Discoveries at launch. The platform creates a research “neighborhood” around a selected publication, allowing users to investigate related papers, including research that seeks to reproduce earlier findings.
The tool is part of a broader NIH initiative focused on improving replication and reproducibility across research supported by the agency. NIH noted that scientists can struggle to evaluate the significance of a particular finding when relevant studies and supporting information are distributed across multiple resources.
Linked Discoveries uses an AI-informed system to identify connections between publications. Researchers can view these relationships through graphical and timeline-based displays, making it possible to follow how research develops and how individual studies relate to earlier or subsequent work.
The platform also offers filters based on information involving conditions, genes and chemicals. In addition, users can identify relationships involving citations, review articles, retractions and publications connected to NIH funding.
NLM Director Peter Embi said the platform is intended to move researchers beyond viewing scientific literature as a fixed collection of information. By examining the surrounding network of studies, researchers can gain additional context about work published before and after a particular finding.
However, NIH emphasized that Linked Discoveries does not determine whether an individual study is scientifically reliable or establish whether a finding has been successfully replicated. Instead, the platform organizes relationships among publications and provides researchers with additional information they can use when evaluating evidence themselves.
The September 24 launch represents the tool’s first public version. NLM plans to continue testing and improving Linked Discoveries, with future development informed by feedback submitted by users through the platform.
