Originated versus imported knowledge: Revisiting the relationship between novelty and academic impact

Knowledge innovation is considered one key driver of scientific impact. However, current studies overlook how cross-field knowledge flows shape novel knowledge creation, and our understanding of its creation patterns and relationship to scientific impact remains incomplete. By introducing a broader scientific context of over 30 million biomedical publications, this study uses biomedical entities and entity combinations identified through automated recognition as proxies for knowledge and quantifies paper impact by citation counts, using Alzheimer’s disease as an empirical case. We classify novel knowledge into either originated or imported based on whether the original source is the focal field, and subsequently analyze its distribution patterns and relationships with impact. The results show that 85.82% of novel entities and 55.84% of novel combinations are imported, indicating the dominance of imported knowledge. Regardless of knowledge sources, higher novel entity proportions are associated with reduced paper impact, and novel combination proportions show an inverted U-shaped relationship with impact, but the source affects relationship intensities. It is also evidenced that originated knowledge is more strongly associated with internal impact, while imported knowledge is more strongly associated with external impact. This study provides new perspectives and more comprehensive observations on the role of novel knowledge in shaping scientific impact.