Prestigious but less interdisciplinary: a network analysis on top-rated journals in medicine

TL;DR

The study addresses whether prestige in medical journals correlates with the dissemination of interdisciplinary knowledge. It builds temporal, topic-level networks from PubMed MeSH co-occurrences for impactful versus non-impactful journals across 1999, 2010, and 2022, using cosine-normalized weights $w_{ij}=c_{ij}/\sqrt{c_{ii}c_{jj}}$ and Louvain-based community detection. The findings show that prestigious journals are generally less interdisciplinary than their non-impactful counterparts, with cancer research acting as a principal bridge between topics; differences between journal types are modest and tend to co-occur in specific areas such as infections, nervous system diseases, and cancer. The work suggests that prestige could foster interdisciplinarity by easing disciplinary boundaries, and provides a framework for comparing interdisciplinarity across journals and time. Overall, the paper offers a novel network-based lens on how journal prestige shapes knowledge integration in medicine, with implications for research policy and scholarly communication.

Abstract

Interdisciplinary research, a process of knowledge integration, is vital for scientific advancements. It remains unclear whether prestigious journals that are highly impactful lead in disseminating interdisciplinary knowledge. In this paper, by constructing topic-level correlation networks based on publications, we evaluated the interdisciplinarity of more and less prestigious journals in medicine. We found research from prestigious medical journals tends to be less interdisciplinary than research from other medical journals. We also established that cancer-related research is the main driver of interdisciplinarity in medical science. Our results indicate a weak tendency for differences in topic correlations between more and less prestigious journals to be co-located. Accordingly, we identified that interdisciplinarity in prestigious journals mainly differs from interdisciplinarity in other journals in areas such as infections, nervous system diseases and cancer. Overall, our results suggest that interdisciplinarity in science could benefit from prestigious journals easing rigid disciplinary boundaries.

Figures (4)

• Figure 1: Impactful network in 1999. Links with strength less than $0.080$ were filtered out for better visualisation. Communities were detected based on the Louvain algorithm blondelFastUnfoldingCommunities2008. We find a very similar community breakdown for the non-impactful network in 1999 (not shown).
• Figure 2: Link Strength Distribution for (a) I, NI, and NI-June networks, and (b) I and NI-June networks. Logarithmic binning was applied on x-axis with 30 bins. A power law distribution with exponent $\alpha=1.9$ and minimum value of link strength $x_{min}=10^{-1.5}$
• Figure 3: Node strength distribution of the I and NI networks. An exponential distribution with decay rate $\lambda=0.6$ with node strength $s\in[1,6]$ is plotted as the dotted line against the distributions.
• Figure 4: The relationship between link strength in Diff networks (x-axis) and neighbouring link strength (y-axis) in 1999 (left), 2010 (middle), and 2022 (right). The red line is a fitted linear regression line, with slopes of $0.08$ in 1999, $0.03$ in 2010 and $0.03$ in 2022, and $R^2$ of $0.007$ in 1999, $0.001$ in 2010 and $0.001$ in 2022. All three slopes reported are statistically significant ($p<2^{-16}$). Blue points represent the averages over bins, with error bars representing one standard error.