New scientific ideas drive progress, yet measuring scientific novelty remains challenging. We use natural language processing to detect the origin and impact of new ideas in scientific publications. To validate our methods, we analyze No-bel Prize-winning papers, which likely pioneered impactful new ideas, and lit-erature review papers, which typically consolidate existing knowledge. We also show that novel papers have more intellectual neighbors published after them, indicating they are ahead of their intellectual peers. Finally, papers introducing new ideas, particularly those with greater follow-on reuse, attract more citations.

The paper contributes to the debate on top performance in research productivity, its persistence over time, and the impact of gender. It uses a panel data set comprising the publications of all biomedical and exact scientists at the University of Leuven in the period 1992 to 2001. We find that women have a significant lower probability of reaching top performance for the first time in their career, particularly for top performance measured through citations, but there is no evidence for a gender bias hindering repeated top performance. On the contrary, women seem to persist in top performance more easily than men do.