Epistemic integration and social segregation of AI in neuroscience

TL;DR

This paper investigates how AI diffuses into neuroscience and whether it leads to epistemic integration or social segregation. It analyzes a Microsoft Academic Graph dataset covering $1970$–$2019$, constructing a temporal egocentric citation network, a journal-level activity map, and a time-aggregated co-authorship network to compare AI-related neuroscience papers ($\mathcal{P}\cap AI$) with non-AI neuroscience papers ($\mathcal{P}\cap \overline{AI}$). By computing an AI-involvement score for authors ($f_{AI}$) and partitioning them into quartiles ($\overline{Q}, Q_0, Q_1, Q_2$), the study reveals a diffusion that creates a dedicated AI ecosystem within neuroscience, with AI researchers forming a socially segregated subspace and publishing in a small set of AI-active journals. The results indicate partial diffusion consistent with the transverse science view: AI acts as a research-technology that diffuses selectively, reinforcing specialized epistemic practices rather than fully reconfiguring core neuroscience. These findings have implications for understanding interdisciplinarity, diffusion of AI tools, and science-policy design aimed at broadening AI's cross-disciplinary impact.

Abstract

In recent years, Artificial Intelligence (AI) shows a spectacular ability of insertion inside a variety of disciplines which use it for scientific advancements and which sometimes improve it for their conceptual and methodological needs. According to the transverse science framework originally conceived by Shinn and Joerges, AI can be seen as an instrument which is progressively acquiring a universal character through its diffusion across science. In this paper we address empirically one aspect of this diffusion, namely the penetration of AI into a specific field of research. Taking neuroscience as a case study, we conduct a scientometric analysis of the development of AI in this field. We especially study the temporal egocentric citation network around the articles included in this literature, their represented journals and their authors linked together by a temporal collaboration network. We find that AI is driving the constitution of a particular disciplinary ecosystem in neuroscience which is distinct from other subfields, and which is gathering atypical scientific profiles who are coming from neuroscience or outside it. Moreover we observe that this AI community in neuroscience is socially confined in a specific subspace of the neuroscience collaboration network, which also publishes in a small set of dedicated journals that are mostly active in AI research. According to these results, the diffusion of AI in a discipline such as neuroscience didn't really challenge its disciplinary orientations but rather induced the constitution of a dedicated socio-cognitive environment inside this field.

Figures (16)

• Figure 1: Left: Classification of the papers in the extracted corpus. The intersection between the blue and red zones corresponds to the set $\mathcal{P}\cap AI$ in the main text, while the blue zone excluded from this intersection represents the set $\mathcal{P}\cap\overline{AI}$. The green set $\overline{\mathcal{P}}$ includes all papers added after the building of the ego-centered citation network that are not published in neuroscience journal. Right: Cumulative number $N_c$ of publications in the two main subsets considered in this paper. The inset shows the instantaneous part of AI-related publications in the whole neuroscience corpus.
• Figure 2: Citation ecosystem centered here around three papers (grey) published at year $y_0$ and that are included either in $\mathcal{P}\cap AI$ or $\mathcal{P}\cap\overline{AI}$. One dashed arrow represents the citation of a target paper by a source one. Hence red points are the papers that are cited by the papers of our corpus (reference) while the blue ones are citing them (impact). The rankings are shown in decreasing order with the associated number of citations of each JSC $d_i$.
• Figure 3: Coordinate system to characterize the distribution of disciplines in the references and citations of the corpora $\mathcal{P}\cap \overline{AI}$ and $\mathcal{P}\cap AI$.
• Figure 4: Distribution of the AI score $f_{AI}$ of the authors.
• Figure 5: Instantaneous similarity between the references' or citations' (also called impact) rankings of the two corpora $\mathcal{P}\cap AI$ and $\mathcal{P}\cap\overline{AI}$ at year $t$.