P330 Quantitative Analysis of Artificial Intelligence Integration in Neuroscience
1Cate School, 1960 Cate Mesa Road, Carpinteria, CA, USA 2Department of Computer Science, Missouri State University, Springfield, MO, USA
*Email: trojancz@hotmail.com
Introduction
This study aimed to quantitatively assess the integration of artificial intelligence (AI) into neuroscience. By analyzing ~50,000 sample papers from the OpenAlex database[1], this study captured the breadth of AI applications across diverse disciplines of neuroscience and gauged emerging trends in research.
Methods A dual-query strategy was applied. One query targeted neuroscience papers (2001-2022) mentioning AI‐related terms (Figure 1), while a control query used “neuroscience.” An automated classification pipeline, built on a prompted GPT‑4o model[2], dynamically processed titles and abstracts, and classified the papers into 6 categories: Behavioral Neuroscience, Cognitive Neuroscience, Computational Neuroscience, Neuroimaging, Neuroinformatics, and Unrelated to Neuroscience. Following classification, papers were aggregated by publication year and normalized via three strategies: division by totals in each discipline, division by annual OpenAlex counts, and a combined normalization method of the above two. See Figure 1 for the workflow chart. Results Analysis revealed a dramatic surge from 2015 to 2022 in Computational Neuroscience (12% increase per year), Neuroinformatics (18 % increase per year), and Neuroimaging (10% increase per year), whereas Cognitive and Behavioral Neuroscience displayed a plateau from 2013 to 2022 with slight declines afterward (Figure 1). Discussion Findings underscore the heterogeneous integration of AI across neuroscience disciplines, suggesting distinct developmental trajectories and new avenues for interdisciplinary research. The surge in AI applications post-2015 appears driven by advances in computational power, algorithmic innovations, and data availability, accelerating research in Computational Neuroscience, Neuroinformatics, and Neuroimaging[3]. Conversely, the plateau in Cognitive and Behavioral Neuroscience after 2013 may reflect shifting priorities or methodological challenges. These results can guide future studies to target underexplored intersections and inform strategic investments in emerging fields.
Figure 1. Data processing and analysis workflow (left); Number of Publication per year (top right); Yealy number of publication normalized by total publications (2001-2022) of each corresponding category. Acknowledgements We gratefully acknowledge the resources provided by OpenAlex and OpenAI. Their platforms enabled the data acquisition and automated classification essential to this bibliometric study. References [1] OpenAlex. (n.d.). OpenAlex: A comprehensive scholarly database. Retrieved from https://openalex.org [2] OpenAI. (2024, May 13). GPT‑4o API [Large language model]. Retrieved from https://openai.com/api [3] Tekin, U., & Dener, M. (2025). A bibliometric analysis of studies on artificial intelligence in neuroscience.Frontiers in Neurology,16:1474484.https://doi.org/10.3389/fneur.2025.1474484
