Large Language Models in Neuroinformatics

Large Language Models (LLMs), such as GPT, LLaMA, Gemini, and other frontier and open-weight transformer-based architectures, are revolutionising how we represent and process information. In neuroscience, the application of LLMs has begun to unlock powerful new capabilities, from decoding brain activity to automating data annotation and enhancing literature-based discovery.This Research Topic invites interdisciplinary contributions exploring the integration of LLMs and LLM-based multimodal models into neuroinformatics. We are particularly interested in how LLMs can facilitate data curation, assist in the synthesis of scientific findings, and help bridge the gap between natural language and biological or computational neural representations.Potential directions include fine-tuning LLMs on domain-specific neuroscience corpora, integrating LLMs with neurosemantic ontologies, automating metadata annotation for neurodata repositories, and decoding naturalistic language representations from brain recordings. We also welcome work on agentic and retrieval-augmented workflows, and on the validation, interpretability, and responsible use of LLMs in the neuroscience domain.We welcome Original Research, Methods, Reviews, Perspectives, and Data Reports on topics including but not limited to:• Fine-tuning LLMs on neuroscience literature and structured neurodata• LLMs for metadata generation, annotation, and schema alignment in neuroinformatics• Brain-to-language and language-to-brain decoding models using LLMs• LLMs as tools for scientific hypothesis generation and research synthesis• Agentic and tool-augmented LLM workflows for neurodata analysis and discovery• Retrieval-augmented generation for literature synthesis and evidence grounding• Foundation and multimodal models for neural, imaging, and behavioural data• Integration of LLMs with neurosemantic ontologies and knowledge graphs• Applications of LLMs in data curation, dataset discovery, and publication linking• Evaluation, benchmarking, and validation of LLM-based neuroinformatics tools• Interpretability, explainability, and error mitigation in LLM usage for neuroscience• Responsible use, reproducibility, and governance of LLM-driven workflows in neuroscience Topic editors: Man Fai Leung, Hao SUN, Zhiyu Wan Source: https://www.frontiersin.org/research-topics/82714/large-language-models-in-neuroinformatics