发表机构
Case Western Reserve University; University of West Florida; University of Central Florida（凯斯西储大学; 西佛罗里达大学; 中佛罗里达大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
NS-Copilot是LLM驱动的多智能体系统，可自主完成神经科学分析全流程，支持EEG等模态，在多任务基准8次试验中核心指标优于强基线，实现高效可扩展分析。
AI 中文摘要
AI正快速推动神经科学发展，但许多实验室因显著的跨学科壁垒未能充分发挥其潜力。尽管面向生理数据的预训练神经模型进展迅速，但其异构架构和模态特定约束阻碍了系统级整合、选择与评估。尽管近期基于大语言模型（LLM）的智能体系统在智能科学应用方面取得进展，现有方法仍缺乏必要的领域专业知识，无法有效选择和协调多样的神经科学预训练模型，也难以处理该领域的独特数据类型。我们提出NS-Copilot，一种用于神经科学分析的LLM驱动多智能体系统，可自主支持各类专业任务的端到端工作流。它统一了领域特定的预训练模型，支持脑电（EEG）、细胞外尖峰数据等关键神经科学模态，通过自然语言接口实现交互。给定原始数据和任务描述，NS-Copilot会协调具备规划、自适应控制、代码生成、结果合成等专门角色的智能体，无需依赖特定数据集的启发式方法即可完成分析。我们在涵盖阿尔茨海默病、帕金森病及工作记忆尖峰解码的神经科学基准上对NS-Copilot进行评估，每个任务开展8次试验，该系统在核心指标上始终优于强基线，展现出NS-Copilot在高效且可扩展的神经科学分析方面的能力。
英文摘要
AI is rapidly advancing neuroscience, yet many laboratories fail to fully unleash its potential due to significant interdisciplinary barriers. While pre-trained neural models for physiological data are progressing quickly, their heterogeneous architectures and modality-specific constraints hinder systematic integration, selection, and evaluation. Despite recent advances in large language model (LLM)-based agent systems for intelligent scientific applications, existing approaches often still lack the domain expertise required to effectively select and coordinate diverse neuroscience pre-trained models and handle unique data types in this domain. We present NS-Copilot, an LLM-driven multi-agent system for neuroscience analysis that autonomously supports end-to-end workflows for diverse professional tasks. It unifies domain-specific pre-trained models and supports key neuroscience modalities, including EEG and extracellular spike data, through a natural-language interface. Given raw data and a task description, NS-Copilot orchestrates agents with specialized roles for planning, adaptive control, code generation, and result synthesis, enabling analysis without dataset-specific heuristics. We evaluate NS-Copilot on neuroscience benchmarks spanning Alzheimer's disease, Parkinson's disease, and working memory spike decoding. Across 8 trials per task, the system consistently outperforms strong baselines on the primary metric, demonstrating the ability of NS-Copilot for effective and scalable neuroscience analysis. Our code is publicly available at https://github.com/FrankLiu1102/ns-copilot.
CommentsAccepted to Findings of EMNLP 2026. 21 pages, 9 figures, 7 tables. v2 adds the public code URL to the abstract and the Reproducibility Statement, and corrects that statement's description of automated checkpoint downloads; results and analysis are unchanged
