Beyond representational alignment with brain-guided language models for robust reasoning

这项研究发表在《自然·机器智能》上，核心突破在于：科学家首次发现，人类大脑在推理时发出的神经信号，可以直接用来提升AI大模型的推理能力，准确率最高提升13%。通俗地说，此前AI与大脑的对比研究大多停留在“相似性”层面——发现AI内部活动模式与大脑有几分相像。但这篇论文更进一步，直接把脑扫描获得的信号“喂”给AI，当作指导AI思考的额外线索。研究人员让10个不同规模的大模型在处理逻辑推理题的同时，参考人类大脑负责推理区域的活跃模式，结果这些模型的表现明显变好，且这种提升不受语言训练数据的影响，说明脑信号提供了语言之外的全新信息。这项成果的意义在于，它把AI与大脑从“长得像”推进到了“能指导”的阶段。未来，这类技术可能让AI在复杂决策、法律推理、医疗诊断等需要严谨思考的场景中更可靠，出错的概率更低。长远看，结合人类认知规律训练的AI，也会让普通用户在使用智能助手时感觉它“更懂人、更靠谱”，而不是机械地回答问题。

Article Published: 03 August 2026 Beyond representational alignment with brain-guided language models for robust reasoning Mingqing Xiao (肖命清) ORCID: orcid.org/0000-0001-6191-77261,2, Kai Du (杜凯) ORCID: orcid.org/0000-0002-7505-15613 & Zhouchen Lin (林宙辰) ORCID: orcid.org/0000-0003-1493-75691 Nature Machine Intelligence (2026) Cite this article Save article View saved research A preprint version of the article is available at arXiv. AbstractThe correspondence between large language models (LLMs) and the neural mechanisms underlying human higher-order cognition remains insufficiently characterized. Given that language and reasoning in the human brain appear dissociable, an open question is whether LLMs align with neural signals from reasoning-related regions and whether such signals can improve them. Here, focusing on deductive reasoning, we show that LLM internal representations are not only partially aligned with task-based functional magnetic resonance imaging activity but can also be directly enhanced by these signals. Using a neural predictivity metric, we find that LLMs explain a substantial fraction of the explainable variance in reasoning-related regions at the aggregate level, whereas predictivity within specific reasoning types is lower, indicating both alignment and divergence. Building on this, we propose a brain-guided framework: we steer model representations along directions induced by the joint structure of model and brain representations, applying intervention at inference and fine tuning during training. We demonstrate that task-evoked brain signals can directly enhance LLM reasoning, yielding gains orthogonal to language-only supervision across ten LLMs (1.5B–72B parameters), with transfer across reasoning types and up to 13% absolute accuracy gain. Our results advance LLM–brain correspondences from correlation to guidance, establishing a brain-signal-driven pathway towards more robust and cognitively aligned artificial intelligence. This is a preview of subscription content, access via your institution Access options Access through your institution /* style specs start */ /* style specs end */ Access Nature and 54 other Nature Portfolio journals Get Nature+, our best-value online-access subscription $32.99 / 30 days cancel any time Learn more Subscribe to this journal Receive 12 digital issues and online access to articles $119.00 per year only $9.92 per issue Le