脑机接口（Brain Computer Interface, BCI）通过在脑与外部设备间建立直接的通信通道，为探索神经机制、监测大脑状态及实现中枢神经系统功能的修复与增强提供了关键技术支撑。传统BCI研究主要依赖从局部脑区提取的活动特征，在解码的准确性与稳定性方面仍面临挑战。鉴于高级认知与行为源于大规模神经元集群间的动态交互而非孤立脑区活动，BCI研究日益关注解析脑网络的整体拓扑特性，为突破其性能瓶颈带来了新机遇。为全面阐述脑网络分析对BCI技术的赋能价值并提供领域参考，本综述系统梳理了其在BCI领域的核心应用与前沿进展。首先，在机制理解层面，揭示BCI学习与调控过程涉及大规模脑网络动态重组的神经可塑性，而非仅局限于局部活动变化；其次，在状态解码层面，既可提供先验知识以指导与优化解码模型设计，也可作为表征整体拓扑结构的新颖特征，显著提升解码性能；随后，在神经调控方面，为BCI干预提供了更精准的靶点选择依据和效果评估指标，促进了其在神经精神疾病诊疗中的应用效能。最后，本文总结了当前脑网络BCI研究所面临的挑战，包括虚假连接的辨识、个性化脑网络建模、解码性能瓶颈及实时计算需求，并对未来方向进行了展望，以推动该领域的发展。

 Brain Computer Interface (BCI) establishes a direct communication pathway between the brain and external devices, offering critical technological support for exploring neural mechanisms, monitoring brain states, and enabling restoration and enhancement of central nervous system functions. Traditional BCIs have largely relied on signal features extracted from localised brain activities and continue to face challenges in terms of decoding accuracy and stability. Given that higher-order cognition and behaviour emerge from dynamic interactions among large-scale neuronal ensembles—rather than isolated regional activity-BCI is increasingly focusing on analysing the global topological properties of brain networks. This shift presents new opportunities for overcoming performance bottlenecks. To comprehensively elucidate how brain network analysis can promote the development of BCIs and to provide field references, this review systematically surveyed the core applications and recent progress of brain network analysis in BCI. First, from a mechanistic perspective, brain network analysis reveals that BCI learning and modulation involve neuroplasticity characterised by large-scale networ reorganisation, extending beyond mere changes in local activity. Second, for state decoding, brain network analysis can serve
as either a priori knowledge to guide and optimise decoding models or as novel features to characterise the overall topological structure, thereby significantly enhancing performance. Third, in neuromodulation, brain network analysis provides a basis for more precise target selection and more comprehensive outcome assessment for BCI-based interventions, facilitating their
therapeutic efficacy in neuropsychiatric disorders. Finally, this review summarises the current challenges in brain networkbased BCI-including identification of spurious connections, individualised network modelling, decoding performance bottlenecks, and real-time computational demands-and offers an outlook on future directions to promote the development of this field.