摘要儿童癫痫具有临床表现高度异质、诊断复杂、治疗反应个体差异显著等特点，传统诊疗模式面临误诊率高、疗效评估主观性强等挑战。近年来，人工智能技术快速发展，尤其是机器学习(ML)在儿童癫痫的多模态数据分析中展现出显著优势。现系统梳理ML在儿童癫痫筛选分型、辅助诊断、发作预测、治疗决策及术后管理等关键环节的应用进展，总结其技术路径与临床价值，为未来研究方向提供理论依据。尽管目前研究仍面临数据标准化不足、模型泛化能力有限及临床验证不充分等挑战，但ML技术通过整合多模态数据显著提升了儿童癫痫诊疗的精准性与效率。未来需聚焦跨机构数据共享、模型可解释性提升及临床嵌入式系统的开发，以推动人工智能技术从实验室向临床实践的转化。

abstractsPediatric epilepsy is characterized by high clinical heterogeneity, complex diagnostic challenges, and significant inter-individual variability in treatment response.Conventional diagnosis and treatment models are challenged by high misdiagnosis rates and strong subjectivity in efficacy evaluation.Over the years, artificial intelligence technology has developed rapidly.In particular, machine learning (ML) has shown significant advantages in multimodal data analysis of pediatric epilepsy.This article systematically reviews recent advances in ML applications for pediatric epilepsy, focusing on seizure classification, auxiliary diagnosis, seizure prediction, surgical decision-making, and postoperative management, and summarizes its technical pathways and clinical value, providing a theoretical basis for future research directions.While challenges remain in data standardization, model generalizability, and clinical validation, ML has significantly improved the accuracy and efficiency of pediatric epilepsy diagnosis and treatment by integrating multimodal data.Future research should prioritize cross-institutional data sharing, enhanced model interpretability, and the development of clinically embedded systems, so as to promote the translation of artificial intelligence technology from the laboratory to clinical practice.