机器学习在孤独症谱系障碍中的应用与展望——我国近十年研究的可视化分析

Application and Prospect of Machine Learning in Autism Spectrum Disorder—Visual Analysis of Domestic Research in the Past Decade

摘要: 目的：探讨我国近10年机器学习在孤独症谱系障碍(ASD)诊疗等领域的发展历史和研究现状，为人工智能赋能ASD儿童的诊断与干预提供指导。方法：在中国知网(CNKI)检索相关文献，采用Citespace构建知识图谱进行可视化分析。结果：我国对于人工智能技术赋能ASD的关注程度持续上升，近年来发文数量增长迅速。已有研究主要关注机器学习在孤独症谱系障碍领域的筛查与干预，如比较Attention-AlexNet和AlexNet模型等在ASD筛查中的不同效能。后续研究可结合深度学习与多模态数据，构建更精准的ASD分类体系，同时拓展人工智能在个性化干预中的应用场景，为临床实践提供更具普适性的解决方案。

Abstract: Objective: To explore the development history and current research status of machine learning in the diagnosis, treatment and other related fields of Autism Spectrum Disorder (ASD) in China over the past decade, and to provide guidance for the diagnosis and intervention of ASD children empowered by artificial intelligence. Methods: Relevant literature was retrieved from China National Knowledge Infrastructure (CNKI), and Citespace was used to construct knowledge graphs for visual analysis. Results: The attention paid to the empowerment of ASD by artificial intelligence technology in China has been on a continuous rise, with the number of published papers growing rapidly in recent years. Existing studies mainly focus on the application of machine learning in the screening and intervention of ASD—for instance, comparing the different efficacies of models such as Attention-AlexNet and AlexNet in ASD screening. Future studies may integrate deep learning with multimodal data to build a more accurate ASD classification system, while expanding the application scenarios of artificial intelligence in personalized intervention, so as to provide more universal solutions for clinical practice.