人工智能技术在学习障碍儿童中的应用及展望

Application and Prospects of Artificial Intelligence Technology in Children with Learning Disabilities

摘要: 目的：旨在梳理人工智能技术在学习障碍领域的应用现状，探明当前人工智能赋能学习障碍诊断和干预的途径和应用形式，后续相关研究与运用提供理论和实践指导。方法：在Web of Science数据库中，以artificial intelligence、learning disabilities及相近词作为检索词，检索至2025年6月。采用CiteSpace对本领域发文量、国家、机构、作者、共现关键词、关键词聚类和时区图谱进行可视化分析。结果：筛选后共纳入56篇文献。可视化分析发现：发文量最多的机构来自法国，在人工智能赋能阅读障碍儿童的诊断和干预领域有较大影响。关键词聚类分析得到8个聚类集群。结果表明，人工智能技术在学习障碍领域的应用场景主要包括预测、分类、诊断与评估、干预、个性化学习、心理支持等。突现词共现图谱表明本领域的演化路径及发展趋势表现为：人工智能技术在学习障碍儿童的诊断和干预服务中，最初的主要涉及阅读障碍，逐渐扩展到学习障碍的多种类型；从最初的分类和诊断，发展到多方面的特征提取和干预；从最初的计算模型，发展到深度学习等新兴算法。结论：人工智能在学习障碍领域的应用研究呈现出快速发展态势，尤其在针对学习障碍儿童的评估与治疗方向。研究热点从传统的症状分类体系向整合型诊疗系统转化，研究重心已逐步转向多维度的评估与个性化干预措施的开发，且值得关注的是，以深度学习、机器学习为代表的新兴技术手段的引入。后续研究可以针对学习障碍进行更为精准的分类研究，以提高学习障碍筛查的精准性以及干预方案的个体适配性。

Abstract: Objective: This study aims to systematically review the current application status of artificial intelligence (AI) technology in the field of learning disabilities (LD), exploring the pathways and application forms through which AI enables the diagnosis and intervention of learning disabilities, thereby providing theoretical and practical guidance for subsequent related research and implementation. Methods: Literature was retrieved from the Web of Science core database using search terms such as “artificial intelligence”, “learning disabilities”, and their synonymous or related keywords, covering publications up to June 2025. CiteSpace was employed to visually analyze the annual publication volume, countries, institutions, authors, keyword co-occurrence, keyword clustering, and timezone mapping in this field. Results: After screening, 56 articles were included in the analysis. Visualized analysis revealed that the institution with the highest number of publications is from France, which has significant influence in the field of AI-enabled diagnosis and intervention for children with dyslexia. Keyword clustering analysis identified 8 clusters. The results indicate that the application scenarios of AI technology in the field of learning disabilities primarily include prediction, classification, diagnosis and assessment, intervention, personalized learning, and psychological support. The keyword burst detection map demonstrates the evolutionary path and development trends in this field: AI technology in diagnostic and intervention services for children with learning disabilities initially focused mainly on dyslexia but has gradually expanded to various types of learning disabilities; its functions have evolved from initial classification and diagnosis to multifaceted feature extraction and intervention; and the methodologies have progressed from early computational models to emerging algorithms such as deep learning. Conclusion: Research on the application of artificial intelligence in the field of learning disabilities is developing rapidly, particularly in the assessment and intervention directions for children with learning disabilities. The research focus has shifted from traditional symptom classification systems to integrated diagnostic and intervention systems, with the core emphasis gradually turning to multidimensional assessment and the development of personalized intervention measures. Notably, the introduction of emerging technological approaches represented by deep learning and machine learning is particularly noteworthy. Future research could focus on more precise classification studies of learning disabilities to enhance the accuracy of screening and the individual adaptability of intervention programs.