Ferguson Emily F, Steele Morgan, Schuck Rachel K, Millan Maria Estefania, Libove Robin A, Phillips Jennifer M, Gengoux Grace W, Hardan Antonio Y
Division of Child and Adolescent Psychiatry, Department of Psychiatry and Behavioral Sciences, School of Medicine, Stanford University, Stanford, CA, USA.
Ross University School of Medicine, Miramar, FL, USA.
J Autism Dev Disord. 2025 Mar 1. doi: 10.1007/s10803-025-06740-z.
Despite the importance of monitoring changes in expressive language in early intervention, existing approaches to language assessment are often costly, time-intensive, and capture limited variability in autistic children. The Language ENvironmental Analysis (LENA) system has thus received considerable attention as an automated approach that may hold promise for capturing fine-grained changes in language development in a more efficient and cost-effective manner. However, evaluations of the utility of the LENA system for tracking response to early intervention in unstructured contexts are currently limited.
This study aimed to build on prior research through evaluating the use of LENA in the context of a well-defined clinical sample from a randomized controlled trial (RCT) of Pivotal Response Treatment (PRT) that demonstrated expressive language gains across standardized and manually-coded measures.
Exploration of automatically-derived LENA metrics (i.e., child vocalizations, conversational turns) revealed no significant association with standardized language assessments (i.e., Mullen expressive language subscale, MacArthur Bates Communicative Development Inventory, Vineland-II expressive language subscale). Furthermore, relative to the delayed treatment group, children participating in PRT did not show significantly greater improvement in the number of vocalizations or conversational turns during naturalistic, daylong LENA recordings collected in home settings from baseline to post-intervention.
Implications and future directions for natural language sampling and the measurement of expressive language in early intervention are discussed.
尽管在早期干预中监测表达性语言变化很重要，但现有的语言评估方法往往成本高昂、耗时且只能捕捉自闭症儿童有限的语言变异性。因此，语言环境分析（LENA）系统作为一种自动化方法受到了广泛关注，它有望以更高效、更具成本效益的方式捕捉语言发展的细微变化。然而，目前对于LENA系统在非结构化环境中跟踪早期干预反应效用的评估有限。
本研究旨在基于先前的研究，通过在一项关键反应治疗（PRT）随机对照试验（RCT）的明确临床样本背景下评估LENA的使用情况，该试验在标准化和人工编码测量中均显示出表达性语言的提高。
对自动得出的LENA指标（即儿童发声、对话轮次）的探索表明，其与标准化语言评估（即马伦表达性语言分量表、麦克阿瑟-贝茨交流发展量表、文兰适应行为量表第二版表达性语言分量表）无显著关联。此外，相对于延迟治疗组，参与PRT的儿童在家庭环境中从基线到干预后的全天自然LENA记录期间，发声次数或对话轮次的改善并不显著。
讨论了自然语言抽样以及早期干预中表达性语言测量的意义和未来方向。
