Smolander Sini, Laasonen Marja, Lahti-Nuuttila Pekka, Arkkila Eva, Thordardottir Elin, Kunnari Sari
Department of Logopedics, School of Humanities, Philosophical Faculty, University of Eastern Finland, Joensuu, Finland.
Research Unit of Logopedics, University of Oulu, Oulu, Finland.
Int J Lang Commun Disord. 2025 Nov-Dec;60(6):e70125. doi: 10.1111/1460-6984.70125.
Differentiating typical language development (TD) and developmental language disorder (DLD) in a bilingual context is difficult. The societal language is often the only mutual language of the child and the SLT. It has been shown that when assessing second language (L2) performance using tools developed for monolingual children, substantial differences between typical and disordered development can be found. There is a need for a systematic understanding of the applicability of tests across different language domains, taking language exposure into account. Sentence comprehension is an important part of language development, and receptive difficulties have often been considered to have prognostic value. Yet, the diagnostic value of sentence comprehension has received little research attention.
In the Helsinki longitudinal SLI study (HelSLI), we investigated L2 exposure effects on sentence comprehension in sequentially bilingual typically developing children (BiTD) and bilingual children with DLD (BiDLD). In addition to group-level comparisons, we examined the classification accuracy of two L2 sentence comprehension tests, taking several explanatory factors into account.
A total of 100 six-year-old children were recruited from day care centres and a hospital clinic (54 BiTD children and 46 bilingual children with DLD). Two offline tests with multiple-choice and act-out tasks (Sentence Comprehension Test (Lausetesti) and Reynell Developmental Language Scales III, Verbal Comprehension Scale) were used to investigate sentence-level comprehension in Finnish. Multiple regression analysis was used to compare BiTD and BiDLD group performance while considering the effects of relative lifetime exposure to L2. Covariate-specific receiver operating characteristic (ROC) analysis was used to study the classification accuracy of the two tests and estimate the thresholds for optimal sensitivity and specificity of the tests.
The TD bilingual children performed significantly better than their peers with DLD in both sentence comprehension tests. The effect of L2 exposure was significant but small and affected both groups similarly. Both tests classified the groups with fair sensitivity and specificity at their best, but the accuracy depended greatly on exposure. Depending on the age and exposure to L2, a sensitivity of 0.80 yielded a specificity of 0.16-0.87 on the Sentence Comprehension Test and a specificity of 0.19-0.87 on the RDLS III Verbal Comprehension Scale.
Sentence comprehension in L2 is promising in informing the detection of language difficulties in L2 Finnish learners with several first-language backgrounds. However, interpretation must take L2 exposure into account. In addition, no one assessment tool or domain can be considered enough for the identification of DLD. In the future, the importance of sentence comprehension tests as classifiers should be considered as part of a larger assessment battery.
What is already known on this subject Differentiating developmental language disorder (DLD) from typical language development (TD) is difficult in bilingual children. Assessment tools are needed for more reliable detection of disordered development. It has been shown that assessing both languages in bilingual children adds to the classification accuracy. The societal language is, however, often the only mutual language between the SLT and the child, and the only available tests are frequently in the child's L2. It has been found that substantial differences in second language (L2) performance between bilingual typically developing children (BiTD) and bilingual children with DLD (BiDLD) groups can be found in different language domains. Language exposure effects in BiTD and BiDLD children as well as classification accuracy of the L2 tests vary, however. What this paper adds to existing knowledge L2 sentence comprehension has received little attention in the field of communication disorders. In the current study, sentence comprehension tests showed substantial differences in L2 performance between bilingual TD and DLD children. On a group level, the effect of L2 exposure was small relative to some other language domains and appeared similar in BiTD and BiDLD children. Classification accuracy, however, depended greatly on exposure. In addition, it was found that some tests classify the children with more exposure better and some those with less exposure. Such patterns are related to the difficulty of the test items and the age range targeted by each test. What are the potential or actual clinical implications of this work? Sentence comprehension is a promising language domain informing the identification of DLD in bilingual children with different L1 backgrounds. Even though the BiTD and BiDLD groups differ significantly in their test performance, it is advisable to consider the possible differences in sensitivity and specificity of the test depending on exposure. In the domain of sentence comprehension, despite the persistent difference in performance between BiTD and BiDLD children, DLD children seem to be able to benefit from the exposure in a similar manner as TD children instead of lagging behind in the course of accumulating exposure.
在双语环境中区分典型语言发展（TD）和发展性语言障碍（DLD）很困难。社会语言通常是儿童与语言治疗师（SLT）唯一的共同语言。研究表明，使用为单语儿童开发的工具评估第二语言（L2）表现时，典型发展和发育障碍之间会存在显著差异。有必要系统地了解各种测试在不同语言领域的适用性，并考虑语言接触情况。句子理解是语言发展的重要组成部分，接受性困难通常被认为具有预后价值。然而，句子理解的诊断价值很少受到研究关注。
在赫尔辛基纵向语言迟缓研究（HelSLI）中，我们调查了顺序双语的典型发展儿童（BiTD）和患有DLD的双语儿童（BiDLD）中L2接触对句子理解的影响。除了进行组间比较外，我们还考虑了几个解释因素，检验了两种L2句子理解测试的分类准确性。
从日托中心和医院诊所招募了100名六岁儿童（54名BiTD儿童和46名患有DLD的双语儿童）。使用两个包含多项选择和表演任务的离线测试（句子理解测试（Lausetesti）和雷诺尔发展语言量表III，语言理解量表）来研究芬兰语句子层面的理解。在考虑相对终生L2接触影响的同时，使用多元回归分析比较BiTD和BiDLD组的表现。使用协变量特异性接受者操作特征（ROC）分析来研究这两种测试的分类准确性，并估计测试最佳敏感性和特异性的阈值。
在两项句子理解测试中，TD双语儿童的表现均显著优于患有DLD的同龄人。L2接触的影响显著但较小，且对两组的影响相似。两项测试在最佳状态下对两组的分类都具有一定的敏感性和特异性，但准确性很大程度上取决于接触情况。根据年龄和L2接触情况，句子理解测试的敏感性为0.80时，特异性在0.16 - 0.87之间；在雷诺尔发展语言量表III语言理解量表上，特异性在0.19 - 0.87之间。
对于具有多种第一语言背景的L2芬兰语学习者，L2句子理解有助于发现语言困难。然而，解释时必须考虑L2接触情况。此外，没有一种评估工具或领域足以识别DLD。未来，应将句子理解测试作为分类工具的重要性视为更大评估组合的一部分。
关于该主题的已知信息 在双语儿童中区分发展性语言障碍（DLD）和典型语言发展（TD）很困难。需要评估工具来更可靠地检测发育障碍。研究表明，评估双语儿童的两种语言可提高分类准确性。然而，社会语言通常是语言治疗师和儿童之间唯一的共同语言，可用的测试通常是儿童的L2。研究发现，在不同语言领域中，双语典型发展儿童（BiTD）和患有DLD的双语儿童（BiDLD）组在第二语言（L2）表现上存在显著差异。然而，BiTD和BiDLD儿童的语言接触影响以及L2测试的分类准确性各不相同。本文对现有知识的补充 在交流障碍领域，L2句子理解很少受到关注。在当前研究中，句子理解测试显示双语TD和DLD儿童在L2表现上存在显著差异。在组层面上，相对于其他一些语言领域，L2接触的影响较小，并且在BiTD和BiDLD儿童中表现相似。然而，分类准确性很大程度上取决于接触情况。此外，发现一些测试对接触较多的儿童分类更好，而一些对接触较少的儿童分类更好。这种模式与测试项目的难度以及每个测试针对的年龄范围有关。这项工作的潜在或实际临床意义是什么？句子理解是一个有前景的语言领域，有助于识别具有不同第一语言背景的双语儿童中的DLD。尽管BiTD和BiDLD组在测试表现上存在显著差异，但建议根据接触情况考虑测试敏感性和特异性的可能差异。在句子理解领域，尽管BiTD和BiDLD儿童在表现上持续存在差异，但DLD儿童似乎能够与TD儿童以类似的方式从接触中受益，而不是在积累接触的过程中落后。
