Department of Cognitive Science, Budapest University of Technology and Economics.
Neuropsychology. 2014 May;28(3):472-83. doi: 10.1037/neu0000052. Epub 2014 Mar 3.
Grammar-specific accounts of specific language impairment (SLI) have been challenged by recent claims that language problems are a consequence of impairments in domain-general mechanisms of learning that also play a key role in the process of language acquisition. Our studies were designed to test the generality and nature of this learning deficit by focusing on both sequential and nonsequential, and on verbal and nonverbal, domains.
Twenty-nine children with SLI were compared with age-matched typically developing (TD) control children using (a) a serial reaction time task (SRT), testing the learning of motor sequences; (b) an artificial grammar learning (AGL) task, testing the extraction of regularities from auditory sequences; and (c) a weather prediction task (WP), testing probabilistic category learning in a nonsequential task.
For the 2 sequence learning tasks, a significantly smaller proportion of children showed evidence of learning in the SLI than in the TD group (χ2 tests, p < .001 for the SRT task, p < .05 for the AGL task), whereas the proportion of learners on the WP task was the same in the 2 groups. The level of learning for SLI learners was comparable with that of TD children on all tasks (with great individual variation).
Taken together, these findings suggest that domain-general processes of implicit sequence learning tend to be impaired in SLI. Further research is needed to clarify the relationship of deficits in implicit learning and language.
特定语言损伤(SLI)的语法特异性解释受到了最近的观点的挑战,该观点认为语言问题是学习的领域一般机制损伤的结果,而这些机制在语言习得过程中也起着关键作用。我们的研究旨在通过关注顺序和非顺序、言语和非言语领域,来测试这种学习缺陷的普遍性和本质。
我们使用(a)序列反应时间任务(SRT),测试运动序列的学习;(b)人工语法学习(AGL)任务,测试从听觉序列中提取规律;以及(c)天气预报任务(WP),测试非序列任务中的概率类别学习,将 29 名 SLI 儿童与年龄匹配的典型发育(TD)对照组儿童进行比较。
对于 2 个序列学习任务,在 SLI 组中,明显有较少比例的儿童表现出学习证据(χ2 检验,SRT 任务的 p <.001,AGL 任务的 p <.05),而在 WP 任务中,2 组的学习者比例相同。对于所有任务,SLI 学习者的学习水平与 TD 儿童相当(具有很大的个体差异)。
这些发现表明,领域一般的隐式序列学习过程在 SLI 中往往受损。需要进一步研究来阐明隐式学习和语言缺陷之间的关系。