Auditory constraints on infant language learning

描述（由申请人提供）：本研究的目的是探讨婴儿语言习得背后的一般学习机制（GLMs）的听觉约束。glm在分词中的重要性已经得到了大量的证明，最近这种方法已经扩展到音素获取。对glm的主要批评是它们的计算能力太强。如果没有约束，就很难提取有意义的规则，这使得语言习得的任务成为一个潜在的棘手问题。听觉约束可能为glm的运作提供了一个良好的框架，促进了语言习得的任务。本研究的第一个具体目的是研究音素习得过程中听觉敏感性与分布学习之间的关系。7.5个月大的婴儿将接触到单峰或双峰的语音分布。一个习惯化过程将被用来测试一个假设，即在类别边界存在一个听觉敏感性增加的区域将促进音素习得。第二个具体目标是研究节奏分组偏差与统计学习在流利语音分词中的关系。在熟悉了包含单词边界的统计和节奏线索的人工语言之后，如果统计和感知到的节奏分组不一致，6.5个月和8.5个月大的婴儿预计会错误地分割语言中的统计单词。只有当语言经验是听觉节奏分组偏见出现的必要条件时，年龄效应才会出现。第三个具体目标是评估第一个和第二个目标中描述的机制的领域普遍性。非言语刺激将被用来评估在相同类型的分类学习和分割任务中的分布和统计学习的一般性。这项研究的目的是了解一个典型的发育中的婴儿的听觉系统是如何组织语言学习的。在一些非典型人群中，尚不清楚语言习得延迟是因为听觉系统受损还是因为负责学习的机制受损。未来的研究将使用这些任务来评估非典型语言学习者的学习情况（例如，有语音障碍的幼儿，植入人工耳蜗的幼儿，有语言障碍风险的婴儿），最终目标是开发用于临床人群的新干预措施。

DESCRIPTION (provided by applicant): The goal of the current research is to investigate auditory constraints on general learning mechanisms (GLMs) underlying infant language acquisition. There have been numerous demonstrations of the importance of GLMs in word segmentation, and more recently this approach has been extended to phoneme acquisition. The primary critique of GLMs is that they are too computationally powerful. Without constraints, it would be difficult to extract only meaningful regularities, making the task of language acquisition a potentially intractable problem. Auditory constraints may provide a good framework on which GLMs can operate, facilitating the task of language acquisition. The first Specific Aim of the research proposed in this application is to examine the relationships between auditory sensitivities and distributional learning during phoneme acquisition. 7.5-month-old infants will be exposed to either a unimodal or a bimodal distribution of speech sounds. A habituation procedure will be used to test the hypothesis that the presence of a region of increased auditory sensitivity at a category boundary will facilitate phoneme acquisition. The second Specific Aim is to investigate the relationship between rhythmic grouping biases and statistical learning during word segmentation from fluent speech. Following familiarization to an artificial language that contains both statistical and rhythmic cues to words boundaries, 6.5- and 8.5-month-old infants are expected to mis-segment the statistical words in the language if the statistics and the perceived rhythmic grouping are inconsistent. Age effects are expected only if linguistic experience is necessary for auditory rhythmic grouping biases to emerge. The third Specific Aim is to assess the domain generality of the mechanisms described in the first and second aims. Non-speech stimuli will be used to assess the generality of distributional and statistical learning in the same types of category learning and segmentation tasks. The goal of this research is to understand how the auditory system of a typically developing infant structures language learning. In some atypical populations, it is unclear whether language acquisition is delayed because the auditory system is compromised or because the mechanisms responsible for learning are compromised. Future studies will use these tasks to assess learning in atypical language learners (e.g., young children with phonological disorders, toddlers with cochlear implants, infants at risk for language impairments), with the eventual goal of developing new interventions for use in clinical populations.