Multiple Cues and Statistical Learning in Infancy

婴儿期的多重线索和统计学习

• 批准号: 7093151
• 负责人: NATASHA Z KIRKHAM
• 金额: $ 7.62万
• 依托单位: STANFORD UNIVERSITY
• 依托单位国家: 美国
• 财政年份: 2005
• 资助国家: 美国
• 起止时间: 2005-07-15 至 2008-12-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/7093151
• 关键词: behavioral /social science research tag; child behavior; clinical research; cognition; cues; early experience; eye movements; human subject; infant human (0-1 year); learning; visual perception; visual tracking

DESCRIPTION (provided by applicant): An important aspect of the ability to perceive the visual environment as coherent and intelligible is a facility to integrate multiple sources of information within and across sensory modalities. Integration of this type is essential for pattern detection, and for the understanding of object relations, dynamic social relations, and action sequences. An inability to integrate multiple sources of information could be a vital aspect in disorders of cognitive development. There is a great deal of evidence suggesting that infants in the first year of life are adept at detecting statistical regularites in the perceptual environment (e.g., Fiser & Aslin, 2003; Kirkham, Slemmer, & Johnson, 2002; Saffran, Aslin, & Newport, 1996). In addition, young infants are equally sensitive to correlations across modalities (e.g., Bahrick & Lickliter, 2000; Lewcowicz, 2000); indeed, they are capable of using this sensitivity to direct eye movements in anticipation of upcoming events (Richardson & Kirkham, 2004). Given the multimodal, dynamic nature of the natural environment, infants must be able to integrate the important correlations in order to establish coherent perceptual representations. Infants must be not only sensitive to these correlations, but also capable of learning from them. Using eye tracking technology and habituation procedures, I plan to investigate the importance and usefulness of multiple cue integration in statistical learning over the first year of life. How do multiple perceptual cues aid in pattern detection? What do infants do with their ability to detect these statistical relationships? Understanding the impact of multiple cue integration on infant learning in a probabilistic environment will help inform accounts of the relative role of learning and innate knowledge in development. Moreover, cognitive disorders, such as autism, may be related to problems with integrating different sources of information (Just, Cherkassy, Keller, & Minshew, 2004). Information about the parameters of multiple cue integration across the first year of life, therefore, might provide a valuable tool for clinical assessment.

描述（由申请人提供）：将视觉环境视为连贯且可理解的能力的重要方面是将多个信息来源整合到感官方式内和跨感官方式中的设施。这种类型的集成对于模式检测以及对对象关系，动态社会关系和行动序列的理解至关重要。 无法整合多种信息来源可能是认知发展疾病的重要方面。有很多证据表明，生命第一年的婴儿擅长在感知环境中检测统计规则材料（例如Fiser＆Aslin，2003; Kirkham，Slemmer，Slemmer，＆Johnson，＆Johnson，2002; Saffran，Aslin，Aslin和Newport，＆Newport，1996年）。此外，年轻婴儿对跨模态的相关性同样敏感（例如，Bahrick＆Lickliter，2000； Lewcowicz，2000）。确实，他们能够在预期即将发生的事件时使用这种敏感性来直接眼动（Richardson＆Kirkham，2004）。鉴于自然环境的多模式，动态性质，婴儿必须能够整合重要的相关性，以建立连贯的感知表示。婴儿不仅必须对这些相关性敏感，还必须能够向它们学习。我计划使用眼睛跟踪技术和习惯性程序，研究在生命的第一年中，多重提示整合在统计学习中的重要性和实用性。多个感知线索如何帮助模式检测？婴儿如何检测这些统计关系？了解多重提示整合对概率环境中婴儿学习的影响将有助于向说明学习和先天知识在发展中的相对作用。此外，认知障碍（例如自闭症）可能与整合不同信息来源的问题有关（Just，Cherkassy，Keller和Minshew，2004年）。因此，有关在生命第一年的多个提示集成参数的信息可能会为临床评估提供宝贵的工具。