Perception / action interactions in children's spatial representations

The project will explore the striking phenomenon that 18-24 month old children, when disoriented, will search for a hidden toy by using the overall shape of the room but not coloured landmarks such as a blue wall.
Past theories of this effect have suggested that children cannot access landmarks to guide search, until they can encode them by language. We will test an alternative theory based on current understanding of the organization of the visual brain. Colour and object recognition, and geometrical information about locations and actions, are analysed in the separate 'ventral' and 'dorsal' streams. Even though both streams are functioning, the children's search performance may therefore be limited by immaturity in integrating information about objects and planned actions from the two streams.
We will test the generality of this idea by testing children's performance on a range of other tasks that require children to remember how 'dorsal' and 'ventral' features are linked. The studies will provide new information on how visuospatial systems interact in cognitive development, and should refine and deepen theoretical debates on the 'modularity' of cognition and its development.

本项目将探究一项引人注目的现象：18至24个月龄的幼儿在迷失方向时，会借助房间的整体轮廓搜寻隐藏的玩具，而非依赖蓝色墙面这类带有色彩的地标。

过往针对该现象的理论认为，幼儿只有在能够通过语言对地标进行编码后，才能借助地标指引搜寻行为。本项目将基于当前对视觉大脑组织机制的认知，检验另一套替代性理论。

色彩与物体识别、以及有关位置与动作的几何信息，分别由大脑独立的腹侧视觉通路（ventral stream）与背侧视觉通路（dorsal stream）进行处理。即便两条通路均已正常运作，幼儿的搜寻表现仍可能因无法成熟整合来自两条通路的物体信息与计划动作信息而受到限制。

我们将通过一系列要求幼儿记住背侧与腹侧特征关联方式的任务，测试幼儿的表现，以此检验该观点的普适性。本研究将为视觉空间系统在认知发展中的交互机制提供全新见解，同时有望完善并深化有关认知及其发展的「模块化」的理论争论。