最佳注视位置效应的认知神经机制

最佳注视位置（Optimal Viewing Position, OVP）效应同人类的视觉信息加工效率密切相关，然而我们对该效应的认知神经机制知之甚少。 本项目拟解决两个重要问题：（1）考察OVP效应的时间进程并推断注视位置影响信息加工的哪些阶段；（2）OVP效应在不同类型视觉刺激加工中具有何种共同机制。本项目将从知觉经验和信息结构两个角度切入，以简体汉字、繁体汉字和物体图像等不同类型材料为载体，结合自由眼动和强迫注视两种范式，采用眼动、脑电和计算建模等技术对OVP效应认知加工的时间进程进行系统研究。 在实验的基础上，我们还将采用线性混合模型的方法和基于贝叶斯算法的理想观察者模型假设，探讨注视位置和信息特征对加工效率的预测力。 该项目将揭示OVP效应的神经活动规律，并对OVP效应在信息加工中的作用做出系统的阐释，为提高人脑的信息加工效率和增强计算领域的模式识别提供必要的科学依据。

The optimal viewing position (OVP) effects refer to the longstanding finding that the effectiveness of information processing is linked to the initial eye fixation within an isolated unit (e.g, word, face, object, etc.). However, there is still limited understanding on the cognitive neural mechanisms of the OVP effects. In this project, we aim to investigate two key issues related to the cognitive neural mechanisms: (1) the time course of the OVP effects and how these effects influence different processing stages (i.e., pre-attentive visual processing, early and later stages of information processing). (2) whether there is similar neural basis of OVP effects on processing across-domain stimuli. Our research methodology will combine paradigms of free eye movements and forced fixation, using eye tracking, computation and electrophysiological techniques. Our experiments will be designed on two aspects, one on perceptual experience through age development and one on information structure of lexical units through different types of Chinese characters as well as graphic objects as stimuli. Furthermore, the linear mixed model and an optimal fixation strategy based on a rational Bayesian observer will be used to help develop prediction models with respect to viewing position and features on performance. This project constitutes the first attempt to systematically investigate the cognitive processes and temporal dynamics of OVP effects. The findings of this project will not only offer better understanding of how OVP effects influence information processing in our mind, but also provide insights to some psychological basis and scientific suggestions for improving the efficiency of information processing. Furthermore, the findings will help to provide scientific basis for pattern recognition as well as to help in the construction and development of computational models for prediction.