摘要
我们通常利用多种感觉通道的信息对周围世界进行感知,如视觉、听觉、嗅觉、前庭感觉和本体感觉信息。为了确保对环境的正确感知,大脑必须把对同一物体特征进行表征的不同感觉信息整合成一致的、稳定的信息。以往的研究证实不同感觉信息之间是以统计最优化的模式结合的。文章首先综述了在贝叶斯理论的基础上建立起来的统计最优化模型以及其行为实验验证方法。然后,介绍了相关的神经成像研究结果和具有生理意义的神经网络模型。
In our daily lives, people utilize various kinds of information from different perceptive medals to perceive properties of surroundings, such as visual, auditory, olfactive, vestibular or proprioceptive cues： All of these types of information from various medals describe the same property of objects by diverse fashions. It is necessary to integrate the varions information when the brain processes properties of objects. Previous studies support that distinct sensory modals are integrated by the mode of the statistic optimization. The current article firsdy summarized the statistic optimization model based on the theory of Bayesian, and the methods of validating the model. Additionally, it introduced the neural network integrated model, which was based on several neural imaging evidences and owned its physiological meanings. Meanwhile, we also discussed the rationality and physiological accessibility of the statistic optimization modal.
出处
《心理科学》
CSSCI
CSCD
北大核心
2008年第4期1021-1023,共3页
Journal of Psychological Science
基金
西南师范大学青年基金项目SWNUQ2005036
西南大学国家重点学科一般项目NSKD06006
关键词
贝叶斯
最大似然估计
群编码
神经网络
Bayesian, Maximum Likelihood Estimate, population code, neural network
