Priors in perception: Top-down modulation, Bayesian perceptual learning rate, and prediction error minimization

详细信息    查看全文

• 作者：Jakob Hohwy ^{Jakob.Hohwy@monash.edu}
• 关键词：Cognitive penetrability ; Bayesian inference ; Hierarchical Gaussian filter ; Prediction error minimization ; Variable learning rate ; Attention ; Active inference ; Illusion
• 刊名：Consciousness and Cognition
• 出版年：2017
• 出版时间：January 2017
• 年：2017
• 卷：47
• 期：Complete
• 页码：75-85
• 全文大小：356 K
• 卷排序：47

文摘

Top-down modulation of perception can be quantified in terms of a variable Bayesian learning rate, revealing a wide range of prior expectations that can modulate perception. The prediction error minimization framework can be used to define cognitive penetration specifically as prediction error minimization deviations from a variable Bayesian learning rate. Cognitive penetration is retained as a category somewhat distinct from other top-down effects, and carves a reasonable route between penetrability and impenetrability. Rampant, relativistic cognitive penetration of perception is prevented, and yet cognition and perception can be viewed as continuous.