【Day37 文献泛读】Bayesian decision theory in sensorimotor control_bayesian learning sensorimotor control-优快云博客

本文链接：https://blog.youkuaiyun.com/u012127014/article/details/122798752

这篇文献探讨了神经系统如何在信号处理中考虑变异性与噪声，并指出人类行为接近于贝叶斯决策理论所预测的理想行为。文章讨论了在不确定性环境下，贝叶斯统计提供了解决问题的系统方法，以及在运动控制和感知中如何进行贝叶斯整合。研究发现，随着感官反馈中的噪声增加，个体会增加先验权重，减少对感官反馈的依赖，这与最优的贝叶斯统计预测相符。同时，贝叶斯过程可能也是感觉处理的基本元素，有助于理解不同模态线索如何被整合成单一估计。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

阅读文献：

Körding, K. P., & Wolpert, D. M. (2006). Bayesian decision theory in sensorimotor control. trends in cognitive sciences, 10(7), 319-326.

文献链接：

Bayesian decision theory in sensorimotor control - ScienceDirect

Abstract

1) Nervous system needs to estimate variability and noise when processing signials.

2) Human behaviour is close to that predicted by Bayesian Decision Theory, which defines optimal behaviour in a world characterized by uncertainty, and provides a coherent way of describing sensorimotor processes.

Introduction

1) Determining movement is a decision process in the central nervous system always companying with uncertainty.

→ Bayesian statistics provides a systematic way of solving problems in the presence of uncertainty.

2) Decision theory: The cost of each movement (such as energy consumed) must be weighed against the potential rewards that can be obtained by moving.

→ Rational choice of the movement is that maximizes utility according to decision theory.

Estimation using Bayes rule

Bayesian integration in motor control

1) With increasing noise in the sensory feedback subjects should increase the weight of the prior and decrease the weight of their sensory feedback in their final estimate of the location.

（详见Day36阅读文献）

在Day36文献中，上图b的纵轴为slope，而本文献中同样的图纵轴为weight of prior

猜测原因：在上图a中，deviation from target即为posterior的估计值，根据贝叶斯模型posterior最优估计的均值为：E(posterior)=αE(prior)+(1-α)E(likelihood)，因此上图a和b中的slope即prior的weight，值为