Investigating the representation of uncertainty in neuronal circuits.
Investigating the representation of uncertainty in neuronal circuits.
复制标题
研究神经元回路中不确定性的表征。
DOI：
10.1371/journal.pcbi.1008138
复制
发表时间：
2021-03
影响因子：
4.3
通讯作者：
Pouget A
中科院分区：
文献类型：
--
作者：
Dehaene GP;Coen-Cagli R;Pouget A
熟练的行为往往显示出贝叶斯推理的特征。为了让大脑执行所需的计算，神经元活动必须携带关于感觉输入的不确定性的准确信息。两种主要的方法已经被提出来研究神经表征的不确定性。第一种是贝叶斯解码方法，主要目的是使用贝叶斯规则解码人口活动刺激的后验概率分布，并间接产生不确定性估计作为副产品。第二种方法，我们称之为相关方法，搜索与不确定性相关的神经元活动的特定特征（如调谐曲线宽度和最大放电率）。为了比较这两种方法，我们推导出一个新的标准模型的声源定位的耳间时间差（ITD），再现了丰富的行为和神经的观察。我们发现，神经元活动的几个特征平均与不确定性相关，但没有一个在逐个试验的基础上提供准确的不确定性估计，这表明相关方法可能无法可靠地识别神经元反应的哪些方面代表不确定性。与此相反，贝叶斯解码方法表明，整个人口的活动模式，需要重建试验到试验后验分布与贝叶斯规则。这些结果表明，不确定性是不可能在一个单一的功能的神经元活动，并强调使用贝叶斯解码方法时，探索神经基础的不确定性的重要性。为了优化它们的行为，动物必须不断地表现出与它们的信念相关的不确定性。了解这种不确定性的神经代码是神经科学中一个紧迫而关键的问题。遵循悠久的传统，一些研究通过测量神经反应的平均统计数据（如调谐曲线）与刺激特征变化时的不确定性之间的关系来研究这种代码。我们表明，这种方法可能非常具有误导性。另一种方法是对神经元的反应进行解码，以恢复编码的感觉变量的后验分布，并使用该分布的方差作为不确定性的度量。我们证明，这种解码方法确实可以避免传统方法的陷阱，同时导致更准确的估计不确定性。
Skilled behavior often displays signatures of Bayesian inference. In order for the brain to implement the required computations, neuronal activity must carry accurate information about the uncertainty of sensory inputs. Two major approaches have been proposed to study neuronal representations of uncertainty. The first one, the Bayesian decoding approach, aims primarily at decoding the posterior probability distribution of the stimulus from population activity using Bayes’ rule, and indirectly yields uncertainty estimates as a by-product. The second one, which we call the correlational approach, searches for specific features of neuronal activity (such as tuning-curve width and maximum firing-rate) which correlate with uncertainty. To compare these two approaches, we derived a new normative model of sound source localization by Interaural Time Difference (ITD), that reproduces a wealth of behavioral and neural observations. We found that several features of neuronal activity correlated with uncertainty on average, but none provided an accurate estimate of uncertainty on a trial-by-trial basis, indicating that the correlational approach may not reliably identify which aspects of neuronal responses represent uncertainty. In contrast, the Bayesian decoding approach reveals that the activity pattern of the entire population was required to reconstruct the trial-to-trial posterior distribution with Bayes’ rule. These results suggest that uncertainty is unlikely to be represented in a single feature of neuronal activity, and highlight the importance of using a Bayesian decoding approach when exploring the neural basis of uncertainty. In order to optimize their behavior, animals must continuously represent the uncertainty associated with their beliefs. Understanding the neural code for this uncertainty is a pressing and critical issue in neuroscience. Following a long tradition, some studies have investigated this code by measuring how average statistics of neural responses (like the tuning curves) correlate with uncertainty as stimulus characteristics are varied. We show that this approach can be very misleading. An alternative consists in decoding the neuronal responses to recover the posterior distribution over the encoded sensory variables and using the variance of this distribution as the measure of uncertainty. We demonstrate that this decoding approach can indeed avoid the pitfalls of the traditional approach, while leading to more accurate estimates of uncertainty.
登录
查看更多内容
影响因子：
16.2
作者：
Beck JM;Ma WJ;Pitkow X;Latham PE;Pouget A
通讯作者：
Pouget A
影响因子：
25
作者：
Goris, Robbe L. T.;Movshon, J. Anthony;Simoncelli, Eero P.
通讯作者：
Simoncelli, Eero P.
影响因子：
16.2
作者：
Ecker AS;Berens P;Cotton RJ;Subramaniyan M;Denfield GH;Cadwell CR;Smirnakis SM;Bethge M;Tolias AS
通讯作者：
Tolias AS
影响因子：
16.6
作者：
Burge, Johannes;Geisler, Wilson S.
通讯作者：
Geisler, Wilson S.
DOI：
10.1037/h0061495
发表时间：
1948-01-01
期刊：
JOURNAL OF COMPARATIVE AND PHYSIOLOGICAL PSYCHOLOGY
影响因子：
--
作者：
JEFFRESS, LA
通讯作者：
JEFFRESS, LA