Second order dimensionality reduction using minimum and maximum mutual information models.
Second order dimensionality reduction using minimum and maximum mutual information models.
复制标题
DOI：
10.1371/journal.pcbi.1002249
复制
发表时间：
2011-10
影响因子：
4.3
通讯作者：
Sharpee TO
中科院分区：
文献类型：
--
作者：
Fitzgerald JD;Rowekamp RJ;Sincich LC;Sharpee TO
用于表征多维神经特征选择性的传统方法，例如尖峰触发协方差（STC）或最大信息维度（MID），仅限于高斯刺激或由于维数灾难而仅能够识别少量特征。为了克服这些问题，我们提出了两个新的降维方法，使用最小和最大信息模型。这些方法是STC的信息论扩展，可用于非高斯刺激分布，以找到任意维数的相关线性子空间。我们将这些新方法与传统方法进行了两种方式的比较：生物启发的模拟神经元对自然图像的反应，以及猕猴视网膜和丘脑细胞对自然随时间变化的刺激的反应。对于非高斯刺激，最小和最大信息方法在所有情况下都显著优于STC，而MID在低维特征空间的范围内表现最好。神经元能够在其锋电位中同时编码关于感觉刺激的多个特征的信息。目前存在的降维方法，以提取这些相关的功能是有偏见的非高斯刺激或受害者的维数灾难。在本文中，我们介绍了两个信息理论扩展的尖峰触发协方差方法。这些新方法使用最小和最大互信息的概念来识别编码在神经元尖峰中的刺激特征。使用模拟和实验的神经数据，这些方法被证明在传统的方法是适当的，他们失败的情况下，表现良好。这些新技术应改善表征的神经功能的选择性，在大脑的区域，目前可用的方法的应用受到限制。
Conventional methods used to characterize multidimensional neural feature selectivity, such as spike-triggered covariance (STC) or maximally informative dimensions (MID), are limited to Gaussian stimuli or are only able to identify a small number of features due to the curse of dimensionality. To overcome these issues, we propose two new dimensionality reduction methods that use minimum and maximum information models. These methods are information theoretic extensions of STC that can be used with non-Gaussian stimulus distributions to find relevant linear subspaces of arbitrary dimensionality. We compare these new methods to the conventional methods in two ways: with biologically-inspired simulated neurons responding to natural images and with recordings from macaque retinal and thalamic cells responding to naturalistic time-varying stimuli. With non-Gaussian stimuli, the minimum and maximum information methods significantly outperform STC in all cases, whereas MID performs best in the regime of low dimensional feature spaces. Neurons are capable of simultaneously encoding information about multiple features of sensory stimuli in their spikes. The dimensionality reduction methods that currently exist to extract those relevant features are either biased for non-Gaussian stimuli or fall victim to the curse of dimensionality. In this paper we introduce two information theoretic extensions of the spike-triggered covariance method. These new methods use the concepts of minimum and maximum mutual information to identify the stimulus features encoded in the spikes of a neuron. Using simulated and experimental neural data, these methods are shown to perform well both in situations where conventional approaches are appropriate and where they fail. These new techniques should improve the characterization of neural feature selectivity in areas of the brain where the application of currently available approaches is restricted.
登录
查看更多内容
影响因子：
2.5
作者：
Cavanaugh, JR;Bair, W;Movshon, JA
通讯作者：
Movshon, JA
影响因子：
16.2
作者：
Atencio, Craig A.;Sharpee, Tatyana O.;Schreiner, Christoph E.
通讯作者：
Schreiner, Christoph E.
影响因子：
2.5
作者：
Fairhall, Adrienne L.;Burlingame, C. Andrew;Berry, Michael J., II
通讯作者：
Berry, Michael J., II
DOI：
10.1073/pnas.0706938104
发表时间：
2007-11-27
影响因子：
11.1
作者：
Chen, Xiaodong;Han, Feng;Dan, Yang
通讯作者：
Dan, Yang
果蝇嗅觉的感觉神经元的系统鉴定。
影响因子：
1.2
作者：
Kim AJ;Lazar AA;Slutskiy YB
通讯作者：
Slutskiy YB