Restricted Boltzmann machines for neuroimaging: an application in identifying intrinsic networks.
Restricted Boltzmann machines for neuroimaging: an application in identifying intrinsic networks.
复制标题
DOI：
10.1016/j.neuroimage.2014.03.048
复制
发表时间：
2014-08-01
期刊：
影响因子：
5.7
通讯作者：
Plis SM
中科院分区：
文献类型：
--
作者：
Hjelm RD;Calhoun VD;Salakhutdinov R;Allen EA;Adali T;Plis SM
矩阵分解模型是目前解决神经影像数据中有意义的数据驱动特征的主要方法。其中，独立成分分析（伊卡）可以说是最广泛用于识别功能网络，它的成功导致了一些通用的扩展组和多模态数据。然而，有迹象表明，伊卡的灵活性和代表性的能力可能已经达到了极限，因为大多数这样的扩展是案例驱动的，定制的解决方案，仍然包含在类的混合模型。在这项工作中，我们寻求一个原则性和自然可扩展的方法，并考虑一个概率模型被称为限制玻尔兹曼机（RBM）。RBM通过将概率分布模型拟合到数据来从功能性脑成像数据中分离线性因子。重要的是，该解决方案可以用作更复杂（深度）模型的构建块，使其自然适用于分层和多模态扩展，这些扩展在单独使用线性因子分解时不容易捕获。我们调查的能力RBM识别内在的网络和比较其性能的著名的线性混合模型，特别是伊卡。使用合成和真实的任务功能磁共振成像数据，我们表明，RBM可以用来识别网络和他们的时间激活的准确性是等于或大于因子分解模型。RBM的有效性证明支持其作为更深层次模型的构建块的使用，这是未来神经影像学研究的重要前景。
Matrix factorization models are the current dominant approach for resolving meaningful data-driven features in neuroimaging data. Among them, independent component analysis (ICA) is arguably the most widely used for identifying functional networks, and its success has led to a number of versatile extensions to group and multimodal data. However there are indications that ICA may have reached a limit in flexibility and representational capacity, as the majority of such extensions are case-driven, custom-made solutions that are still contained within the class of mixture models. In this work, we seek out a principled and naturally extensible approach and consider a probabilistic model known as a restricted Boltzmann machine (RBM). An RBM separates linear factors from functional brain imaging data by fitting a probability distribution model to the data. Importantly, the solution can be used as a building block for more complex (deep) models, making it naturally suitable for hierarchical and multimodal extensions that are not easily captured when using linear factorizations alone. We investigate the capability of RBMs to identify intrinsic networks and compare its performance to that of well-known linear mixture models, in particular ICA. Using synthetic and real task fMRI data, we show that RBMs can be used to identify networks and their temporal activations with accuracy that is equal or greater than that of factorization models. The demonstrated effectiveness of RBMs supports its use as a building block for deeper models, a significant prospect for future neuroimaging research.
登录
查看更多内容
影响因子：
4.8
作者：
Kiviniemi, Vesa;Starck, Tuomo;Tervonen, Osmo
通讯作者：
Tervonen, Osmo
DOI：
10.1098/rstb.2005.1634
发表时间：
2005-05-29
影响因子：
6.3
作者：
Beckmann, CF;DeLuca, M;Smith, SM
通讯作者：
Smith, SM
影响因子：
4.8
作者：
Abou-Elseoud, Ahmed;Starck, Tuomo;Kiviniemi, Vesa
通讯作者：
Kiviniemi, Vesa
DOI：
10.1073/pnas.0601417103
发表时间：
2006-09-12
影响因子：
11.1
作者：
Damoiseaux, J. S.;Rombouts, S. A. R. B.;Beckmann, C. F.
通讯作者：
Beckmann, C. F.
影响因子：
3.3
作者：
BISWAL, B;YETKIN, FZ;HYDE, JS
通讯作者：
HYDE, JS