Non-parametric representation and prediction of single- and multi-shell diffusion-weighted MRI data using Gaussian processes.
Non-parametric representation and prediction of single- and multi-shell diffusion-weighted MRI data using Gaussian processes.
复制标题
DOI：
10.1016/j.neuroimage.2015.07.067
复制
发表时间：
2015-11-15
期刊：
影响因子：
5.7
通讯作者：
Sotiropoulos SN
中科院分区：
文献类型：
--
作者：
Andersson JL;Sotiropoulos SN
磁共振弥散成像在研究人脑的微观结构和连通性方面具有很大的潜力。然而，扩散图像会受到图像失真和杂散信号丢失等技术问题的影响。纠正这些问题并不是一件容易的事，它依赖于拥有一种能够预测未来的机制。在本文中，我们描述了一种新的方法来表示和预测扩散磁共振数据。它基于一个或几个球体上的高斯过程，类似于“克里格法”的地统计学方法。我们提出了一种协方差函数的选择，它允许我们即使从具有复杂纤维图案的体素中也能准确地预测信号。对于多壳数据(多个非零b值)，协方差函数跨壳扩展，这意味着在对另一个壳进行预测时使用来自一个壳的数据。我们建议使用高斯过程(GP)来模拟扩散MRI数据。GP由一个协方差函数和一组直接从数据学习的超参数定义。它可以用来进行预测，这些预测可以随后用于例如失真校正。
Diffusion MRI offers great potential in studying the human brain microstructure and connectivity. However, diffusion images are marred by technical problems, such as image distortions and spurious signal loss. Correcting for these problems is non-trivial and relies on having a mechanism that predicts what to expect. In this paper we describe a novel way to represent and make predictions about diffusion MRI data. It is based on a Gaussian process on one or several spheres similar to the Geostatistical method of “Kriging”. We present a choice of covariance function that allows us to accurately predict the signal even from voxels with complex fibre patterns. For multi-shell data (multiple non-zero b-values) the covariance function extends across the shells which means that data from one shell is used when making predictions for another shell. We suggest using a Gaussian process (GP) to model diffusion MRI data. The GP is defined by a covariance function and a set of hyperparameters learned directly from the data. It can be used to make predictions that can subsequently be used in for example distortion correction.