Nonparametric model for a tensor field based on high angular resolution diffusion imaging (HARDI)
Nonparametric model for a tensor field based on high angular resolution diffusion imaging (HARDI)
复制标题
基于高角分辨率扩散成像 (HARDI) 的张量场非参数模型
DOI：
10.1007/s11203-020-09236-y
复制
发表时间：
2021
影响因子：
0.8
通讯作者：
David C. Zhu
中科院分区：
文献类型：
--
作者：
L. Sakhanenko;M. DeLaura;David C. Zhu
我们开发了一种使用HARDI数据估计曲线轨迹的非参数技术。对于大脑的各个区域，我们考虑成像信号处理，并将多变量核平滑技术应用于描述从MRI图像获得的信号处理的一般函数f。在大脑的每一个位置，我们搜索的方向上的最大扩散的单位球，然后跟踪的积分曲线驱动的向量场，以获得曲线轨迹的估计。我们建立了适当的归一化曲线估计高斯过程的收敛性。这种方法是计算效率高的曲线跟踪的每一步，我们构建一个逐点的信心椭球区域，而不是穷举迭代采样方法。这些曲线轨迹是轴突纤维的模型，其位置和几何形状在神经科学中很重要。
We develop a nonparametric technique for the estimation of curve trajectories using HARDI data. For various regions of the brain, we consider the imaging signal process and apply multivariate kernel smoothing techniques to a general function f describing the signal process obtained from the MRI image. At each location in the brain we search for the direction of maximum diffusion on the unit sphere, and then trace the integral curve driven by the vector field to obtain the estimates of curve trajectories. We establish the convergence of the properly normalized curve estimators to a Gaussian process. This method is computationally efficient as with each step of the curve tracing we construct a pointwise confidence ellipsoid region as opposed to exhaustive iterative sampling methods. These curve trajectories are models of axonal fibers whose location and geometry are important in neuroscience.
影响因子：
14.5
作者：
Chang, Soo-Eun;Zhu, David C.
通讯作者：
Zhu, David C.
影响因子：
1.1
作者：
Carmichael, Owen;Sakhanenko, Lyudmila
通讯作者：
Sakhanenko, Lyudmila