头部异物两步核磁共振电阻抗成像仿真研究

Simulation on two-step magnetic resonance electrical impedance tomography of brain anomaly tissues

在三层球和真实形状头模型(包括头皮、颅骨和大脑)上,采用基于径向基函数(RBF)神经网络的两步核磁共振电阻抗成像(MREIT)算法,对颅内病变组织阻抗成像进行了仿真研究.使用高分辨率的核磁共振成像系统对人体头部进行三维构建和不同组织边界区分,利用两步基于径向基函数神经网络的MREIT算法分别对颅内病变组织均匀阻抗和非均匀阻抗(每个单元的电阻抗值)分布进行估计.该两步MREIT算法适用于人体头部复杂组织结构的阻抗成像,仿真实验表明,采用MREIT技术对颅内占位性病变组织(血肿或脑瘤)的阻抗成像过程简单,重构的阻抗图像具有较高的精确性,算法具有一定的抗噪性能.

Two-step magnetic resonance electrical impedance tomography （MREIT） algorithm based on radius basic function （RBF） neural network was used to reconstruct the electrical impedance distribution of the encephalic pathological tissues on the three-sphere and realistic head model. The high resolution magnetic resonance imaging system was used to construct the three-dimensional head model and identify the boundary of different tissues. Then the two-step MREIT algorithm was applied to estimate the piece wise homogeneous and inhomogeneous impedance of the pathological tissue respectively. The simulation verified that the two-step MREIT algorithm is a feasible means to reconstruct the continuous electric impedance distribution, especially for the complicated human head tissues, with simple imaging process, robustness against noise, and high spatial resolution.