基于变密度螺旋轨迹采集的亚毫米级等体素高分辨率fMRI技术研究

Functional magnetic resonance imaging (fMRI) has become an important method for brain science research since its ionizing-radiation-free and non-invasive features. However, due to the physical limitations of imaging characteristics and acquisition methods, traditional fMRI images usually have low spatial resolution. On the other hand, with the development of brain science, there are growing requirements for high-resolution fMRI, as the accuracy of locating brain activation areas is more and more demanded. The existing high-resolution fMRI technologies faces problems such as low time resolution, low signal-to-noise ratio, and severe distortions. To address these issues, this study aims to develop a new robust high-resolution fMRI method, through using variable-density spiral trajectory acquisition, and designing the method for reconstruction from highly undersampled data based on time-frame sharing and L0 norm constraints. Furthermore, the variable excitation pulse and inter-slice super-resolution reconstruction in the slice-selective direction are used to finally realize high-resolution fMRI with sub-millimeter (~0.8mm) isotropic voxels. Simulation, phantom and in vivo experiments will be carried out to validate the acquisition and reconstruction, and to analyze the accuracy of brain activation. This study will break through the limitations of traditional fMRI methods to achieve high-quality, practical high-resolution fMRI, which can help to locate the brain activation more accurately, and to promote the development of brain science research as well as related clinical applications.

功能磁共振成像(fMRI)因其无电离辐射、无创性等特点已经成为脑科学研究的重要手段。然而，由于成像特点和采集方法的物理限制，传统fMRI成像空间分辨率低；而随着脑科学领域的发展，对脑区激活定位的精准度要求越来越高，因此对高分辨率fMRI图像的需求日益增长。目前已有的高分辨率fMRI技术面临着时间分辨率低、信噪比低和变形严重等问题。针对这些问题，本研究以实现高分辨率fMRI成像为目标，使用变密度螺旋轨迹采集，设计基于时间帧共享和L0范数约束的高倍降采重建方法，提高空间分辨率；进一步采用层编码方向的变脉冲激发和层间超分辨率重建，最终实现亚毫米级（~0.8mm）的等体素高分辨率成像。并通过仿真、仿体和脑功能实验进行采集、重建方法验证和脑激活准确性分析。本研究将突破传统fMRI方法的局限，实现高质量、实用性强的高分辨率fMRI成像，帮助实现脑区激活的精准定位，推进脑科学研究和相关临床应用的发展。

功能磁共振成像(fMRI)因其无电离辐射、无创性等特点已经成为脑科学研究的重要手段。然而，由于成像特点和采集方法的物理限制，传统fMRI成像空间分辨率低；而随着脑科学领域的发展，对脑区激活定位的精准度要求越来越高，因此对高分辨率fMRI图像的需求日益增长。目前已有的高分辨率fMRI技术面临着时间分辨率低、信噪比低和变形严重等问题。针对这些问题，本研究以实现高分辨率fMRI成像为目标，使用变密度螺旋轨迹采集，设计基于时间帧共享和L0范数约束的高倍降采重建方法，提高空间分辨率；在重建算法中，将多线圈稀疏约束应用于时间总变分域，使用基于L0范数最小化的非线性重建算法代替传统的L1范数最小化算法，进一步提高了欠采样率进一步采用层编码方向的变脉冲激发和层间超分辨率重建，最终实现亚毫米级高分辨率成像。并通过仿真、仿体和脑功能实验进行采集、重建方法验证和脑激活准确性分析。本研究将突破传统fMRI方法的局限，实现高质量、实用性强的高分辨率fMRI成像，帮助实现脑区激活的精准定位，推进脑科学研究和相关临床应用的发展。