Department of Systems Neuroscience, University Medical Center Hamburg-Eppendorf, Geb. W34, Hamburg, 20246, Germany.
Department of Systems Neuroscience, University Medical Center Hamburg-Eppendorf, Geb. W34, Hamburg, 20246, Germany.
Neuroimage. 2023 Mar;268:119868. doi: 10.1016/j.neuroimage.2023.119868. Epub 2023 Jan 13.
Cortico-spinal fMRI acquisitions aim to investigate direct interactions between brain and spinal cord, e.g. during motor output or pain processing, by covering both regions in a single measurement. Due to their large distance and location in the body, a dynamic shim update of constant and linear shim terms is required when using echo-planar imaging (EPI) to achieve reasonable image quality in both target regions. A previously presented approach with region-wise shim settings is based on a standard single-region shim algorithm and suffers from (i) non-optimal shim settings because it combines linear and second-order shim terms optimized for different volumes, and (ii) significant user interactions making it rather cumbersome, time consuming, and error-prone. Here, a dedicated ("CoSpi") shim algorithm for cortico-spinal fMRI is presented that performs joint optimization of static second-order shim terms and one set of linear and constant shim terms for each region in a single run and with minimal user interaction. Field map and T2*-weighted EPI measurements were performed on a clinical 3 T whole-body MR system in water phantoms and five healthy volunteers using the conventional region-wise and CoSpi shim settings as well as "gold standard" shim settings optimized for one of the target regions only. With CoSpi shim settings, (i) overall field inhomogeneity was reduced by about 65% / 75% (brain / spinal cord volume) compared to the conventional region-wise approach and in vivo was within 5% of the values obtained with the single-volume shim settings, (ii) geometric distortions derived from voxel displacement maps were reduced on average by about 35% / 70%, (iii) the temporal SNR determined from an EPI time series that may reflect the impact of through-slice dephasing, was increased by about 17% / 10%, and (iv) the variation of the mean field between slices, a measure targeting the predisposition to insufficient fat saturation and GRAPPA-related ghosting artifacts, was reduced by about 90% / 45%. Thus, the presented algorithm not only speeds up and simplifies the shim procedure considerably, but also provides a better field homogeneity and image quality, which both could help to significantly improve the applicability of cortico-spinal fMRI.
皮质脊髓 fMRI 采集旨在通过单次测量同时覆盖这两个区域，从而研究大脑和脊髓之间的直接相互作用，例如在运动输出或疼痛处理期间。由于它们之间的距离较大且位置在身体内部，因此在使用回波平面成像（EPI）时，需要对恒定和线性匀场项进行动态匀场更新，才能在两个目标区域中实现合理的图像质量。之前提出的一种基于区域的匀场设置方法基于标准的单区域匀场算法，存在以下问题：(i) 由于它将为不同体积优化的线性和二阶匀场项结合在一起，因此匀场设置并不理想；(ii) 由于需要大量的用户交互，因此该方法非常繁琐、耗时且容易出错。这里提出了一种专门用于皮质脊髓 fMRI 的 ("CoSpi") 匀场算法，该算法在单次运行中针对每个区域的静态二阶匀场项和一组线性和恒定匀场项进行联合优化，并且仅需要最小的用户交互。在临床 3T 全身 MR 系统上，在水模体和五名健康志愿者中进行了场图和 T2*-加权 EPI 测量，分别使用常规的基于区域的和 CoSpi 匀场设置以及仅针对一个目标区域优化的“金标准”匀场设置。使用 CoSpi 匀场设置，(i) 与常规基于区域的方法相比，整体场不均匀性降低了约 65%/75%（大脑/脊髓体积），并且在体内与单体积匀场设置获得的值相差 5%以内；(ii) 来自体素位移图的几何变形平均减少了约 35%/70%；(iii) 从可能反映切片间去相位影响的 EPI 时间序列中确定的时间 SNR 提高了约 17%/10%；(iv) 平均切片间场的变化，这是针对脂肪饱和度不足和 GRAPPA 相关鬼影伪影的倾向的度量，降低了约 90%/45%。因此，该算法不仅大大加快和简化了匀场过程，而且还提供了更好的场均匀性和图像质量，这两者都有助于显著提高皮质脊髓 fMRI 的适用性。
