Department of Psychiatry, University of Wisconsin Madison, Madison, WI, USA.
Department of Biomedical Engineering, University of Wisconsin Madison, Madison, WI, USA; Department of Medical Physics, University of Wisconsin Madison, Madison, WI, USA.
Neuroimage. 2014 Feb 1;86:221-30. doi: 10.1016/j.neuroimage.2013.09.001. Epub 2013 Sep 8.
Functional MRI blood oxygen level-dependent (BOLD) signal changes can be subtle, motivating the use of imaging parameters and processing strategies that maximize the temporal signal-to-noise ratio (tSNR) and thus the detection power of neuronal activity-induced fluctuations. Previous studies have shown that acquiring data at higher spatial resolutions results in greater percent BOLD signal changes, and furthermore that spatially smoothing higher resolution fMRI data improves tSNR beyond that of data originally acquired at a lower resolution. However, higher resolution images come at the cost of increased acquisition time, and the number of image volumes also influences detectability. The goal of our study is to determine how the detection power of neuronally induced BOLD fluctuations acquired at higher spatial resolutions and then spatially smoothed compares to data acquired at the lower resolutions with the same imaging duration. The number of time points acquired during a given amount of imaging time is a practical consideration given the limited ability of certain populations to lie still in the MRI scanner. We compare acquisitions at three different in-plane spatial resolutions (3.50×3.50mm(2), 2.33×2.33mm(2), 1.75×1.75mm(2)) in terms of their tSNR, contrast-to-noise ratio, and the power to detect both task-related activation and resting-state functional connectivity. The impact of SENSE acceleration, which speeds up acquisition time increasing the number of images collected, is also evaluated. Our results show that after spatially smoothing the data to the same intrinsic resolution, lower resolution acquisitions have a slightly higher detection power of task-activation in some, but not all, brain areas. There were no significant differences in functional connectivity as a function of resolution after smoothing. Similarly, the reduced tSNR of fMRI data acquired with a SENSE factor of 2 is offset by the greater number of images acquired, resulting in few significant differences in detection power of either functional activation or connectivity after spatial smoothing.
功能磁共振成像(fMRI)血氧水平依赖(BOLD)信号变化可能较为细微,因此需要采用影像参数和处理策略来最大化时间信号噪声比(tSNR),从而提高神经元活动诱导波动的检测能力。先前的研究表明,提高空间分辨率可以带来更大的 BOLD 信号变化百分比,并且对更高分辨率的 fMRI 数据进行空间平滑处理可以进一步提高 tSNR,超越原本在较低分辨率下采集的数据。然而,更高分辨率的图像需要更长的采集时间,并且图像数量也会影响可检测性。我们的研究目标是确定在更高空间分辨率下采集并进行空间平滑处理的神经元诱导 BOLD 波动的检测能力与使用相同成像时间采集的较低分辨率数据相比如何。在给定的成像时间内采集的时间点数量是一个实际考虑因素,因为某些人群在磁共振成像(MRI)扫描仪中保持静止的能力有限。我们在三个不同的平面空间分辨率(3.50×3.50mm(2)、2.33×2.33mm(2)、1.75×1.75mm(2))下进行比较,分别从 tSNR、对比噪声比以及检测任务相关激活和静息态功能连接的能力方面进行比较。还评估了 SENSE 加速的影响,SENSE 加速可以加快采集时间,增加采集的图像数量。我们的结果表明,在对数据进行空间平滑处理以达到相同的固有分辨率后,在一些但不是所有脑区,较低分辨率采集的任务激活检测能力略高。在平滑处理后,功能连接的分辨率没有显著差异。同样,SENSE 因子为 2 时采集的 fMRI 数据的 tSNR 降低,但由于采集的图像数量增加,因此在空间平滑处理后,功能激活或连接的检测能力几乎没有显著差异。