Aiello Marco, Salvatore Elena, Cachia Arnaud, Pappatà Sabina, Cavaliere Carlo, Prinster Anna, Nicolai Emanuele, Salvatore Marco, Baron Jean-Claude, Quarantelli Mario
IRCCS SDN, Istituto di Ricerca Diagnostica e Nucleare, Via E. Gianturco 113, 80143 Naples, Italy.
Department of Neurosciences, Reproductive Sciences and Odontostomatology, Federico II University, via S. Pansini 5-ed. 17, I-80131 Naples, Italy.
Neuroimage. 2015 Jun;113:111-21. doi: 10.1016/j.neuroimage.2015.03.017. Epub 2015 Mar 17.
Recently introduced hybrid PET/MR scanners provide the opportunity to measure simultaneously, and in direct spatial correspondence, both metabolic demand and functional activity of the brain, hence capturing complementary information on the brain's physiological state. Here we exploited PET/MR simultaneous imaging to explore the relationship between the metabolic information provided by resting-state fluorodeoxyglucose-PET (FDG-PET) and fMRI (rs-fMRI) in neurologically healthy subjects. Regional homogeneity (ReHo), fractional amplitude of low frequency fluctuations (fALFF), and degree of centrality (DC) maps were generated from the rs-fMRI data in 23 subjects, and voxel-wise comparison to glucose uptake distribution provided by simultaneously acquired FDG-PET was performed. The mutual relationships among each couple of these four metrics were explored in terms of similarity, both of spatial distribution across the brain and the whole group, and voxel-wise across subjects, taking into account partial volume effects by adjusting for grey matter (GM) volume. Although a significant correlation between the spatial distribution of glucose uptake and rs-fMRI derived metrics was present, only a limited percentage of GM voxels correlated with PET across subjects. Moreover, the correlation between the spatial distributions of PET and rs-fMRI-derived metrics is spatially heterogeneous across both anatomic regions and functional networks, with lowest correlation strength in the limbic network (Spearman rho around -0.11 for DC), and strongest correlation for the default-mode network (up to 0.89 for ReHo and 0.86 for fALFF). Overall, ReHo and fALFF provided significantly higher correlation coefficients with PET (p=10(-8) and 10(-7), respectively) as compared to DC, while no significant differences were present between ReHo and fALFF. Local GM volume variations introduced a limited overestimation of the rs-fMRI to FDG correlation between the modalities under investigation through partial volume effects. These novel results provide the basis for future studies of alterations of the coupling between brain metabolism and functional connectivity in pathologic conditions.
最近推出的混合型正电子发射断层扫描/磁共振成像（PET/MR）扫描仪提供了同时测量大脑代谢需求和功能活动的机会，且二者在空间上直接对应，从而获取有关大脑生理状态的互补信息。在此，我们利用PET/MR同步成像技术，探究神经功能正常受试者静息态氟代脱氧葡萄糖正电子发射断层扫描（FDG-PET）提供的代谢信息与功能磁共振成像（rs-fMRI）之间的关系。从23名受试者的rs-fMRI数据生成了局部一致性（ReHo）、低频波动分数振幅（fALFF）和中心度（DC）图，并与同时采集的FDG-PET提供的葡萄糖摄取分布进行了体素级比较。考虑到灰质（GM）体积的部分容积效应，从大脑空间分布、整个组以及受试者间体素级的相似性方面，探究了这四个指标两两之间的相互关系。尽管葡萄糖摄取的空间分布与rs-fMRI衍生指标之间存在显著相关性，但在所有受试者中，仅有有限百分比的GM体素与PET相关。此外，PET与rs-fMRI衍生指标的空间分布之间的相关性在解剖区域和功能网络中均存在空间异质性，在边缘网络中相关性强度最低（DC的斯皮尔曼相关系数约为-0.11），在默认模式网络中相关性最强（ReHo高达0.89，fALFF高达0.86）。总体而言，与DC相比，ReHo和fALFF与PET的相关系数显著更高（分别为p = 10^(-8)和10^(-7)），而ReHo和fALFF之间无显著差异。局部GM体积变化通过部分容积效应，对所研究的两种模式之间rs-fMRI与FDG的相关性产生了有限的高估。这些新结果为未来研究病理状态下脑代谢与功能连接之间耦合的改变奠定了基础。
