Mapping brain metabolic connectivity in relation to structural connectivity

Objectives: Positron emission tomography (PET) with 18F-fluorodeoxyglucose (FDG) is a valuable tool for recording baseline neural activity via measurement of regional glucose metabolism. Recent studies have shown promising results for so-called metabolic connectivity (MC), an interregional covariance of FDG-PET measures as proxy of brain connectivity. It is still unclear, however, to what extent MC reflects actual structural connectivity. Thus, the aim of this study was examine the degree of overlap between metabolic and structural connectivity networks at a whole brain as well as regional level.

Methods: Structural MRI, diffusion tensor imaging (DTI) and FDG-PET data of 303 subjects were acquired on a hybrid PET/MR system. Thirty two were healthy individuals, while other subjects have been imaged as part of a diagnostic work-up for causes of cognitive impairment. PET data were corrected for partial volume effects. Both PET and DTI images were spatially normalized to a standard space and segmented into 62 anatomical regions using a probabilistic atlas. Sparse inverse covariance estimation was employed to quantify metabolic connectivity. Structural connectivity was measured by streamline tractography through fiber assignment by continuous tracking. Finally, a metabolic and structural network with the same number of connections was generated.

Results: Both networks (see supporting figure) had a significantly higher similarity metrics and convergence ratio compared to random networks by chance. At the whole brain level, on average 90% of all possible connections were similar, i.e. either present or absent in both metabolic and structural network. On average 50% of detected connections were found to be convergent, i.e. present in both types of networks. Convergent were 80% of intrahemispheric intralobe connections, while only 30% of interhemispheric interlobe connections. A 100% similarity and convergence were found for connections between regions of the occipital lobe and perirolandic cortex.

Conclusion: On a gross spatial scale, structural connectivity may underline around 50% of interregional metabolic connections. This proportion is substantially higher for intrahemispheric and intralobe connections. Regions with a perfect overlap of connectivity are those myelinating earliest. The results support MC as a relevant concept in the field of brain connectivity. Research Support: MindView project (ID: 603002) within the FP7-HEALTH of the European Comission