Spatial distribution of resting-state BOLD regional homogeneity as a predictor of brain glucose uptake: A study in healthy aging

- PET-FDG accurately represents glucose metabolism (CMRGlu), but remains complex and invasive;
- Finding a quantitative relationship between CMRGlu and an fMRI metric has been elusive;
- We used a cognitively healthy aging dataset to compare fMRI to CMRGlu in the grey matter;
- fMRI local synchrony (ReHo) correlated well to CMRGlu in both young and older groups;
- ReHo could be used as a semi-quantitative proxy of CMRGlu and as a biomarker of aging.

Positron emission tomography using [18F]-fluorodeoxyglucose (PET-FDG) is the primary imaging modality used to measure glucose metabolism in the brain (CMRGlu). CMRGlu has been used as a biomarker of brain aging and neurodegenerative diseases, but the complexity and invasive nature of PET often limits its use in research. There is therefore great interest in developing non-invasive metrics for estimating brain CMRGlu. We therefore investigated resting state fMRI metrics such as regional homogeneity (ReHo), amplitude of low-frequency fluctuations (ALFF) and regional global connectivity (Closeness) with multiple analytical approaches to determine their relationship to CMRGlu. We investigated this relation in two distinct cognitively healthy populations separated by age (27 young adults and 35 older adults). Overall, we found that both regionally and across participants, ReHo strongly correlated with CMRGlu in healthy young and older adults. Moreover, ReHo demonstrated the same age-related differences as CMRGlu throughout all cortical regions, particularly in the default network and frontal areas.