Associations between FDG metabolism and fMRI connectivity in the default mode network depend on Aβ status in MCI patients

Abstract

AbstractBackgroundBrain function can be assessed using both resting‐state functional magnetic resonance imaging (rs‐fMRI) and Fluorodeoxyglucose (FDG) PET imaging. The associations between these biomarkers and their evolution with Alzheimer’s disease (AD) are unclear. We aimed to evaluate the associations between the default mode network (DMN) connectivity and glucose metabolism in mild cognitive impairment (MCI) patients with (Aβ+) and without (Aβ‐) AD pathology.MethodForty MCI patients (19 Aβ‐/21 Aβ+) performed between 2 to 4 rs‐fMRI acquisitions with an average follow‐up duration of 15 months. At baseline, DMN metabolism was estimated using [18F]‐FDG‐PET SUVr of the isthmus cingulate region‐of‐interest extracted from the Desikan‐Killiany atlas; and Aβ was measured using [18F]‐Flutemetamol‐PET. We used a Centiloid‐threshold>26 to define Aβ+. Multiple regressions were used to assess the FDG‐fMRI associations in the DMN, comparing the two subgroups using FDG‐Aβ interaction models. Table 1 provides demographics and follow‐up durations. All analyses were adjusted for age and sex.ResultsIn the entire sample, DMN connectivity and metabolism were not significantly associated to each other. However, we observed an interaction between Aβ status and FDG metabolism (0.23, p=0.03) such that, in the high‐Aβ subgroup, lower FDG metabolism was signiticantly associated with lower DMN connectivity (Fig. 1). In the low‐Aβ subgroup, no associations were observed. We also observed an FDG‐Aβ interaction when predicting longitudinal changes in DMN (‐0.27, p=0.04) such that, in the high‐Aβ subgroup, lower FDG metabolism was associated with a subsequent increase in DMN connectivity over time (Fig. 2).ConclusionIn high‐Aβ MCI patients, DMN connectivity is positively associated with FDG metabolism measured at baseline. However, the longitudinal association is negative with lower metabolism predicting increasing connectivity over time. Future research is needed to understand the mechanisms behind this inverse relationship. Hypothetically, a synchronous decrease in metabolism in several regions of the DMN could lead to an increase in the correlation between resting‐state activity measured in these regions.