Abstract
Functional MRI (fMRI) studies have reported altered integrity of large-scale neurocognitive networks (NCNs) in dementing disorders. However, findings on the specificity of these alterations in patients with Alzheimer disease (AD) and behavioral-variant frontotemporal dementia (bvFTD) are still limited. Recently, NCNs have been successfully captured using PET with $^{18}$F-FDG. This study measured network integrity in 72 individuals using simultaneous resting-state fMRI and $^{18}$F-FDG PET. Results showed that while fMRI integrity was reduced across multiple networks in both patient groups, PET data revealed highly disease-specific patterns. Specifically, PET-derived integrity of the anterior default-mode network (DMN) accurately differentiated AD from bvFTD with 94% accuracy, suggesting metabolic connectivity as a superior diagnostic tool.
Introduction
Distinguishing between Alzheimer’s Disease (AD) and behavioral-variant frontotemporal dementia (bvFTD) is a significant clinical challenge, particularly in the early stages where symptoms overlap. Functional MRI (fMRI) has been the primary tool for studying neurocognitive networks (NCNs), but its results in dementia are often inconsistent due to low signal-to-noise ratios and the indirect nature of the BOLD signal, which measures hemodynamic changes rather than direct neural activity.
This study investigates whether "metabolic connectivity"—measured via $^{18}$F-FDG PET—provides a more specific and reliable diagnostic fingerprint for these disorders than traditional fMRI-based functional connectivity.
Methods
The researchers utilized a simultaneous PET/MRI scanner to capture metabolic and functional data under identical physiological conditions in 72 subjects (29 AD, 21 bvFTD, and 22 healthy controls).
Independent Component Analysis (ICA) was used to decompose the complex brain signals into distinct spatial networks. The analysis focused on four major hubs: the anterior Default Mode Network (DMN), posterior DMN, Salience Network (SN), and the right Central Executive Network (CEN). Integrity metrics were calculated for each subject to quantify how well their brain activity matched healthy network templates.
Key Results and Insights
The Modality Gap
A striking finding was the low correlation between fMRI and PET network indices, suggesting that functional and metabolic connectivity capture distinct biological processes. While fMRI showed reduced integrity across almost all networks in AD patients, PET data revealed much cleaner, disease-specific patterns.
Disease-Specific Metabolic Signatures
PET imaging demonstrated that AD primarily affects the posterior DMN, whereas bvFTD is characterized by significant hits to the anterior DMN, Salience Network, and Central Executive Network. This specificity allowed for high diagnostic accuracy:
- Classification Accuracy: PET-derived integrity of the anterior DMN alone distinguished AD from bvFTD with 94% accuracy (97% sensitivity).
- Combined Models: Adding the Salience Network to the PET model increased diagnostic accuracy to 96%.
- fMRI Performance: In contrast, the best fMRI-based predictor (posterior DMN) achieved only 64% accuracy.
Vascular Interference in fMRI
Post-hoc analysis suggested that the global impairment seen in fMRI for AD patients might be due to vascular issues, such as white matter hyperintensities, which "muddy" the BOLD signal. Metabolism, as measured by PET, appears to be a more stable and specific marker of network health in neurodegenerative contexts.
Conclusion
Simultaneous PET/fMRI reveals that metabolic connectivity is a more effective biomarker for differentiating AD and bvFTD than functional connectivity. The high disease specificity of PET-derived network integrity supports its potential as a powerful diagnostic tool. Future work should focus on validating these findings in larger cohorts and developing a standardized atlas for PET-based metabolic networks to facilitate clinical adoption.
Summary updated