Alzheimer’s disease (AD) is a neurodegenerative disorder characterized by neuronal loss and cognitive decline. Resting-state fMRI enables the assessment of brain alterations by examining large-scale functional networks. In this study, we investigated functional connectivity changes across stages of cognitive impairment using Independent Component Analysis (ICA) and dual regression. fMRI data from 94 AD, 89 cognitive normal (CN), and 249 mild cognitive impairment (MCI) subjects were preprocessed with DPARSF and analyzed with FSL’s Melodic to obtain 20 independent components. Dual regression was performed on them to evaluate connectivity differences between groups and two group-level comparisons were carried out: (i) AD vs. MCI vs. CN, and (ii) AD vs. early MCI (eMCI, n=100) vs. late MCI (lMCI, n=55) vs. CN. Voxel-wise statistical tests were run with FSL’s Randomise tool, correcting for multiple comparisons. Results showed a trend toward higher functional connectivity in AD relative to MCI (i), and greater connectivity in AD compared with eMCI, as well as in lMCI compared with CN (ii). These findings suggest stage-dependent alterations in resting-state networks, with AD exhibiting enhanced connectivity that may reflect compensatory mechanisms, while the distinct involvement in lMCI relative to CN points to an intermediate position between early impairment and AD.