Alzheimer's disease (AD) can be conceptualized as a network-based syndrome. Network alterations are linked to the molecular hallmarks of AD, involving amyloid-beta and tau accumulation, and consecutively neurodegeneration. By combining molecular and resting-state functional magnetic resonance imaging, we assessed whether different biological patterns of AD identified through a data-driven approach matched specific abnormalities in brain dynamic connectivity. We identified three main patient clusters. The first group displayed mild pathological alterations. The second cluster exhibited typical behavioral impairment alongside AD pathology. The third cluster demonstrated similar behavioral impairment but with a divergent tau (low) and neurodegeneration (high) profile. Univariate and multivariate analyses revealed two connectivity patterns encompassing the default mode network and the occipito-temporal cortex, linked respectively with typical and atypical patterns. These results support the key association between macro-scale and molecular alterations. Dynamic connectivity markers can assist in identifying patients with AD-like clinical profiles but with different underlying pathologies.