In the last ten years the literature on Alzheimer's disease has focused on in vivo neurobiological changes. Extracellular beta-amyloid and intracellular hyperphosphorilated tau deposition are pivotal features and several authors have described their progression over time in pathological specimens. The revised criteria for Alzheimer's disease suggest that in vivo biomarkers reflecting neurobiological changes are useful for early diagnosis in the clinical practice. The most widely used biomarkers for the Alzheimer's disease are: Abeta42 and Tau levels in the cerebrospinal fluid (CSF); brain glucose hypometabolism detected by positron emission tomography (PET) using the 2-deoxy-2-[18F]fluoro-D-glucose ([18F]FDG); brain structural and functional changes detected by magnetic resonance imaging and amyloid burden by PET using Carbon-11-labelled Pittsburgh compound-B [(11)C]PiB). In order to explain the latest in vivo observations, the dynamic biomarker hypothesis has been recently developed, integrating both pathological and clinical knowledge, and indicating which biomarkers might be more sensitive to disease state and progression at different stages. In this review, we will outline studies that support the dynamic hypothesis by: 1) testing slope differences among biomarkers; 2) describing how biomarkers map all neuropathological and clinical changes starting from the detection of amyloid burden, to neurodegeneration, and symptoms; and finally 3) identifying the best combination of biomarkers sensitive to prodromal AD in clinical practice.