Alzheimer's disease (AD) is the most prevalent neurodegenerative disease in the growing population of elderly people. A hallmark of AD is the accumulation of plaques in the brain of AD patients. The plaques predominantly consist of aggregates of amyloid-beta (Abeta), a peptide of 39-42 amino acids generated in vivo by specific, proteolytic cleavage of the amyloid precursor protein. There is a growing body of evidence that Abeta aggregates are ordered oligomers and the cause rather than a product of AD. The analysis of the assembly pathway of Abeta in vitro and biochemical characterization of Abeta deposits isolated from AD brains indicate that Abeta oligomerization occurs via distinct intermediates, including oligomers of 3-50 Abeta monomers, annular oligomers, protofibrils, fibrils and plaques. Of these, the most toxic species appear to be small Abeta oligomers. This article reviews the current knowledge of the mechanism of Abeta assembly in vivo and in vitro, as well as the influence of inherited amino acid replacements in Abeta and experimental conditions on Abeta aggregation. Challenges regarding the reproducible handling of the Abeta peptide for in vitro assembly studies are discussed.