The brains of patients with Alzheimer disease (AD) are characterized by an abundance of amyloid plaques that contain the Abeta peptide. Numerous studies have established that the soluble oligomers, and not the insoluble amyloid plaques, constitute the real culprits responsible for the AD associated neuronal death. This thesis undertakes two projects that focus on discerning the molecular mechanisms of Abeta aggregation into soluble oligomers. The first project was designed to test whether a specific hydrophobic side-chain interaction occurs during the early stages of Abeta aggregation. A working model of the Abeta amyloid fibril structure (based on solid-state NMR constraints) has the methyl group of Met35 positioned above the Phe19 aromatic ring (within 5 angstrom), and we reasoned that switching the aromatic (single) ring to a more hydrophobic (two) ring naphthyl system would stabilize this interaction and increase the random coil to beta sheet conversion that occurs during Abeta aggregation. Four modified Abeta(1-40) peptides were prepared with the Phe19 or Phe20 rings substituted with either 1- or 2-naphthyl rings. Circular dichroism revealed that the Phe19 modified peptides underwent more rapid random to beta sheet conversions. The 1H NMR spectra of the naphthyl peptides were not appreciably different from the wild-type peptide, and the chemical shift of the Met35 methyl signal did not change, suggesting that it may not reside above the naphthyl ring. These results suggest that the Phe19-Met35 interaction does not occur in the early stages of Abeta aggregation, and instead is involved in the later stages of association into beta sheet fibrils. For the second project, the goal was to determine a high-resolution structural model of the Abeta oligomer. The NMR spectra of monomers and oligomers (separated by size-exclusion chromatography) were essentially identical and consistent with random structure. However, the monomer had significantly lower surface tension and was random structured by circular dichroism, while the oligomer had higher surface tension and was beta sheet. These results suggest that the early-formed, soluble Abeta aggregates may associate into micelle-like structures and that the micelles act as reservoirs that eventually break down and lead to amyloid fibril formation.