The role of Aβ accumulation in the pathogenesis of Alzheimer’s disease (AD) is supported by genetic studies showing that mutations in the amyloid-β precursor protein (APP) that alter Aβ production are linked to a subset of familial AD (FAD) cases with autosomal penetrance (reviewed in ref. 1). Several of these FAD-associated APP mutations, as well as FAD-associated mutations in the presenilin 1 (PS1) and presenilin 2 (PS2) genes, lead to an increase in the production of Aβ₁–₄₂ relative to Aβ₁_₄₀. This, combined with the observation that these peptides are differentially deposited in senile plaques (SPs) in vivo, suggests that differential production of Aβ₁–₄₀ and Aβ₁_₄₂ may be crucially important in the pathogenesis of AD. Thus, it is important to use techniques that not only quantitate Aβ production, but also specifically differentiate between these two peptides in a variety of experimental paradigms. Here we describe the use of a highly sensitive sandwich-ELISA (enzyme-linked immunosorbent assay) to quantitate both Aβ₁–₄₀ and Aβ₁–₄₂ in soluble pools, after secretion by cultured cells into the medium or in human cerebrospinal fluid (CSF) samples, as well as in insoluble pools, as found intracellularly in cultured cells, or deposited in the brain parenchyma.