Mutations in presenilin 1 gene (PS1) account for the majority of early-onset familial Alzheimer's disease (FAD) cases. The disease is characterized by intracellular neurofibrillary tangles and extracellular amyloid fibrils composed of amyloid beta peptides (Abeta). Two successive cleavages are necessary to free the Abeta peptide from the amyloid precursor protein (APP). Gamma-secretase catalyzes the final cleavage of APP to generate Abeta peptides. PS1 is a catalytic subunit of gamma-secretase and is also involved in the cleavage of many membrane proteins. PS1 also has functional interactions with many other proteins. The use of animal models of AD has initiated the deciphering of these molecular pathways and mechanisms. Transgenic mouse models are useful to study the features of FAD and to investigate the nature of the neural-tissue changes of the disease and their evolution during aging. When expressed alone, mutations in human PS1 do not induce any detectable lesions, although they do increase Abeta peptides. This absence has led to the criticism that PS1 mouse models are not valuable for the study of AD. In this review we present how studies using PS1 transgenic mice have raised new questions related to pathological mechanisms of AD and are useful models for the study of (1) progressive cognitive decline, (2) early-occurring synaptic dysfunction, and (3) mechanisms other than amyloidogenesis that can be involved in disease pathogenesis.