Alzheimer's disease (AD) generally results in neuronal loss due to protein dysfunction in various brain regions. Genome-wide data have provided new opportunities to analyze the underlying mechanisms of AD. Here, we present a novel network-based systems biology framework to identify and analyze differentially activated pathways by integrating human protein-protein interaction data and gene expression profile data in six brain regions. Specifically, we propose a new scoring system by ranking the edges associated with AD. Then, an edge expansion algorithm is designed to identify the dysfunctional pathways implicated in AD pathogenesis in six brain regions respectively. The analyses of the similarities and differences of these dysfunctional pathways provide insights into understanding the dynamics of AD progression in six brain regions from a network perspective, which will further shed light on the pathogenesis of AD.