Arteriovenous malformations (AVMs) are dangerous vascular lesions of largely unknown etiology. Those that occur in the brain (BAVMs) are particularly devastating, since they often lead to hemorrhagic stroke. Current treatment strategies are limited and may be as risky as the disease itself. At their core, AVMs are an aberration of normal vascular hierarchy caused by enlarged direct connections between the artery and the vein, which shunt blood past the capillary bed. Influenced by the discovery of a Notch-regulated program of arteriovenous (AV) specification, I hypothesized that disruption of this program could cause BAVM. My thesis work identifies increased endothelial Notch signaling as a candidate lesion in brain AVMs by showing that it is sufficient to cause BAVM-like abnormalities in a mouse model, and that it is upregulated in human BAVM tissue samples. Furthermore, it demonstrates that a sustained increase in Notch signaling is required to maintain the BAVMs in our mouse model, since high-flow AVMs rapidly regress to capillaries once Notch transgene is turned off. This work will be presented as three chapters, the first two of which have been published (PMID: 18667694 and PMID: 19546852), and the third of which is submitted for publication.