Patent foramen ovale (PFO) is found in 40% of patients with acute ischemic stroke in whom the etiology of stroke is unclear (cryptogenic stroke patients). The incidence of new cryptogenic stroke patients with PFO is estimated to be at least 48,000 annually. Many are treated with warfarin or aspirin. However the risk of stroke recurrence or systemic embolization in medically treated cryptogenic stroke patients with PFO is unknown. An available option of PFO closure, either percutaneously or surgically exists and the interest in these modalities is high with proven efficacy. If used for all cryptogenic stroke patients with PFO, the potential national cost will be enormous. Therefore, before these procedures can be considered for use, there is a clear need to determine the rate of stroke recurrence or systemic embolization in medically treated patients. Such a trial will be extremely costly. A solution is to formulate a study that uses the structure of an existing stroke trial. The Warfarin Antiplatelet Recurrent Stroke Study (WARSS) is an NINDS- funded multi-center trial to assess the efficacy of warfarin versus aspirin in reducing stroke recurrence or systemic embolization after an initial stroke. In WARSS, 1,920 patients from 30 centers are randomly, double-blindly assigned to warfarin or aspirin and followed for a minimum of two-years. However, WARSS has no provision for PFO identification. Therefore, the study proposed here, PFO in Cryptogenic Stroke Study (PICSS), will use the WARSS structure in identifying cryptogenic stroke patients who will then undergo transesophageal echocardiography for PFO characterization. Patients will be randomized and followed per WARSS protocol. Twenty-seven of 30 WARSS centers will participate. Expected enrollment in PICSS is 474 cryptogenic stroke patients. The hypothesis tested is: PFO is an important risk factor for stroke recurrence or systemic embolization in medically-treated cryptogenic stroke patients which doubles the two-year rate of stroke recurrence or systemic embolization. Since patients are randomized to warfarin or aspirin the rate of stroke recurrence or systemic embolization on each agent may be compared. Transesophageal echocardiography allows for assessment of PFO size and detection of other potential cardioembolic sources. Therefore, pilot data on the natural history of medically treated cryptogenic stroke patients with these variables will also be available. Additionally, a subset of cryptogenic stroke patients with PFO in whom the stroke recurrence or systemic embolization rate is particularly high may be identified. This group could then serve as the target group for future trials involving other treatment modalities such as percutaneous closure.