The problem of stroke of unknown origin is still unsolved. Since the report of the high prevalence of patent foramen ovale (PFO) in young stroke populations, paradoxical embolism has been suggested as a possible etiology. In this paper, we discuss the studies which support this hypothesis and, on the basis of our own experience, we propose the systematic search of this malformation by contrast echocardiography in each patient under age 60 with cerebral infarction of unknown origin after completion of the neurologic work-up. The transesophageal approach is the most sensitive method for detecting a PFO and must be used whenever possible. Deep venous thrombosis should be systematically searched for soon after admission, even in patients with no clinical signs. Unfortunately, specific therapeutic guidelines are not yet available for these cases, and the current approach to therapy is still empiric. Further studies are still necessary to define the best treatment.