Summary: Project 3 Stroke is the leading cause of disability in the United States. Speech and language deficits (aphasia) following stroke are common. Yet, some individuals with aphasia experience rapid spontaneous recovery while others are left with persistent deficits. Likewise, rehabilitation appears to help some individuals but not others. Beyond the severity of initial impairments, we have no proven way to predict the extent of recovery one can expect. We suggest that better prognostic measures could help tailor individual treatment. The objective of this proposal is to determine whether the combination of neuroimaging, acute behavior and other biomarkers can improve our ability to map and predict recovery from stroke. First, we will use a new technique to map the anatomy of apraxia of speech, aphasia, a common symptom of left hemisphere injury. Second, we will see if we can predict the effectiveness of treatment provided to patients with chronic aphasia. Third, we will see whether these methods can also predict recovery in acute/subacute patients who receive either rehabilitation or a combined rehabilitation and brain stimulation therapy. Finally, we will explore if this computer-aided prognosis can be fully automated, providing an objective tool to help predict recovery from stroke.