Stroke affects more than 15,000 Veterans per year and is the leading cause of acquired adult disability in the United States. Improving functional disability for Veterans with stroke is a high priority for the VA healthcare system. Upper extremity weakness after stroke is a major source of functional disability with over two-thirds of stroke survivors unable to incorporate their affected arm into activities of daily life at six months post-stroke. In contrast to transformative advances in acute stroke care in recent years, novel therapies for rehabilitation after stroke remain limited. The current clinical state-of-the-art for post-stroke arm motor rehabilitation remains occupational and physical therapies. These standard approaches are not currently informed by post-stroke neuroanatomic injury or neurophysiology. Brain computer interfaces (BCIs) are a promising novel technology for stroke rehabilitation. Beyond substituting for lost motor function by allowing people to control computer cursors to regain communication or robot prosthetics to restore movement, BCIs also have the potential to enhance neurorehabilitation. The underlying principle of a rehabilitative BCI is that linking the neural activity of intended arm movement (i.e. via EEG) to the sensorimotor feedback of actual arm movement (i.e. via arm orthosis) recreates the Hebbian environment needed to maximally engage neural circuits and restore neurologic function. Small studies of BCIs to improve rehabilitation for patients with stroke have shown some promise, but the mechanisms by which BCIs enable recovery have been understudied. As a result, significant questions remain before meaningful clinical translation of BCIs for stroke rehabilitation can occur. Who are the patients who would optimally benefit from BCI neurorehabilitation? What is the optimal time period post-stroke for BCI training? In this proposed career development application, I will leverage a unique and ongoing clinical-research infrastructure at Providence VA Medical Center and Massachusetts General Hospital, which has recruited over 120 acute stroke patients with arm weakness in 2.5 years and followed these patients through the course of their first year of recovery with arm motor outcome measures including Fugl-Meyer and upper extremity kinematics, neuroimaging, and neurophysiology. I will directly extend this study by adding EEG- BCI arm orthosis sessions at four study time points, with the first session occurring within days of stroke. My first aim is to identify longitudinal changes in cortical functional connectivity induced by single- sessions of EEG-BCI arm orthosis training. I will investigate specific neural circuits that are strengthened by EEG-BCI training and the post-stroke time window in which circuits are most sensitive to training. My second aim is to evaluate how baseline arm motor severity and stroke neuroanatomy modulate the effects of EEG- BCI arm orthosis training on cortical functional connect...