Project Summary/Abstract Aphasia is one of the most common and debilitating consequences of stroke. Aphasia is caused by damage to language regions of the brain, which are usually localized to the left hemisphere. Fortunately, most individuals with aphasia after a stroke experience some degree of recovery of language function over time. The pace of recovery is greatest in the first weeks and months, but clinically meaningful gains in language function are possible even years after stroke. Recovery from aphasia is thought to depend on neural plasticity, that is, functional reorganization of surviving brain regions such that they take on new or expanded roles in language processing. However, despite much research, the mechanisms that underlie this process of functional reorganization remain poorly understood. The overall goals of this project are to better characterize the neural correlates of recovery from aphasia after stroke, and to determine which patterns of functional reorganization are associated with more or less favorable language outcomes. To address major limitations of prior studies, we will use a range of innovative approaches. Adaptive language mapping paradigms will be used to identify language regions in a reliable and valid manner, while minimizing performance confounds. We will recruit large numbers of patients at two established sites, enabling rigorous statistical approaches such as linear mixed models and permutation testing. Advanced machine learning algorithms will allow us to disentangle complex relationships between structural damage, neurofunctional changes, and language outcomes. We will study two complementary cohorts of patients—acute and chronic—longitudinally using the same multimodal functional and structural MRI protocol, and the same language evaluations. In the acute cohort, we will investigate the dynamics of early recovery and functional reorganization, while in the chronic cohort, we will identify relationships between patterns of reorganization and language outcomes in a larger sample, and track the neural correlates of recovery in the chronic phase. Our first specific aim is to build predictive models of the trajectory of evolving aphasia profiles based on structural neuroimaging, including lesion location and extent, as well as multiple measures of the integrity of surviving tissue, and other clinical variables. Our second specific aim is to characterize the brain regions recruited for language processing in people with aphasia, and to identify patterns that are associated with more or less favorable outcomes. Critically, the key analyses will use the predictive models from Aim 1 to determine which patterns of functional reorganization result in better or worse outcomes relative to what would be expected on the basis of structural and clinical factors alone. Our third aim is to identify changes in functional activity associated with gains in language function over time, again in the context of the predictive...