PROJECT SUMMARY/ABSTRACT The ability to read is fundamental to living in modern society. Loss of reading ability due to stroke, called alexia, likely affects over a million Americans at any given time and causes difficulty performing many daily life functions. To improve diagnosis and treatment of alexia, we must understand the neurocognitive basis of reading deficits after stroke. However, prior small-scale studies using broad diagnostic categories and older neuroimaging methods have yielded only general lesion-behavior associations in alexia. Here, we propose the largest study to date of both alexia and typical reading in older adults, using detailed measures of reading ability and the most advanced multimodal neuroimaging methods available to test a new neurocognitive model. Reading relies on brain networks that evolved for speech and language processes, but neurocognitive models of reading have not yet incorporated recent advances in our understanding of these networks. We propose a new model of Reading Integrated with Speech and Semantics (RISS) that provides a more specific neurocognitive architecture for reading than prior models. We hypothesize that lesions of specific RISS network processors and connections account for specific reading deficit patterns after stroke, and that restoration of the injured RISS pathways or compensation in uninjured pathways underlie alexia recovery. Although brain networks for reading have been extensively mapped in typical and atypical young populations, stroke tends to occur in the aging brain and in people of low education and socioeconomic status (SES) who are too often left out of cognitive neuroscience research. Pathological patterns of reading in alexia are also observed to a lesser degree in typical readers, and age, education and SES are all known to affect reading abilities. Therefore, alexia can only be fully understood by examining how these factors relate to reading behavior and the brain in typical older adults, and referencing reading deficits to this personal context. In the first study of this project, we will collect an extensive battery of reading and language tests along with advanced multimodal MRIs in 100 older adults demographically matched to stroke survivors. We will test hypotheses based on RISS and examine how age, education, and SES relate to both behavior and the brain. We will freely disseminate all testing materials and both behavioral and imaging data to facilitate further research in this area. In the second study, we will perform the same behavioral battery in 200 chronic stroke survivors prospectively selected based on lesion attributes from a new imaging database of thousands of stroke survivors. We will model the effects of the lesions on processors and connections in RISS and test brain-behavior hypotheses using lesion-network mapping analyses. In the third study, we will collect detailed behavioral data and multimodal MRIs in 50 stroke survivors during the subacute p...