Current understanding of the origins of cerebral specialization and of hemispheric interactions is fairly limited. For example, it is unclear which recognized cortical asymmetries lead to lateralization, whether the net influence of one hemisphere on the other is excitatory or inhibitory, and to the extent to which the intact contralateral hemisphere contributes to recovery following a cortical lesion such as a stroke. The long-term goal of this work is to gain a better understanding of these issues by developing and studying neural models of emergent hemispheric lateralization and of hemispheric interactions as those models recover from simulated cortical lesions. The models consist of networks of paired left and right cortical regions connected by a simulated corpus callosum. The specific aims are: 1. to test the hypothesis that models have excitatory callosal connections and indirect interhemispheric competition can better explain data from biological/behavioral experiments than previous models; 2. to determine how learning one behavioral task can influence the direction/extent of another task's lateralization; 3. to determine how multiple underlying hemispheric asymmetries in a single model interact, altering the direction/extent of lateralization produced by each alone; and 4. to examine lateralization and post-lesion hemispheric interactions in a neurobiologically-grounded model of associative word learning that is directly comparable to behavioral, clinical and functional imaging data. This is the first systematic attempt to better understand cerebral specialization and transcallosal diaschisis using computational models. The results will directly relate to ongoing experimental work, have important implications for current theories of the mechanisms underlying hemispheric functional asymmetries and post-stroke recovery, and may suggest new therapeutic concepts for stroke patients.