Interactive-specialization of language impairment

This project will examine the neural basis of language development using functional magnetic resonance imaging (fMRI). In a longitudinal design, young children will be given word level and sentence level tasks that tap into phonological, semantic and morpho-syntactic processing. We hypothesize that temporo-parietal cortex becomes more specialized over development for distinct linguistic representations through increasing interaction (i.e. connectivity) with frontal cortex. Specialization should be characterized by increases in differential responses in temporo-parietal cortex to tasks tapping into distinct aspects of linguistic function. This project will test the interactive-specialization model that makes clear predictions of developmental brain differences [1, 2], but has little direct neuroimaging evidence to support it [10]. The central tenet of this model is that interaction between brain regions drives the specialization of brain regions for making certain computations. An extension of the interactive-specialization model predicts that children with LI should have lack of specialization in temporo-parietal cortex perhaps due to altered interaction between temporo-parietal and frontal regions. Functional connectivity between brain regions will be measured with Psychophysiological Interaction (PPI) and specialization will be measured with activation differences, but also with Representational Similarity Analysis (RSA). Examination of language development in the transition from preschool to elementary school is important because this transition is marked by mastery of complex morpho-syntactic principles, elaboration and refinement of semantic representations, and by increases in children’s ability to process phonemic information. Central to testing our overarching hypothesis is quantifying brain changes in young children, and differences that emerge in language impairment (LI). Previous fMRI studies have generally used cross-sectional designs and have only examined one age group, have examined a wide age range (i.e. from kindergarten through adolescence) and/or have not examined multiple linguistic processes. Moreover, only a handful of neuroimaging studies have examined the brain basis of LI in children. Our project will overcome these shortcomings by using a longitudinal approach to examine the development of multiple linguistic processes in young children. Longitudinal approaches are rarely used in the developmental neuroimaging literature, but are crucial for gaining insights into what mechanisms drive language development. A significant strength of our project is the use of theoretically motivated experimental tasks based on extensive behavioral literature to investigate innovative ideas regarding brain development in young children. The use of sophisticated change models will also allow us to determine whether the longitudinal growth in language impairment represents a quantitative versus qualitative difference and whether neuroimaging measures can be used to uniquely predict those children who are likely to fall further behind in language processing.