Language and Reading

A major focus of our research has been on age-related changes in the orthographic (spelling), phonological (speech) and semantic (meaning) representations involved in reading and other linguistic processes. Our research has supported several general conclusions about the brain development of these representational systems. First, orthographic, phonological and semantic systems appear to become more elaborated with development. Age-related increases in activation are seen in brain regions implicated in processing each of these types of representation. Second, orthographic, phonological and semantic systems seem to become more integrated as there are developmental increases in activation in the brain regions involved in mapping between representations. Third, orthographic, phonological and semantic systems appear to become more specialized suggested by our findings that there is less overlap with development in the systems for processing auditory and visual information. There is also greater left hemisphere specialization as shown by age-related increases in activation in left inferior frontal gyrus. We have used effective connectivity to show that left inferior frontal gyrus also exhibits increases in top-down modulation on posterior brain regions involved in integrating task relevant information. Fourth, compensatory mechanisms seem to be used less over development, as there are age-related decreases in brain regions that are not critical for performing a specific task. In general, our research in language and reading is consistent with the interactive-specialization approach of brain development. An increasing focus of our research involves an examination of how the brains of children with dyslexia are different. We have shown that children with dyslexia exhibit less activation in several nodes of the reading network and show abnormalities in effective connectivity between these regions, particularly areas involved in orthographic and phonological processing. We have also demonstrated that children with dyslexia have subtle deficits in brain regions implicated in semantic processing.

In more recent work, we are using the lexicality effect (i.e. comparing words and pseudowords) to test dual-route versus connectionist models of reading acquisition. Our results are not consistent with dual-route models because we do not observe a dissociation between a ventral lexical route and a dorsal non-lexical route. Rather our results are consistent with connectionist models arguing for a progression from word specific representations to abstract coding that allows for generalization to novel stimuli. Our recent work is also comparing multimodal stimuli (audio-visual) to unimodal stimuli (auditory or visual) to test the central assumption that reading acquisition critically involves multisensory integration. We show developmental increases in multisensory integration but that it is specific to familiar stimuli. We are also currently testing the long-standing assumption that a critical deficit in dyslexia is in multisensory integration. In our previous work, we have extensively used effective connectivity methods to examine network dynamics, and we are now extending that work using graph theoretic analytical methods. We are showing skill related increases in segregation within an anatomical region and integration between anatomical regions in the reading network, but also that there is increasing separation of the reading network from other networks in the brain. We expect that children with dyslexia will show abnormalities in segregation, integration and separation. Our future work will examine whether we can predict who will have a language or reading impairment in second grade from neural measures in kindergarten. We have already shown in school age children that we can predict changes in reading skill 4- to 6-years later. Our results are consistent with models of reading development arguing for the early importance of phonological and later importance of orthographic skills in becoming a skilled reader. In the future, we will test contrasting models that argue for the differential role of morpho-syntax, phonology and short-term memory in the development of language and reading.