Neuroanatomical outcomes of early literacy intervention in kindergarteners at risk for dyslexia

Learning to read is one of the most important developmental milestones and an instrumental skill in modern society. Approximately 7% of the population is diagnosed with developmental dyslexia, a specific learning disability associated with severe and persistent difficulties with word reading and/or spelling. Literacy interventions can be beneficial for individuals with dyslexia, and even more so if provided early, such as in kindergarten or first grade. Yet in current practice, targeted support is typically provided much later, which carries critical implications for literacy development and prospective health outcomes.  In parallel, emerging neuroimaging research suggests that early neuroanatomical differences are evident even before the onset of reading (difficulties), potentially contributing to poor literacy outcomes. Coupled with evidence on intervention-induced neuroplasticity in dyslexia and the presumably heightened brain plasticity in early childhood, these findings raise the question of whether preventive intervention that precedes formal education may promote better literacy outcomes in at-risk children. To date however, the neural underpinnings of such intervention in at-risk children have not been investigated.

To address this gap, this PhD project characterized neuroanatomy and learning-related plasticity in five-year-old preliterate kindergarteners. The first objective was to examine neuroanatomical profiles related to an increased risk for dyslexia, compared to typical development, before reading onset. Therefore, the investigation focused on pre-readers with and without a cognitive risk, as identified by behavioral screening in kindergarten. Early risk identification is critical in allowing us to provide support during the initial stages of reading development, when interventions are most effective. Driven by this intention, the second objective of this work was to investigate the neuroanatomical outcomes of a preventive game-based training focusing on early literacy skills. To tackle these objectives, multimodal MRI data were collected before and after a 12-week period in (1) at-risk pre-readers engaging with tablet-based early literacy training, (2) at-risk pre-readers engaging with tablet-based games unrelated to literacy, and (3) typically developing children.

At baseline, we observed bilateral risk-related differences in diffusion properties of reading-related white matter tracts, suggesting less organized white matter in at-risk children. Moreover, we found evidence of regionally larger and more myelinated right dorsal regions in at-risk children relative to typical controls. These findings demonstrate that, already prior to reading onset, differences in profiles of early literacy skills are reflected in structural brain measures. Thus, our results support the hypothesis that early neural systems may predispose certain individuals to struggle with reading and at the same time raise new questions on the emergence of potential right-hemispheric protective mechanisms.

Our longitudinal analyses revealed that at-risk children who engaged with the early literacy game exhibited a training-specific increase in cortical thickness of the left supramarginal gyrus, a perisylvian region with a proposed role in grapheme-phoneme integration. In addition, our analysis of white matter tracts revealed an overall bilateral increase in myelin estimates, albeit in the absence of training-specific changes in diffusion measures. Together, these findings offer a first indication that early training of grapheme-phoneme associations is associated with localized gray matter plasticity, presumably directly related to the trained skills, as well as more widespread microstructural changes. Given that myelin water fraction and cortical thickness change largely independently from each other, the training-related changes we observed in these measures likely reflect distinct, but complementary processes.

Lastly, within the framework of this research we aimed to evaluate the potential of myelin-specific imaging in providing a more detailed view of the relationship between white matter and reading in school-aged children, thereby laying the foundation for the use of this technique to study myelin plasticity in the context of literacy intervention. This study revealed a negative association between reading and myelin water fraction in left-hemispheric dorsal and ventral tracts, which contrasts previous hypotheses that dyslexia might be rooted in poor myelination. As myelin-specific imaging techniques gain further appeal, it would be valuable for future longitudinal work to address how myelin contributes to individual differences in reading ability, ideally across varying levels of reading experience.

To conclude, this dissertation made novel contributions to our understanding of the emergence of dyslexia and the impact of preventive intervention during the sensitive period of literacy acquisition. Our findings revealed that early neuroanatomy may distinguish children that differ in early literacy performance already before school entry, which underscores the importance of early risk identification. Our multimodal MRI approach enabled a nuanced view of the structural changes occurring during early literacy acquisition, revealing both localized as well as global plasticity. This training-related plasticity seems to reflect the positive behavioral gains observed in the intervention group, hence supporting the potential of digital games to support early literacy. Building on these insights, future work can help bridge our findings to educational practice by further investigating how early anatomy and plasticity shape long-term reading development in at-risk populations.