Poster #79 - Examining the Functional and Genetic Mechanisms of Early Reading Disability Using Connectome-Based Predictive Modeling and Polygenic Risk Scores

• In Event: Poster Session 12

Abstract

The goal of this project is to identify genetic risk factors associated with reading disability that are measurable early in a child’s life when intervention is most effective. Current paper and pencil methods for early risk assessment for reading disability lack precision and reliability and require trained test administrators. Genetic risk screening requires a saliva sample, which can be obtained at any age, is scalable, inexpensive, and has the potential to be more precise and reliable than paper and pencil assessments.
We hypothesize that genetic risk factors will be associated with differences in brain function that account for individual differences in children’s reading ability. To test this hypothesis, we will use connectome-based predictive modeling (CPM) to generate a functional map for basic word reading in the 420 children from the New Haven Lexinome Project (NHLP). Then we will generate and utilize polygenic risk scores (PRS) for basic word reading ability to test how genes predict these functional brain maps in order to determine the efficacy of genetic screening for early assessment of risk for reading disability.
Initially, we will generate functional maps of basic reading ability using the NHLP as a learning sample, which we will cross-validate in an independent test sample. The NHLP is a multi-site longitudinal intervention study aimed at identifying the biological, cognitive, neurological and behavioral mechanisms of reading disability in children in grades 1 through 5. Using functional MRI scans, CPM will create a prediction model that selects the most important nodes (brain regions) and edges (connections between brain regions) to predict children’s basic reading ability as measured by the Woodcock Johnson-III Basic Reading Composite (WJ-III BRC). Then, the prediction model will be cross-validated on fMRI scans and WJ-III BRC scores from an independent test sample called the Genes, Reading and Dyslexia (GRaD) Study. The GRaD Study is a cross-sectional study of 1453 children in grades 3 through 5 with WJ-III and fMRI assessments and similar demographics to the NHLP.
To connect functional brain maps to genetic variants, we will then regress the strength of the nodes and edges identified in GRaD onto polygenic risk scores for basic reading ability derived from a meta-analysis of the genetics of reading comprised of 34,000 children ages 5 to 18 (Eising et al., 2021). Significant results will be cross validated in an independent cohort. This will allow us to connect genes, brain function, and reading ability. By identifying the neurological and behavioral manifestations of genetic risk for reading disability, this analysis will expand on the existing clinical tools for early identification of reading disability facilitating early intervention efforts.

Authors

• Mia Daucourt, Yale University
  Presenting Author

• Joan Bosson-Heenan, Yale University

• Jeffrey Gruen, Yale University