Socioeconomic dissociations in the cognitive and neural correlates of reading disability

• In Event: Relations among Socioeconomic Status, Functional Brain Activity, and Neurocognitive Outcomes: Unified Framework Approaches

Abstract

Socioeconomic status (SES) is a robust predictor of children’s reading development, presumably due to variation in early literacy exposure and underlying cognitive development. Furthermore, reading disability (RD) is a highly heterogeneous disorder, with variable neurocognitive profiles of deficits across individuals. However, there has been little investigation into whether the neurocognitive foundations of RD vary by early environments, and thus, whether SES may predispose children to certain RD subtypes. The present study investigates whether SES moderates the neurocognitive patterns underlying RD.

Participants were 155 children aged 6-9 years (76 RD; 79 control) from diverse SES backgrounds (parental education 12-22 years). A median split on parental education yielded a lower SES group (< bachelor’s degree, n=73) and a higher SES group (≥ bachelor’s degree, n=82). Children completed standardized assessments of language, reading, and reading-related skills, including phonological awareness (PA), which indexes knowledge of how sounds form words, and rapid automatized naming (RAN), which reflects the automaticity of visual symbol recognition. On a second day, children completed two functional magnetic resonance imaging (fMRI) tasks indexing brain activation during phonological and orthographic processing. In both tasks, children saw a series of words and were instructed to press a button if two words in a row began with the same first sound (phonological task, n=90) or if two words in a row were the same visually (orthographic task, n=91). Effects of SES (high/low) and group (RD/control) were investigated in a priori language (inferior frontal, superior temporal) and reading (fusiform) regions.

Significant main effects indicated reduced performance and activation in the reading network in children with RD and those from lower SES backgrounds; however, SES moderated the relationship between diagnostic group and both cognitive and neural patterns. Specifically, higher SES children with RD exhibited greater deficits in PA, while lower SES children with RD exhibited greater deficits in RAN (Figure 1). Logistic regressions confirmed a double dissociation whereby the different skills predicted diagnostic group depending on SES. Neuroimaging results revealed commensurate interactions. Specifically, higher SES children with RD exhibited reduced accuracy (d') on the phonological task, as compared to SES-matched control-group peers, as well as reduced activation in language-related left perisylvian regions, including inferior frontal, superior temporal, and supramarginal regions (Figure 2). However, lower SES children with RD exhibited reduced accuracy on the orthographic task, as compared to SES-matched control-group peers, as well as reduced activation in bilateral fusiform regions involved in text processing (Figure 2).

Results indicate that SES systematically moderates the patterns of neurocognitive strengths and weaknesses underlying RD. This suggests that the hallmarks of RD may differ depending on a child’s early environments and experiences, which has important implications for early screening and targeted treatment strategies. More broadly, whereas phonological processing and its neural underpinnings are commonly recognized as the “core deficit” of dyslexia, the present study suggests that other reading-related skills may be more implicated in children from less advantaged backgrounds, and that representationally biased samples may mask critical heterogeneity in the neurocognitive patterns underlying both typical and disordered development.

Authors

• Rachel R. Romeo, University of Maryland - College Park
  Presenting Author

• Joanna A. Christodoulou, MGH Institute of Health Professions
  Non-Presenting Author

• Halie A. Olson, Massachusetts Institute of Technology
  Non-Presenting Author

• John D.E. Gabrieli, Massachusetts Institute of Technology
  Non-Presenting Author