Neurocognitive Processes as Longitudinal Predictors of Reading Fluency
Abstract
Reading fluency involves a complex interaction of different cognitive skills and abilities that develop with instruction and practice and relies on the automaticity of many distinct reading skills (e.g., pacing, word recognition, expression, phonological awareness). Fluent reading frees cognitive resources, such as working memory, for more advanced components of reading including comprehension and synthesis (Kim & Wagner, 2015; Samuels, 2006). Though early identification of students with fluency challenges is critical for providing evidence-based intervention, most screeners assess skills that are dependent upon instruction (e.g., letter identification, sound-symbol correspondence), delaying opportunities for identification. The current study examines the utility for including specific neurocognitive measures and reading error patterns as part of early screeners to predict future reading problems in children with dyslexia. In addition to standard reading screeners, data from neurocognitive measures (e.g., working memory, phonological processing, processing speed, rapid automatic naming) was collected from 61 at-risk children in grades 2-4 at the beginning of the school year. Reading fluency was assessed again after six months of reading instruction. Multiple hierarchical regression models and exploratory analyses investigated the incremental validity and differential strength of neurocognitive predictors in explaining variance in the acquisition of three distinct reading skills (word reading accuracy, word reading fluency, and text reading fluency) and types of reading errors (phonology, orthography, and attention). Consistent with previous literature, baseline reading measures, specifically measures of sight word reading (WJ-IV LWID), consistently emerged as significant predictors for growth in WRA (ß = .534, p = .018), WRF (ß = .824, p < .001), and TRF (ß = .699, p = .001). In these models, neurocognitive measures and types of errors did not significantly predict growth in word reading accuracy or word reading fluency above and beyond baseline reading. A measure of orthographic processing showed significance as a negative predictor of growth in text reading fluency above and beyond other neurocognitive predictors and some baseline reading measures. These findings highlight the importance of investigating the value and nuances of measurement of orthographic processing. Further research is needed to explore different measures of orthographic processing, the interactions of the neurocognitive measures and reading errors with conventional reading assessments, and to explore their specific contributions in predicting reading outcomes.