Combining neuroimaging and behavioral methods to predict developmental trajectory in dyslexic children

This research focused on assessing structural (voxel based morphometry, VBM) and functional (functional magnetic resonance imaging, fMRI) neuroimaging, behavioral assessments, and their combination in their predictive ability of future reading skills in dyslexic children and adolescents. It was predicted that the combined approach would yield a more powerful and precise model. Initially, 22 participants were scanned and assessed using behavioral measures, with a 2.5 year behavioral assessment only at follow-up. The average age of the 22 children was 14.25 at Time1, and 16.77 at Time2, with 9 males and 13 females. Neuroimaging scans were done during a rhyme judgment task, which was in line with the phonological processing theory of dyslexia. The outcome variable was the slope (i.e., change over time) of Word Identification standard scores of the Woodcock Reading Mastery Test Revised. Through multiple regression analysis, it was found that behavioral measures (i.e., Word Attack, a measure of phonetic decoding, and Passage Comprehension, a measure of passage reading comprehension) accounted for 19% of the variance. Two neuroimaging predictor variables were found which predicted 55% of the variance (i.e., hyperactivation in right inferior frontal gyrus and the increased gray matter volume of the right parietal lobule). Combination of these variables was able to account for 78% of the variance, thus represented a significantly stronger prediction model than behavioral and neuroimaging models did on their own. In conclusion, a combination of neuroimaging and behavioral approaches lead to a more precise prediction model of change in reading skills in a dyslexic sample.