Detailed contribution information
| Contribution title | 3217 - Neuroimaging evidence on how simulation of letter-speech sound acquisition predicts initial reading outcome |
|---|---|
| Contribution code | PS03-67 (P) |
| Authors | |
| Form of presentation | Poster |
| Topic |
|
| Abstract |
Children with developmental dyslexia struggle to learn the associations between speech sounds and letters (Blomert, 2011; Hahn, et al., 2014). Although behavioral measures reach substantial classification accuracy, further research is needed to identify reliable and applicable predictors for dyslexia (Catts, et al., 2001; Puolakanaho, et al., 2007). Here, we provide a novel approach to increase specificity in early identification of struggling readers by simulating the process of learning letter-speech sound correspondences in prereaders. In a longitudinal study, 29 prereading children in their last year of kindergarten (6.7±0.3 years) performed an artificial letter training (<40 min). In a consecutive simultaneous EEG-fMRI session, they solved an implicit audiovisual task, including congruent and incongruent presentation of trained pairs. After half a year of formal reading instruction, children’s initial reading fluency was assessed (7.3±0.3 years) and 14 subjects were classified as poor readers (1 SD from mean). Individual learning rates during the artificial grapheme-phoneme training outperformed established behavioural precursors of reading by over 10% in predicting initial reading outcome. We also show how training induced plastic changes in brain networks of prereaders significantly improve the prediction of initial reading skills. A differential event-related potential after 400 ms, reflecting audiovisual integration of trained pairs, revealed a significant difference between normal and poor readers (p<0.001). In addition, region of interest analysis in the ventral occipitotemporal cortex (vOT) revealed an audiovisual integration effect only for normal readers (p<0.05). Taken together, an artificial letter training seems to be most informative for pursuing suitable predictors for dyslexia. It not only allows for quantifying the individual ability to learn correspondences, but also pinpoints training induced plastic changes in brain networks, such as the vOT, adapting to multisensory integration of orthographical and phonological information during reading acquisition (Brem, et al., 2010; Dehaene, et al., 2010). Our results critically extend the notion of an audiovisual integration deficit in poor readers by showing that this deficit is associated to deviating brain functioning in prereaders. To conclude, we propose novel behavioural and neurobiological precursors to identify an audiovisual integration deficit before reading acquisition. |