Saccadic targeting deficits of Chinese children with developmental dyslexia: Evidence from novel word learning in reading

It is reported that, the primary approach for school children to acquire vocabularies is by deriving word meanings from contexts. The typical deficit of developmental dyslexia is that they have smaller vocabulary size than their chronological age-matched children. One recent study has examined the cognitive processes underlying dyslexic children's novel word learning during reading by using eye tracking. This is a method that is well established as a means of investigating reading behaviour by measuring when and where the eye fixates on text as written language is processed naturally. It should be noted that all the studies investigating novel word learning measured fixation durations on novel words fixated by dyslexic children with a view to quantifying the time required for those novel words to be successfully identified within the context of a sentence. In the present study we investigated saccadic targeting in relation to novel word learning in dyslexia children. Each novel word was embedded into eight sentences, each of which provided a context for readers to form a new lexical representation. Two groups of dyslexic children and age-matched control children's eye movements were recorded when they read sentences. Given the ongoing lexical processing difficulty influences the basic decision of "where to target" in Chinese reading, the novel word poses substantial processing difficulty, particular for dyslexic children with inefficient lexical processing, we predict that dyslexic children would be less efficient to target the eyes than control children did in novel word learning. Consistent with our prediction, the mean initial landing positions on novel words were further away from the word center for dyslexic than control children, showing that the basic decision of saccadic targeting on novel words was less efficient for dyslexic than control children. Additionally, we categorized 8 exposures to novel words as being two learning stages: Learning stage 1 including exposures 1 to 4; and learning stage 2 including the exposures from 5 to 8. We aimed to examine whether they were able to modulate their saccadic targeting as the accumulated learning of novel words. The results showed that, control children targeted the initial saccades closer to the word centers with increased exposures, while such effect did not occur for dyslexic children. These findings indicate that control children adjusted the initial saccadic targeting based on lexical familiarity information, while dyslexic children did not. On the basis of the findings above, we argue that, dyslexic children may adopt more careful strategy of saccade-target selection given their lower efficiency in word processing, such that they had lower efficiencies in the basic decision of saccadic targeting, as well as the usage of "lexical familiarity information" to modulate the saccadic targeting to novel words. This might account for their low word acquisition efficiency in reading.