Atypical predictive processing during visual statistical learning in children with developmental dyslexia: an event-related potential study

Abstract

Previous research suggests that individuals with developmental dyslexia perform below typical readers on non-linguistic cognitive tasks involving the learning and encoding of statistical-sequential patterns. However, the neural mechanisms underlying such a deficit have not been well examined. The aim of the present study was to investigate the event-related potential (ERP) correlates of sequence processing in a sample of children diagnosed with dyslexia using a non-linguistic visual statistical learning paradigm. Whereas the response time data suggested that both typical and atypical readers learned the statistical patterns embedded in the task, the ERP data suggested otherwise. Specifically, ERPs of the typically developing children (n = 12) showed a P300-like response indicative of learning, whereas the children diagnosed with a reading disorder (n = 8) showed no such ERP effects. These results may be due to intact implicit motor learning in the children with dyslexia but delayed attention-dependent predictive processing. These findings are consistent with other evidence suggesting that differences in statistical learning ability might underlie some of the reading deficits observed in developmental dyslexia.

Statistical Learning of Orthographic Regularities in Chinese Children With and Without Dyslexia

“…Specifically, children with dyslexia not only associated the phonetic/sematic radical with the primary, but also the secondary, sound/sematic category (i.e., inconsistent items), which may have caused some confusion in distinguishing the target from the distractor and, consequently, led to higher tendencies of mapping a distractor, instead of the correct pseudocharacter, onto the target sound or meaning. In fact, our explanation is supported by a recent study showing that although both children with dyslexia and their typically developing peers were able to learn the statistical patterns of the target and its predictors with various predictabilities (high possibility predictor: 90%; low possibility predictor: 20%; zero possibility predictor: 0%), children with dyslexia relied on both high (90%) and low (20%) possibility predictors to make a prediction, whereas typically developing children made predictions on the basis of the high (90%) possibility predictor only (Singh et al, 2018). Taken together, it seems more likely that children with dyslexia have intact ability to extract and integrate statistical learning from high‐consistency input, but they are not skillful enough to reject the distractors and trim the overfitting regularities learned from inconsistency items.…”

“…A recent and growing interest in the role of statistical learning in literacy acquisition has produced a number of studies that explore statistical learning as a potential source of reading difficulties in children with dyslexia (e.g., for a review, see Schmalz, Altoè, & Mulatti, 2016). The findings, though, are mixed, with some studies demonstrating that individuals with dyslexia have impaired statistical learning, whereasothers report evidence of intact statistical learning (e.g., Gabay, Thiessen, & Holt, 2015; Lum, Ullman, & Conti‐Ramsden, 2013; Sigurdardottir et al, 2017; Singh, Walk, & Conway, 2018). Moreover, these findings are exclusively derived from studies that use either implicit sequence learning tasks or a segmentation paradigm and focus on conditional statistical learning (i.e., learning based on the probability of co‐occurrence), with less attention paid to distributional statistical learning, which refers to the sensitivity to the frequency and variability of input exemplars (Erickson & Thiessen, 2015).…”

Probability learning and feedback processing in dyslexia: A performance and heart rate analysis

“…Other studies found impaired statistical learning in individuals with dyslexia for speech and nonspeech sound learning in adults (Gabay, Thiessen, & Holt, ) and for speech sound learning in children (Vandermosten, Wouters, Ghesquière, & Golestani, ). Additionally, there seems to be some evidence for impaired probabilistic learning of nonlinguistic material in visual (Singh, Walk, & Conway, ) and motor (Howard, Howard, Japikse, & Eden, ) domains.…”

How statistical learning interacts with the socioeconomic environment to shape children’s language development

“…Based on these previous findings [ 51 , 52 , 55 ], in the current study we predicted that if children learned the probabilistic relationships between the two types of predictor stimuli and the target, there should be significant differences in their response times (RTs) to the targets as well as differences in the ERP amplitudes to the predictors based on whether a trial was high or low probability. Specifically, we expected that reaction times would be quicker to the target when it was preceded by the HP predictor compared to the LP predictor and that ERP amplitudes would be significantly larger for the HP compared to the LP predictor.…”

“…Consequently, our final sample included 26 participants ages 8–12 (age mean = 10.12 years, SD = 1.48; 17 males). We chose this age range for four reasons: (a) there has been relatively less research on statistical learning and other forms of implicit learning in middle childhood, with the bulk of the research having been done on infancy through preschool and adolescence through adulthood; (b) there are major changes in the development of both statistical learning and language at the endpoints of this age-range (around 7–8 years and around 12 years), but statistical learning seems to show little change within this age range [ 54 , 58 , 59 ]; (c) after 12 years, statistical learning appears to have reached adult levels [ 52 ] and thus seems less relevant to our focus here on child development; and (d) the pattern learning task (with EEG measures) used in this study has been difficult to use with younger children, but has successfully been used with this age range [ 51 , 55 ].…”