An event-related potential study of awareness of form-sound correspondence in Chinese children with reading disorders: Preliminary data

Abstract

Introduction: Children with developmental dyslexia are usually characterized as having difficulties learning graphemeñphoneme correspondence and applying the mappings. While Chinese characters do not contain components representing phonemes, over 80% of characters are phonetic compound characters containing a radical often serving as a phonetic cue (e.g. phonetic radical 將 zoeng1 in 漿 zoeng1 ëpasteí). The form-sound mapping can be described by (i) regularity ñ whether a phonetic compound character has the same pronunciation as its phonetic radical, and (ii) consistency ñ the degree of reliability of a phonetic radical as a phonetic cue for characters containing it. In light of the observations from normal Chinese adult readers of phonological consistency effects in P200 (180-230 ms) and N400 (300-500 ms) (Lee et al., 2006, 2007), this paper reports preliminary results of a study examining the sensitivity of Chinese reading-impaired children to form-sound correspondence in terms of effects of regularity and consistency using the event-related potentials (ERPs) technique. Methods: The participants included two right-handed Cantonese-speaking Primary 4 male students, one with reading impairment and one with normal reading performance. The third participant was a left-handed male student of the same grade who was formerly diagnosed as dyslexic and underwent reading remediation in the summer of 2009. His reading score at the time of study was within the normal range. The participants carried out a character recognition task in which they pressed separate buttons to indicate whether or not they had learned the character presented in each trial. Unlearned items were pseudocharacters created by rearranging the radicals from real characters. Learned characters were those taught by Primary 2. They were selected in terms of regularity (regular vs. irregular) and consistency (consistent = 0.88 vs. inconsistent = 0.21; values based at Primary 4). Cumulative frequency and visual complexity (i.e. number of strokes) were matched across experimental conditions. Results and Discussion: Behavioral results show that the dyslexic participant had the poorest accuracy in identifying real characters (66%), with comparable accuracy from the control (86%) and former dyslexic child (84%). An effect of lexicality at the N400 was found for the control participant, where pseudocharacters elicited a greater negativity (Hauk et al., 2006), see Figure 1. This was not seen in the other children. Moreover, the former dyslexic child showed the opposite pattern where real character evoked greater negativity. Similarly, regularity and consistency effects were found at the P200 only for the control, with regular and consistent characters eliciting greater positivity (see also Lee et al., 2006, 2007). The dyslexic participant showed no clear difference for regularity, and the opposite pattern of polarity for consistency. While no notable effects of regularity and consistency were observed for the former dyslexic child, his overall ERP waveform pattern was similar to the control. In short, the preliminary ERP findings of lexicality, consistency and regularity effects observed only in the control participant suggest that dyslexic children, remediated or not, may have weaker orthographic representations, and may be less sensitive to the internal structure of characters and its relationship with sounds. Figure 1. Preliminary ERP waveforms of lexicality, regularity and consistency effects at the N400 (top) and P200 (bottom) for the control, dyslexic and former dyslexic participants.