Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Prediction of reading processing time in adults using auditory event-related potentials
Poster Session C, Friday, October 25, 4:30 - 6:00 pm, Great Hall 3 and 4
Eunbi Park1, Yukihide Horiguchi1, Fumitaka Homae1,2; 1Depeartment of Language Sciences,Tokyo Metropolitan University, 2Research Center for Language, Brain and Genetics, Tokyo Metropolitan University
Reading involves various cognitive processes, resulting in variations in individual reading times. In this study, we explored whether processes that differ between individuals with dyslexia (who have specific reading difficulties) and non-dyslexic controls could also explain reading processing times in non-dyslexics. Two types of auditory processing, speech sounds and temporal changes, have been reported to be difficult for dyslexics, and are reflected in neurophysiological indices (Goswami, 2011; Sharma et al., 2006). Herein, we investigated the relationship between reading processing speed and auditory event-related potentials (AERPs) elicited by speech sounds and temporal changes. Data from 50 native Japanese speakers (23 females, mean age: 20.6 ± 1.7) who did not report difficulty in reading were analyzed. Electroencephalography (EEG) recording was performed while two types of oddball sessions were conducted: 1) speech sound condition: foreign phoneme contrast of English /la/ and /ra/, and 2) pure tone condition: different interstimulus intervals of the same pure tone stimulus. The mean amplitude differences between the standard and deviant stimuli in two time windows (mismatch negativity, MMN: 145 – 297 ms; late discriminative negativity, LDN: 297 – 600 ms) were calculated for each condition. After EEG recordings, participants performed a reading assessment designed for Japanese speakers (STRAW-R; Uno et al., 2017), which included four different tasks: Rapid Automatized Naming (RAN), word reading, nonword reading, and sentence reading. Subsequently, we examined the relationship between electrophysiological indicators (MMN and LDN) and the time of each reading task in the STRAW-R using multiple linear regression analyses. We observed significant regression models for RAN and word-reading tasks. For the RAN task, the MMN amplitude of the standard /la/ and deviant /ra/ predicted reading time, indicating that participants with larger MMN amplitude for speech sounds have poorer RAN scores (R2 = 0.12, F(4,45) = 2.67, p = 0.04; β = -0.35, p = 0.02). In the word reading task, the LDN amplitude of the standard /la/ and deviant /ra/ predicted that participants with larger LDN amplitudes had slower reading times (R2 = 0.13, F(4,45) = 2.86, p = 0.03; β = -0.38, p < 0.01). The regression results revealed that AERPs elicited by speech sound differences can predict reading processing time. Phonological processing difficulties have previously been suggested to predict learning the read in children (Snowling, 1995). The current findings suggest that individual phonological processing differences can also affect reading processing speed in non-dyslexic adults. The opposite polarity of the current results compared to prior studies reporting attenuated MMN and LDN amplitudes in dyslexics (Hommet et al., 2009; Gu and Bi, 2020) likely stems from the use of a foreign phoneme contrast in this study. The MMN amplitude induced by speech sounds is reduced when stimuli are distant from the typical phonemes of the native language (Näätänen et al., 1997). Accordingly, participants with larger AERPs to foreign phoneme contrasts in this study may have had unstable native phoneme representations. Overall, we propose that neurophysiological responses to foreign phoneme contrasts represent a novel index for investigating reading processes in adults.
Topic Areas: Reading,